diff --git a/prediction-service/README.md b/prediction-service/README.md
index acde50e..7f0561a 100644
--- a/prediction-service/README.md
+++ b/prediction-service/README.md
@@ -45,34 +45,34 @@ curl -X 'POST' \
   -H 'Content-Type: application/json' \
   -d '{
   "customer": {
-    "customer_id": 4444,
-    "age_range": "46-55",
-    "marital_status": "Married",
-    "family_size": 5,
-    "no_of_children": 3,
-    "income_bracket": 3,
-    "gender": "M",
-    "mean_discount_used": -1.83,
-    "total_discount_used": -7548.79,
-    "total_unique_items_bought": 2040,
-    "total_quantity_bought": 4314,
-    "mean_quantity_bought": 1.04,
-    "mean_selling_price_paid": 61.27,
-    "total_coupons_redeemed": 220,
-    "total_price_paid": 253295.6
+    "customer_id": 54,
+    "age": "old",
+    "credit": 1,
+    "gender": "F",
+    "mean_product_price": 13.63,
+    "unique_coupons_used": 369,
+    "mean_discount_used": 11.93,
+    "unique_items_bought": 1934,
+    "total_items_bought": 42265
   },
   "coupons": [
     {
-      "coupon_id": 1111,
-      "item_selling_price": 70.88,
-      "coupon_discount": -35.62,
-      "item_category": "Men"
+      "coupon_id": 1,
+      "mean_item_selling_price": 7.06,
+      "coupon_discount": 50,
+      "category": "",
+      "how_many_products": 2,
+      "coupon_type": "buy_more",
+      "days_valid": 24
     },
     {
-      "coupon_id": 2323,
-      "item_selling_price": 75.51,
-      "coupon_discount": -26.71,
-      "item_category": "Sport"
+      "coupon_id": 2,
+      "mean_item_selling_price": 7.06,
+      "coupon_discount": 3.78,
+      "category": "",
+      "how_many_products": 2,
+      "coupon_type": "buy_all",
+      "days_valid": 20
     }
   ]
 }'
@@ -83,14 +83,14 @@ Example response:
 ```
 [
   {
-    "coupon_id": 1111,
-    "customer_id": 4444,
-    "prediction": 0.85
+    "coupon_id": 2,
+    "customer_id": 54,
+    "prediction": 0.0409353164
   },
   {
-    "coupon_id": 2323,
-    "customer_id": 4444,
-    "prediction": 0.8
+    "coupon_id": 1,
+    "customer_id": 54,
+    "prediction": 0.0311633506
   }
 ]
 ```
diff --git a/prediction-service/app/encoder.py b/prediction-service/app/encoder.py
index 6dfbcc7..a30d6b3 100644
--- a/prediction-service/app/encoder.py
+++ b/prediction-service/app/encoder.py
@@ -1,14 +1,13 @@
 import pandas
 
-# from .model import PredictionInput
 from app.model import PredictionInput
 
 
 class DataEncoder:
-    _age_range = {'18-25': 0, '26-35': 1, '36-45': 2, '46-55': 3, '56-70': 4, '70+': 5}
-    _marital_status = {'Married': 0, 'Single': 1}
-    _gender = {'F': 0, 'M': 1}
+    _gender = ['F', 'M']
+    _age = ['young', 'mid', 'old']
     _categories = ['Boys', 'Girls', 'Men', 'Sports', 'Women']
+    _coupon_types = ['biy_all', 'boy_more', 'department', 'just_discount']
 
     @classmethod
     def encode(cls, input: PredictionInput) -> pandas.DataFrame:
@@ -16,29 +15,37 @@ def encode(cls, input: PredictionInput) -> pandas.DataFrame:
         for coupon in input.coupons:
             row = {
                 'customer_id': input.customer.customer_id,
-                'age_range': cls._age_range[input.customer.age_range],
-                'marital_status': cls._marital_status[input.customer.marital_status],
-                'family_size': input.customer.family_size,
-                'no_of_children': input.customer.no_of_children,
-                'income_bracket': input.customer.income_bracket,
-                'gender': cls._gender[input.customer.gender],
-                'mean_discount_per_cust': input.customer.mean_discount_used,
-                'unique_items_per_cust': input.customer.total_unique_items_bought,
-                'mean_quantity_per_cust': input.customer.mean_quantity_bought,
-                'mean_selling_price_per_cust': input.customer.mean_selling_price_paid,
-                'total_discount_per_cust': input.customer.total_discount_used,
-                'total_coupons_used_per_cust': input.customer.total_coupons_redeemed,
-                'total_quantity_per_cust': input.customer.total_quantity_bought,
-                'total_selling_price_per_cust': input.customer.total_price_paid,
+                'cust_credit': input.customer.credit,
+                'cust_mean_product_price': input.customer.mean_product_price,
+                'cust_unique_coupons_used': input.customer.unique_coupons_used,
+                'cust_mean_discount': input.customer.mean_discount_used,
+                'cust_unique_products_bought': input.customer.unique_items_bought,
+                'cust_total_products_bougth': input.customer.total_items_bought,
                 'coupon_id': coupon.coupon_id,
                 'coupon_discount': coupon.coupon_discount,
-                'item_selling_price': coupon.item_selling_price
+                'coupon_how_many': coupon.how_many_products,
+                'coupon_days_valid': coupon.days_valid,
+                'coupon_mean_prod_price': coupon.mean_item_selling_price
             }
-            row.update(cls._encode_category(coupon.item_category))
+            # row.update(cls._encode_category(coupon.item_category))
+            row.update(cls._encode_age(input.customer.age))
+            row.update(cls._encode_gender(input.customer.gender))
+            row.update(cls._encode_coupon_type(coupon.coupon_type))
             rows.append(row)
-
         return pandas.DataFrame(rows)
 
     @classmethod
     def _encode_category(cls, category):
         return {f'category_{c}': 1 if category == c else 0 for c in cls._categories}
+
+    @classmethod
+    def _encode_age(cls, age):
+        return {f'cust_age_{a}': 1 if age == a else 0 for a in cls._age}
+
+    @classmethod
+    def _encode_gender(cls, gender):
+        return {f'cust_gender_{g}': 1 if gender == g else 0 for g in cls._gender}
+
+    @classmethod
+    def _encode_coupon_type(cls, coupon_type):
+        return {f'coupon_type_{t}': 1 if coupon_type == t else 0 for t in cls._coupon_types}
diff --git a/prediction-service/app/main.py b/prediction-service/app/main.py
index 712137d..bbdd62e 100644
--- a/prediction-service/app/main.py
+++ b/prediction-service/app/main.py
@@ -15,11 +15,6 @@
 )
 
 
-@app.get('/')
-async def root():
-    return {'message': 'Hello World'}
-
-
 @app.post('/score', response_model=List[PredictionOutput])
 def score_coupon(
     input_data: PredictionInput,
diff --git a/prediction-service/app/model.py b/prediction-service/app/model.py
index f0a3db7..30f5330 100644
--- a/prediction-service/app/model.py
+++ b/prediction-service/app/model.py
@@ -5,28 +5,24 @@
 
 class Coupon(BaseModel):
     coupon_id: int
-    item_selling_price: float
+    mean_item_selling_price: float
     coupon_discount: float
-    item_category: str  # TODO Enum
+    category: str  # TODO this is not used (for now)
+    how_many_products: int
+    coupon_type: str
+    days_valid: int
 
 
 class Customer(BaseModel):
     customer_id: int
-    age_range: str  # TODO Enum
-    marital_status: str # TODO Enum
-    family_size: int
-    no_of_children: int
-    income_bracket: int
+    age: str  # TODO Enum
+    credit: int
     gender: str  # TODO Enum
+    mean_product_price: float
+    unique_coupons_used: int
     mean_discount_used: float
-    total_discount_used: float
-    total_unique_items_bought: int
-    total_quantity_bought: int
-    mean_quantity_bought: float
-    mean_selling_price_paid: float
-    total_coupons_redeemed: int
-    total_price_paid: float
-
+    unique_items_bought: int
+    total_items_bought: int
 
 class PredictionInput(BaseModel):
     customer: Customer
diff --git a/prediction-service/app/model_store/pickled_model_gbm_no_balancing b/prediction-service/app/model_store/pickled_model_gbm_no_balancing
new file mode 100644
index 0000000..3ff584e
Binary files /dev/null and b/prediction-service/app/model_store/pickled_model_gbm_no_balancing differ
diff --git a/prediction-service/app/scorer.py b/prediction-service/app/scorer.py
index bf20246..8047c00 100644
--- a/prediction-service/app/scorer.py
+++ b/prediction-service/app/scorer.py
@@ -14,11 +14,11 @@ def score(self, input_df: pandas.DataFrame) -> pandas.DataFrame:
         input_df.drop(['customer_id', 'coupon_id'], axis=1, inplace=True)
         probs = self._model.predict_proba(input_df)[:, 1]
         output_df['prediction'] = probs.round(decimals=10)
-        output_df.sort_values(by='prediction', ascending=False)
+        output_df.sort_values(by='prediction', ascending=False, inplace=True)
         return output_df
 
 
 def get_scorer():
-    model_path = 'app/model_store/scikit_classifier'
+    model_path = 'app/model_store/pickled_model_gbm_no_balancing'
     with open(model_path, 'rb') as f:
         return Scorer(pickle.load(f))
diff --git a/training-with-artificial-data/01_data_prep_v1.ipynb b/training-with-artificial-data/01_data_prep_v1.ipynb
new file mode 100644
index 0000000..be32e41
--- /dev/null
+++ b/training-with-artificial-data/01_data_prep_v1.ipynb
@@ -0,0 +1,3523 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "antique-place",
+   "metadata": {},
+   "source": [
+    "# Prepare data for training dataframe\n",
+    "\n",
+    "The aim is to produce a training table where each `order_id` is mapped to all coupons available in the departments from which products were selected in that order on the date the order was made, along with information (`True`/`False`) if such available coupon was used in that order.\n",
+    "\n",
+    "The steps are as follows:\n",
+    "1. Create `coupon_date_department` table which maps coupons to applicable departments, to coupon validity dates\n",
+    "2. Create `order_date_department` table which maps orders to departments from which products were ordered, to order dates.\n",
+    "3. Create `order_coupons_available` by merging the tables above on date and department - this will create a table which maps orders and coupons available at that order date, but limited by departments (i.e. rows are selected only if order contained products from a department for which there was a coupon)\n",
+    "4. Create `order_coupons_used` - by dropping columns from `order_details`, leaving only a mapping between order and coupons used in that order.\n",
+    "5. Cobine 3 and 4 to create a dataset with coupons and orders mapped, with information whether coupon was used or not.\n",
+    "6. Add details about customers (such as age, gender, etc., and also some statistics about customer purchases)\n",
+    "7. Add details about coupons (type, discount, mean product pric, etc.)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "seven-middle",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import datetime\n",
+    "import os\n",
+    "\n",
+    "from IPython.display import Image\n",
+    "import numpy as np\n",
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "encouraging-subscription",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<IPython.core.display.Image object>"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "Image('data_diagram.png')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "radical-mining",
+   "metadata": {},
+   "source": [
+    "#### Step 0. Read and prepare tables"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "worth-situation",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dir = 'data_0408_0'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "recorded-intellectual",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1033 entries, 0 to 1032\n",
+      "Data columns (total 7 columns):\n",
+      " #   Column      Non-Null Count  Dtype \n",
+      "---  ------      --------------  ----- \n",
+      " 0   id          1033 non-null   int64 \n",
+      " 1   type        1033 non-null   object\n",
+      " 2   department  62 non-null     object\n",
+      " 3   discount    1033 non-null   int64 \n",
+      " 4   how_many    1033 non-null   int64 \n",
+      " 5   start_date  1033 non-null   object\n",
+      " 6   end_date    1033 non-null   object\n",
+      "dtypes: int64(3), object(4)\n",
+      "memory usage: 56.6+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "coupons = pd.read_csv(os.path.join(data_dir, 'coupons.csv'))\n",
+    "coupons.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "racial-complaint",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "coupons.rename(columns={'id': 'coupon_id'}, inplace=True)\n",
+    "coupons.start_date = pd.to_datetime(coupons.start_date, format='%Y-%m-%d')\n",
+    "coupons.end_date = pd.to_datetime(coupons.end_date, format='%Y-%m-%d')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "prescription-transsexual",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1812 entries, 0 to 1811\n",
+      "Data columns (total 2 columns):\n",
+      " #   Column      Non-Null Count  Dtype\n",
+      "---  ------      --------------  -----\n",
+      " 0   coupon_id   1812 non-null   int64\n",
+      " 1   product_id  1812 non-null   int64\n",
+      "dtypes: int64(2)\n",
+      "memory usage: 28.4 KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "coupon_product = pd.read_csv(os.path.join(data_dir, 'coupon_product.csv'))\n",
+    "coupon_product.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "martial-failure",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 3000 entries, 0 to 2999\n",
+      "Data columns (total 8 columns):\n",
+      " #   Column       Non-Null Count  Dtype  \n",
+      "---  ------       --------------  -----  \n",
+      " 0   id           3000 non-null   int64  \n",
+      " 1   name         3000 non-null   object \n",
+      " 2   category     3000 non-null   object \n",
+      " 3   sizes        3000 non-null   object \n",
+      " 4   vendor       3000 non-null   object \n",
+      " 5   description  3000 non-null   object \n",
+      " 6   buy_price    3000 non-null   float64\n",
+      " 7   department   3000 non-null   object \n",
+      "dtypes: float64(1), int64(1), object(6)\n",
+      "memory usage: 187.6+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "products = pd.read_csv(os.path.join(data_dir, 'products.csv'))\n",
+    "products.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "miniature-pension",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "products.rename(columns={'id': 'product_id'}, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "weekly-heater",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 308313 entries, 0 to 308312\n",
+      "Data columns (total 3 columns):\n",
+      " #   Column       Non-Null Count   Dtype \n",
+      "---  ------       --------------   ----- \n",
+      " 0   id           308313 non-null  int64 \n",
+      " 1   customer_id  308313 non-null  int64 \n",
+      " 2   order_date   308313 non-null  object\n",
+      "dtypes: int64(2), object(1)\n",
+      "memory usage: 7.1+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "orders = pd.read_csv(os.path.join(data_dir, 'orders.csv'))\n",
+    "orders.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "pursuant-scholarship",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>order_date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>166938</th>\n",
+       "      <td>166939</td>\n",
+       "      <td>212</td>\n",
+       "      <td>2011-08-15 16:24:53</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>183539</th>\n",
+       "      <td>183540</td>\n",
+       "      <td>274</td>\n",
+       "      <td>2011-10-14 10:57:56</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>274010</th>\n",
+       "      <td>274011</td>\n",
+       "      <td>158</td>\n",
+       "      <td>2012-08-30 11:15:48</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>46785</th>\n",
+       "      <td>46786</td>\n",
+       "      <td>740</td>\n",
+       "      <td>2010-06-15 15:19:19</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>153207</th>\n",
+       "      <td>153208</td>\n",
+       "      <td>660</td>\n",
+       "      <td>2011-06-28 09:54:40</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            id  customer_id           order_date\n",
+       "166938  166939          212  2011-08-15 16:24:53\n",
+       "183539  183540          274  2011-10-14 10:57:56\n",
+       "274010  274011          158  2012-08-30 11:15:48\n",
+       "46785    46786          740  2010-06-15 15:19:19\n",
+       "153207  153208          660  2011-06-28 09:54:40"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "orders.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "intense-mistake",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "orders.rename(columns={'id': 'order_id', 'order_date': 'date'}, inplace=True)\n",
+    "orders.date = pd.to_datetime(orders.date, format='%Y-%m-%d').dt.date"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "unexpected-desperate",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 2297759 entries, 0 to 2297758\n",
+      "Data columns (total 7 columns):\n",
+      " #   Column            Dtype  \n",
+      "---  ------            -----  \n",
+      " 0   id                int64  \n",
+      " 1   order_id          int64  \n",
+      " 2   product_id        int64  \n",
+      " 3   quantity_ordered  int64  \n",
+      " 4   original_price    float64\n",
+      " 5   buy_price         float64\n",
+      " 6   coupon_id         float64\n",
+      "dtypes: float64(3), int64(4)\n",
+      "memory usage: 122.7 MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "order_details = pd.read_csv(os.path.join(data_dir, 'order_details.csv'))\n",
+    "order_details.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "literary-asbestos",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_details.drop(['id'], axis=1, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "western-clear",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>product_id</th>\n",
+       "      <th>quantity_ordered</th>\n",
+       "      <th>original_price</th>\n",
+       "      <th>buy_price</th>\n",
+       "      <th>coupon_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1789646</th>\n",
+       "      <td>240717</td>\n",
+       "      <td>2642</td>\n",
+       "      <td>4</td>\n",
+       "      <td>7.63</td>\n",
+       "      <td>7.6300</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>940699</th>\n",
+       "      <td>125187</td>\n",
+       "      <td>859</td>\n",
+       "      <td>2</td>\n",
+       "      <td>3.09</td>\n",
+       "      <td>2.9355</td>\n",
+       "      <td>417.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>763067</th>\n",
+       "      <td>100890</td>\n",
+       "      <td>1877</td>\n",
+       "      <td>4</td>\n",
+       "      <td>81.99</td>\n",
+       "      <td>45.0945</td>\n",
+       "      <td>338.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>672868</th>\n",
+       "      <td>88426</td>\n",
+       "      <td>898</td>\n",
+       "      <td>7</td>\n",
+       "      <td>3.60</td>\n",
+       "      <td>3.6000</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1142583</th>\n",
+       "      <td>153580</td>\n",
+       "      <td>2477</td>\n",
+       "      <td>3</td>\n",
+       "      <td>8.12</td>\n",
+       "      <td>4.1412</td>\n",
+       "      <td>514.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  product_id  quantity_ordered  original_price  buy_price  \\\n",
+       "1789646    240717        2642                 4            7.63     7.6300   \n",
+       "940699     125187         859                 2            3.09     2.9355   \n",
+       "763067     100890        1877                 4           81.99    45.0945   \n",
+       "672868      88426         898                 7            3.60     3.6000   \n",
+       "1142583    153580        2477                 3            8.12     4.1412   \n",
+       "\n",
+       "         coupon_id  \n",
+       "1789646        NaN  \n",
+       "940699       417.0  \n",
+       "763067       338.0  \n",
+       "672868         NaN  \n",
+       "1142583      514.0  "
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_details.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "substantial-boards",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1000 entries, 0 to 999\n",
+      "Data columns (total 11 columns):\n",
+      " #   Column       Non-Null Count  Dtype \n",
+      "---  ------       --------------  ----- \n",
+      " 0   id           1000 non-null   int64 \n",
+      " 1   name         1000 non-null   object\n",
+      " 2   gender       1000 non-null   object\n",
+      " 3   age          1000 non-null   int64 \n",
+      " 4   phone        1000 non-null   object\n",
+      " 5   address      1000 non-null   object\n",
+      " 6   city         1000 non-null   object\n",
+      " 7   state        1000 non-null   object\n",
+      " 8   postalCode   1000 non-null   int64 \n",
+      " 9   country      1000 non-null   object\n",
+      " 10  creditLimit  1000 non-null   int64 \n",
+      "dtypes: int64(4), object(7)\n",
+      "memory usage: 86.1+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "customers = pd.read_csv(os.path.join(data_dir, 'customers.csv'))\n",
+    "customers.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "colonial-enforcement",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "customers.rename(columns={'id': 'customer_id'}, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "identical-financing",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>name</th>\n",
+       "      <th>gender</th>\n",
+       "      <th>age</th>\n",
+       "      <th>phone</th>\n",
+       "      <th>address</th>\n",
+       "      <th>city</th>\n",
+       "      <th>state</th>\n",
+       "      <th>postalCode</th>\n",
+       "      <th>country</th>\n",
+       "      <th>creditLimit</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>510</th>\n",
+       "      <td>511</td>\n",
+       "      <td>Travarious Wolff</td>\n",
+       "      <td>M</td>\n",
+       "      <td>37</td>\n",
+       "      <td>(359)461-8875</td>\n",
+       "      <td>138 Davis Road</td>\n",
+       "      <td>Watertown</td>\n",
+       "      <td>Massachusetts</td>\n",
+       "      <td>2272</td>\n",
+       "      <td>US</td>\n",
+       "      <td>5483</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>757</th>\n",
+       "      <td>758</td>\n",
+       "      <td>Jhade Stracke</td>\n",
+       "      <td>F</td>\n",
+       "      <td>56</td>\n",
+       "      <td>(282)145-1638</td>\n",
+       "      <td>47 Central Trail</td>\n",
+       "      <td>Absaraka</td>\n",
+       "      <td>North Dakota</td>\n",
+       "      <td>58002</td>\n",
+       "      <td>US</td>\n",
+       "      <td>4845</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>858</th>\n",
+       "      <td>859</td>\n",
+       "      <td>Treydan MD</td>\n",
+       "      <td>M</td>\n",
+       "      <td>19</td>\n",
+       "      <td>(793)753-7311</td>\n",
+       "      <td>119 View Court</td>\n",
+       "      <td>Savanna</td>\n",
+       "      <td>Oklahoma</td>\n",
+       "      <td>74565</td>\n",
+       "      <td>US</td>\n",
+       "      <td>9810</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>74</th>\n",
+       "      <td>75</td>\n",
+       "      <td>Evelynn DVM</td>\n",
+       "      <td>F</td>\n",
+       "      <td>66</td>\n",
+       "      <td>(945)774-6546</td>\n",
+       "      <td>160 Cherry Court</td>\n",
+       "      <td>Chewelah</td>\n",
+       "      <td>Washington</td>\n",
+       "      <td>99109</td>\n",
+       "      <td>US</td>\n",
+       "      <td>9202</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>815</th>\n",
+       "      <td>816</td>\n",
+       "      <td>Saedie Sr.</td>\n",
+       "      <td>F</td>\n",
+       "      <td>36</td>\n",
+       "      <td>(188)487-2981</td>\n",
+       "      <td>323 4th Trail</td>\n",
+       "      <td>Plattsburgh</td>\n",
+       "      <td>New York</td>\n",
+       "      <td>12901</td>\n",
+       "      <td>US</td>\n",
+       "      <td>1452</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "     customer_id              name gender  age          phone  \\\n",
+       "510          511  Travarious Wolff      M   37  (359)461-8875   \n",
+       "757          758     Jhade Stracke      F   56  (282)145-1638   \n",
+       "858          859        Treydan MD      M   19  (793)753-7311   \n",
+       "74            75       Evelynn DVM      F   66  (945)774-6546   \n",
+       "815          816        Saedie Sr.      F   36  (188)487-2981   \n",
+       "\n",
+       "              address         city          state  postalCode country  \\\n",
+       "510    138 Davis Road    Watertown  Massachusetts        2272      US   \n",
+       "757  47 Central Trail     Absaraka   North Dakota       58002      US   \n",
+       "858    119 View Court      Savanna       Oklahoma       74565      US   \n",
+       "74   160 Cherry Court     Chewelah     Washington       99109      US   \n",
+       "815     323 4th Trail  Plattsburgh       New York       12901      US   \n",
+       "\n",
+       "     creditLimit  \n",
+       "510         5483  \n",
+       "757         4845  \n",
+       "858         9810  \n",
+       "74          9202  \n",
+       "815         1452  "
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "customers.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "crazy-festival",
+   "metadata": {},
+   "source": [
+    "#### Step 1. Create `coupon_date_department` table\n",
+    "\n",
+    "1. Create `coupon_department` mapping coupons to departments for which they are valid\n",
+    "2. Create `coupon_dates` mapping coupons to all dates on which they were valid\n",
+    "3. Merge the two into `coupon_department_dates`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "italian-apple",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>department</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>Men</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>Sport</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1598</th>\n",
+       "      <td>1031</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1599</th>\n",
+       "      <td>1031</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1600</th>\n",
+       "      <td>1031</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1601</th>\n",
+       "      <td>1032</td>\n",
+       "      <td>Men</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1602</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1603 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      coupon_id department\n",
+       "0             1      Girls\n",
+       "1             2       Boys\n",
+       "2             3       Boys\n",
+       "3             4        Men\n",
+       "4             5      Sport\n",
+       "...         ...        ...\n",
+       "1598       1031       Boys\n",
+       "1599       1031      Women\n",
+       "1600       1031      Girls\n",
+       "1601       1032        Men\n",
+       "1602       1033      Women\n",
+       "\n",
+       "[1603 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 18,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Step 1.1 Map coupons to departments\n",
+    "coupon_department = pd.merge(coupon_product, products[['product_id', 'department']], on='product_id')\\\n",
+    "                      .drop(['product_id'], axis=1).drop_duplicates()\n",
+    "\n",
+    "# Coupon_product does not include coupons valid for all products in a department, add this info here\n",
+    "department_coupons = coupons.loc[coupons.type == 'department'][['coupon_id', 'department']]\n",
+    "coupon_department = coupon_department.append(department_coupons).sort_values(by='coupon_id').reset_index(drop=True)\n",
+    "coupon_department"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "exact-exclusion",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Validate that all coupons are present\n",
+    "assert len(coupon_department.coupon_id.unique()) == max(coupon_department.coupon_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "handled-maximum",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Start: 2010-01-01, end: 2013-01-23. 1119 days\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Step 1.2 Map coupons to dates on which they were valid\n",
+    "coupon_dates = coupons.drop(['type', 'department', 'discount', 'how_many'], axis=1)\n",
+    "# Get the earliest and latest date in the dataset\n",
+    "start = coupon_dates.start_date.min()\n",
+    "end = coupon_dates.end_date.max()\n",
+    "days = (end-start).days + 1\n",
+    "print(f'Start: {start.date()}, end: {end.date()}. {days} days')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "connected-swiss",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Create a dataframe with row for each day from the earliest to the latest date in the set\n",
+    "all_dates = pd.DataFrame(pd.date_range(start=start, end=end, freq='D'), columns=['date'])\n",
+    "assert days == len(all_dates)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "future-external",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-02</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-03</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-04</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-05</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1155914</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-11</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1155915</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-12</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1155916</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-13</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1155917</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-14</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1155918</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-15</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>15590 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         coupon_id       date\n",
+       "0                1 2010-01-01\n",
+       "1                1 2010-01-02\n",
+       "2                1 2010-01-03\n",
+       "3                1 2010-01-04\n",
+       "4                1 2010-01-05\n",
+       "...            ...        ...\n",
+       "1155914       1033 2013-01-11\n",
+       "1155915       1033 2013-01-12\n",
+       "1155916       1033 2013-01-13\n",
+       "1155917       1033 2013-01-14\n",
+       "1155918       1033 2013-01-15\n",
+       "\n",
+       "[15590 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Step 1.2.1: Perform a cross join of `all_dates` and `coupon_dates` - which contains info on validity periods\n",
+    "coupon_dates['key'] = 1\n",
+    "all_dates['key'] = 1\n",
+    "coupon_dates = pd.merge(coupon_dates, all_dates, on='key').drop('key', axis=1)\n",
+    "\n",
+    "# Step 1.2.2 Drop rows where a date does not fall within the validity period of a coupon\n",
+    "coupon_dates = coupon_dates[(coupon_dates['date'] >= coupon_dates['start_date']) & \\\n",
+    "                            (coupon_dates['date'] <= coupon_dates['end_date'])]\n",
+    "coupon_dates.drop(['start_date', 'end_date'], axis=1, inplace=True)\n",
+    "coupon_dates"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "activated-entrance",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Validate coupon_dates is consistent with the original data in terms of coupon validity dates\n",
+    "coupons['days_valid'] = (coupons.end_date - coupons.start_date).dt.days + 1\n",
+    "df = pd.merge(coupons, coupon_dates.groupby(by='coupon_id').count().rename(columns={'date': 'days_valid'}), on='coupon_id')\n",
+    "assert 0 == len(df.loc[df.days_valid_x != df.days_valid_y])\n",
+    "coupons.drop(['days_valid'], axis=1, inplace=True)\n",
+    "\n",
+    "# Validate no coupon has been lost\n",
+    "assert len(coupons) == len(coupon_dates.coupon_id.unique())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "charitable-certificate",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "coupon_id    14.0\n",
+       "dtype: float64"
+      ]
+     },
+     "execution_count": 24,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "coupon_dates.groupby('date').count().median()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "ignored-window",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>date</th>\n",
+       "      <th>department</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-02</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-03</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-04</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-05</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>24395</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-11</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>24396</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-12</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>24397</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-13</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>24398</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-14</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>24399</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>2013-01-15</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>24400 rows × 3 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "       coupon_id        date department\n",
+       "0              1  2010-01-01      Girls\n",
+       "1              1  2010-01-02      Girls\n",
+       "2              1  2010-01-03      Girls\n",
+       "3              1  2010-01-04      Girls\n",
+       "4              1  2010-01-05      Girls\n",
+       "...          ...         ...        ...\n",
+       "24395       1033  2013-01-11      Women\n",
+       "24396       1033  2013-01-12      Women\n",
+       "24397       1033  2013-01-13      Women\n",
+       "24398       1033  2013-01-14      Women\n",
+       "24399       1033  2013-01-15      Women\n",
+       "\n",
+       "[24400 rows x 3 columns]"
+      ]
+     },
+     "execution_count": 25,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Step 1.3 Merge coupon_department and coupon_dates\n",
+    "coupon_date_department = pd.merge(coupon_dates, coupon_department, on='coupon_id', how='left')\n",
+    "coupon_date_department.date = coupon_date_department.date.dt.date\n",
+    "coupon_date_department"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "confidential-driver",
+   "metadata": {},
+   "source": [
+    "#### Step 2. Create `order_date_department` table\n",
+    "\n",
+    "1. Map `order_id` and product `department` -> `order_department`\n",
+    "2. Add dates by joining with `orders` -> `order_date_department`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "fitted-difference",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>department</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Sport</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>2</td>\n",
+       "      <td>Girls</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>2</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297745</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297748</th>\n",
+       "      <td>308312</td>\n",
+       "      <td>Women</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297749</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Boys</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297751</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Sport</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297753</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Men</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1096057 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id department\n",
+       "0               1       Boys\n",
+       "1               1      Sport\n",
+       "2               1      Girls\n",
+       "3               2      Girls\n",
+       "4               2       Boys\n",
+       "...           ...        ...\n",
+       "2297745    308311      Women\n",
+       "2297748    308312      Women\n",
+       "2297749    308313       Boys\n",
+       "2297751    308313      Sport\n",
+       "2297753    308313        Men\n",
+       "\n",
+       "[1096057 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 26,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_department = pd.merge(order_details[['order_id', 'product_id']], products[['product_id', 'department']],\n",
+    "                            on='product_id', how='left').drop(['product_id'], axis=1).drop_duplicates()\n",
+    "order_department"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "mature-feedback",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>department</th>\n",
+       "      <th>date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Boys</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Sport</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>Girls</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>2</td>\n",
+       "      <td>Girls</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>2</td>\n",
+       "      <td>Boys</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1096052</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>Women</td>\n",
+       "      <td>2012-12-30</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1096053</th>\n",
+       "      <td>308312</td>\n",
+       "      <td>Women</td>\n",
+       "      <td>2012-12-30</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1096054</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Boys</td>\n",
+       "      <td>2012-12-30</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1096055</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Sport</td>\n",
+       "      <td>2012-12-30</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1096056</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>Men</td>\n",
+       "      <td>2012-12-30</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1096057 rows × 3 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id department        date\n",
+       "0               1       Boys  2010-01-01\n",
+       "1               1      Sport  2010-01-01\n",
+       "2               1      Girls  2010-01-01\n",
+       "3               2      Girls  2010-01-01\n",
+       "4               2       Boys  2010-01-01\n",
+       "...           ...        ...         ...\n",
+       "1096052    308311      Women  2012-12-30\n",
+       "1096053    308312      Women  2012-12-30\n",
+       "1096054    308313       Boys  2012-12-30\n",
+       "1096055    308313      Sport  2012-12-30\n",
+       "1096056    308313        Men  2012-12-30\n",
+       "\n",
+       "[1096057 rows x 3 columns]"
+      ]
+     },
+     "execution_count": 27,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_date_department = pd.merge(order_department, orders[['order_id', 'date']], on='order_id', how='right')\n",
+    "order_date_department"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dynamic-jumping",
+   "metadata": {},
+   "source": [
+    "#### Step 3. Create `order_coupons_available`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "adjacent-mexican",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>coupon_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>3.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>6.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>12.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>13.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4866788</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1015.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4866789</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1019.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4866791</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1028.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4866794</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1029.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4866795</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1032.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>3374643 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  coupon_id\n",
+       "0               1        2.0\n",
+       "1               1        3.0\n",
+       "2               1        6.0\n",
+       "3               1       12.0\n",
+       "4               1       13.0\n",
+       "...           ...        ...\n",
+       "4866788    308313     1015.0\n",
+       "4866789    308313     1019.0\n",
+       "4866791    308313     1028.0\n",
+       "4866794    308313     1029.0\n",
+       "4866795    308313     1032.0\n",
+       "\n",
+       "[3374643 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 28,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupons_available = pd.merge(order_date_department, coupon_date_department, on=['date', 'department'], how='left')\\\n",
+    "                            .drop(['department', 'date'], axis=1).drop_duplicates()\n",
+    "order_coupons_available"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "neural-spouse",
+   "metadata": {},
+   "source": [
+    "#### Step 4. Create `order_coupons_used`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "id": "corresponding-evans",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>coupon_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>2</td>\n",
+       "      <td>14.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>3</td>\n",
+       "      <td>8.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>44</th>\n",
+       "      <td>7</td>\n",
+       "      <td>2.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>58</th>\n",
+       "      <td>9</td>\n",
+       "      <td>11.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>59</th>\n",
+       "      <td>9</td>\n",
+       "      <td>13.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297736</th>\n",
+       "      <td>308310</td>\n",
+       "      <td>1020.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297740</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>1020.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297744</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>1018.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297749</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1020.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2297750</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1021.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>341463 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  coupon_id\n",
+       "8               2       14.0\n",
+       "11              3        8.0\n",
+       "44              7        2.0\n",
+       "58              9       11.0\n",
+       "59              9       13.0\n",
+       "...           ...        ...\n",
+       "2297736    308310     1020.0\n",
+       "2297740    308311     1020.0\n",
+       "2297744    308311     1018.0\n",
+       "2297749    308313     1020.0\n",
+       "2297750    308313     1021.0\n",
+       "\n",
+       "[341463 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 29,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupon_used = order_details[['order_id', 'coupon_id']].dropna().drop_duplicates()\n",
+    "order_coupon_used"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "solar-mounting",
+   "metadata": {},
+   "source": [
+    "#### Step 5. Combine `order_coupons_available` and `order_coupons_used` to get final info on which coupons were used and which, although they were available, were ignored"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "id": "worthy-austria",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupon_used.set_index(['order_id', 'coupon_id'], inplace=True)\n",
+    "order_coupons_available.set_index(['order_id', 'coupon_id'], inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "id": "tough-proportion",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupons_available['coupon_used'] = order_coupons_available.index.isin(order_coupon_used.index)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "capital-twelve",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupons = order_coupons_available.reset_index().dropna().reset_index(drop=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "id": "portuguese-alias",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "False    0.89864\n",
+       "True     0.10136\n",
+       "Name: coupon_used, dtype: float64"
+      ]
+     },
+     "execution_count": 33,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupons.coupon_used.value_counts(normalize=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "informal-launch",
+   "metadata": {},
+   "source": [
+    "## Prepare customer data\n",
+    "- From `customers` table, get:\n",
+    "    - gender\n",
+    "    - age bracket: young < 30, medium >= 30 & < 60, old >= 60\n",
+    "    - credit bracket: '0' < 3000, '1' >= 3000 & < 6000, '2' >= 6000\n",
+    "- From `order_details` get:\n",
+    "    - number of unique products bought\n",
+    "    - mean price paid\n",
+    "    - total quantity bought\n",
+    "    - total coupons used\n",
+    "    - mean discount used"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "id": "billion-ceramic",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "customer_demo = customers.drop(['name', 'phone', 'address', 'city', 'state', 'postalCode', 'country'], axis=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 35,
+   "id": "difficult-mouth",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "customer_demo['age_bracket'] = None\n",
+    "customer_demo.loc[customer_demo.age < 30, 'age_bracket'] = 'young'\n",
+    "customer_demo.loc[(customer_demo.age >= 30) & (customer_demo.age < 60), 'age_bracket'] = 'mid'\n",
+    "customer_demo.loc[(customer_demo.age >= 60), 'age_bracket'] = 'old'\n",
+    "customer_demo.age = customer_demo.age_bracket\n",
+    "customer_demo.drop(['age_bracket'], axis=1, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 36,
+   "id": "enormous-finnish",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "customer_demo['credit'] = None\n",
+    "customer_demo.loc[customer_demo.creditLimit < 3000, 'credit'] = 0\n",
+    "customer_demo.loc[(customer_demo.creditLimit >= 3000) & (customer_demo.creditLimit < 6000), 'credit'] = 1\n",
+    "customer_demo.loc[(customer_demo.creditLimit >= 6000), 'credit'] = 2\n",
+    "customer_demo.drop(['creditLimit'], axis=1, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 37,
+   "id": "going-blowing",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1000 entries, 0 to 999\n",
+      "Data columns (total 4 columns):\n",
+      " #   Column       Non-Null Count  Dtype \n",
+      "---  ------       --------------  ----- \n",
+      " 0   customer_id  1000 non-null   int64 \n",
+      " 1   gender       1000 non-null   object\n",
+      " 2   age          1000 non-null   object\n",
+      " 3   credit       1000 non-null   object\n",
+      "dtypes: int64(1), object(3)\n",
+      "memory usage: 31.4+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "customer_demo.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 38,
+   "id": "ideal-depression",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "customer_demo.rename(\n",
+    "    columns={'age': 'cust_age', 'credit': 'cust_credit', 'gender': 'cust_gender'},\n",
+    "    inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 39,
+   "id": "confused-hampton",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>product_id</th>\n",
+       "      <th>quantity_ordered</th>\n",
+       "      <th>buy_price</th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>discount</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>905646</th>\n",
+       "      <td>1181</td>\n",
+       "      <td>8</td>\n",
+       "      <td>6.4400</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>692</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2192443</th>\n",
+       "      <td>2659</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.0400</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>675</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1130421</th>\n",
+       "      <td>1870</td>\n",
+       "      <td>3</td>\n",
+       "      <td>3.3099</td>\n",
+       "      <td>514.0</td>\n",
+       "      <td>698</td>\n",
+       "      <td>49.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>305352</th>\n",
+       "      <td>2293</td>\n",
+       "      <td>20</td>\n",
+       "      <td>4.9800</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>351</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1712637</th>\n",
+       "      <td>1300</td>\n",
+       "      <td>2</td>\n",
+       "      <td>7.6096</td>\n",
+       "      <td>771.0</td>\n",
+       "      <td>536</td>\n",
+       "      <td>42.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         product_id  quantity_ordered  buy_price  coupon_id  customer_id  \\\n",
+       "905646         1181                 8     6.4400        NaN          692   \n",
+       "2192443        2659                 2    12.0400        NaN          675   \n",
+       "1130421        1870                 3     3.3099      514.0          698   \n",
+       "305352         2293                20     4.9800        NaN          351   \n",
+       "1712637        1300                 2     7.6096      771.0          536   \n",
+       "\n",
+       "         discount  \n",
+       "905646        0.0  \n",
+       "2192443       0.0  \n",
+       "1130421      49.0  \n",
+       "305352        0.0  \n",
+       "1712637      42.0  "
+      ]
+     },
+     "execution_count": 39,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cust_orders = pd.merge(order_details, orders[['order_id', 'customer_id']], on='order_id', how='left')\\\n",
+    "                .drop(['order_id'], axis=1)\n",
+    "cust_orders['discount'] = 100 * ((cust_orders.original_price - cust_orders.buy_price) / cust_orders.original_price)\n",
+    "cust_orders.drop(['original_price'], axis=1, inplace=True)\n",
+    "cust_orders.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 40,
+   "id": "floppy-conflict",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>cust_mean_product_price</th>\n",
+       "      <th>cust_unique_coupons_used</th>\n",
+       "      <th>cust_mean_discount</th>\n",
+       "      <th>cust_unique_products_bought</th>\n",
+       "      <th>cust_total_products_bougth</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>customer_id</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>15.47</td>\n",
+       "      <td>43.0</td>\n",
+       "      <td>10.79</td>\n",
+       "      <td>296</td>\n",
+       "      <td>2381</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>8.91</td>\n",
+       "      <td>132.0</td>\n",
+       "      <td>11.79</td>\n",
+       "      <td>851</td>\n",
+       "      <td>8470</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>10.16</td>\n",
+       "      <td>75.0</td>\n",
+       "      <td>12.59</td>\n",
+       "      <td>486</td>\n",
+       "      <td>4671</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>23.02</td>\n",
+       "      <td>18.0</td>\n",
+       "      <td>9.72</td>\n",
+       "      <td>129</td>\n",
+       "      <td>1155</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>11.81</td>\n",
+       "      <td>157.0</td>\n",
+       "      <td>11.85</td>\n",
+       "      <td>981</td>\n",
+       "      <td>11171</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>996</th>\n",
+       "      <td>9.52</td>\n",
+       "      <td>377.0</td>\n",
+       "      <td>12.60</td>\n",
+       "      <td>2194</td>\n",
+       "      <td>41630</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>997</th>\n",
+       "      <td>9.90</td>\n",
+       "      <td>106.0</td>\n",
+       "      <td>11.26</td>\n",
+       "      <td>648</td>\n",
+       "      <td>6550</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>998</th>\n",
+       "      <td>10.12</td>\n",
+       "      <td>70.0</td>\n",
+       "      <td>10.66</td>\n",
+       "      <td>478</td>\n",
+       "      <td>4511</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>999</th>\n",
+       "      <td>10.89</td>\n",
+       "      <td>208.0</td>\n",
+       "      <td>9.79</td>\n",
+       "      <td>1398</td>\n",
+       "      <td>19659</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1000</th>\n",
+       "      <td>11.09</td>\n",
+       "      <td>284.0</td>\n",
+       "      <td>12.88</td>\n",
+       "      <td>1653</td>\n",
+       "      <td>24460</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1000 rows × 5 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             cust_mean_product_price  cust_unique_coupons_used  \\\n",
+       "customer_id                                                      \n",
+       "1                              15.47                      43.0   \n",
+       "2                               8.91                     132.0   \n",
+       "3                              10.16                      75.0   \n",
+       "4                              23.02                      18.0   \n",
+       "5                              11.81                     157.0   \n",
+       "...                              ...                       ...   \n",
+       "996                             9.52                     377.0   \n",
+       "997                             9.90                     106.0   \n",
+       "998                            10.12                      70.0   \n",
+       "999                            10.89                     208.0   \n",
+       "1000                           11.09                     284.0   \n",
+       "\n",
+       "             cust_mean_discount  cust_unique_products_bought  \\\n",
+       "customer_id                                                    \n",
+       "1                         10.79                          296   \n",
+       "2                         11.79                          851   \n",
+       "3                         12.59                          486   \n",
+       "4                          9.72                          129   \n",
+       "5                         11.85                          981   \n",
+       "...                         ...                          ...   \n",
+       "996                       12.60                         2194   \n",
+       "997                       11.26                          648   \n",
+       "998                       10.66                          478   \n",
+       "999                        9.79                         1398   \n",
+       "1000                      12.88                         1653   \n",
+       "\n",
+       "             cust_total_products_bougth  \n",
+       "customer_id                              \n",
+       "1                                  2381  \n",
+       "2                                  8470  \n",
+       "3                                  4671  \n",
+       "4                                  1155  \n",
+       "5                                 11171  \n",
+       "...                                 ...  \n",
+       "996                               41630  \n",
+       "997                                6550  \n",
+       "998                                4511  \n",
+       "999                               19659  \n",
+       "1000                              24460  \n",
+       "\n",
+       "[1000 rows x 5 columns]"
+      ]
+     },
+     "execution_count": 40,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cust_stats = pd.pivot_table(cust_orders,\n",
+    "                            values=['product_id', 'quantity_ordered', 'buy_price', 'coupon_id', 'discount'],\n",
+    "                            index='customer_id',\n",
+    "                            aggfunc={\n",
+    "                                'product_id': lambda x: len(set(x)),\n",
+    "                                'quantity_ordered': sum,\n",
+    "                                'buy_price': lambda x: np.round(np.mean(x), decimals=2),\n",
+    "                                'coupon_id': lambda x: int(x.nunique()),\n",
+    "                                'discount': lambda x: np.round(np.mean(x), decimals=2)\n",
+    "                            })\n",
+    "cust_stats.rename(columns={\n",
+    "    'product_id': 'cust_unique_products_bought',\n",
+    "    'quantity_ordered': 'cust_total_products_bougth',\n",
+    "    'buy_price': 'cust_mean_product_price',\n",
+    "    'coupon_id': 'cust_unique_coupons_used',\n",
+    "    'discount': 'cust_mean_discount'\n",
+    "}, inplace=True)\n",
+    "cust_stats"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 41,
+   "id": "painted-arbor",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>cust_gender</th>\n",
+       "      <th>cust_age</th>\n",
+       "      <th>cust_credit</th>\n",
+       "      <th>cust_mean_product_price</th>\n",
+       "      <th>cust_unique_coupons_used</th>\n",
+       "      <th>cust_mean_discount</th>\n",
+       "      <th>cust_unique_products_bought</th>\n",
+       "      <th>cust_total_products_bougth</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>F</td>\n",
+       "      <td>young</td>\n",
+       "      <td>1</td>\n",
+       "      <td>15.47</td>\n",
+       "      <td>43.0</td>\n",
+       "      <td>10.79</td>\n",
+       "      <td>296</td>\n",
+       "      <td>2381</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>F</td>\n",
+       "      <td>young</td>\n",
+       "      <td>1</td>\n",
+       "      <td>8.91</td>\n",
+       "      <td>132.0</td>\n",
+       "      <td>11.79</td>\n",
+       "      <td>851</td>\n",
+       "      <td>8470</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>F</td>\n",
+       "      <td>old</td>\n",
+       "      <td>0</td>\n",
+       "      <td>10.16</td>\n",
+       "      <td>75.0</td>\n",
+       "      <td>12.59</td>\n",
+       "      <td>486</td>\n",
+       "      <td>4671</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>F</td>\n",
+       "      <td>old</td>\n",
+       "      <td>2</td>\n",
+       "      <td>23.02</td>\n",
+       "      <td>18.0</td>\n",
+       "      <td>9.72</td>\n",
+       "      <td>129</td>\n",
+       "      <td>1155</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>1</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>157.0</td>\n",
+       "      <td>11.85</td>\n",
+       "      <td>981</td>\n",
+       "      <td>11171</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>995</th>\n",
+       "      <td>996</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>0</td>\n",
+       "      <td>9.52</td>\n",
+       "      <td>377.0</td>\n",
+       "      <td>12.60</td>\n",
+       "      <td>2194</td>\n",
+       "      <td>41630</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>996</th>\n",
+       "      <td>997</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>9.90</td>\n",
+       "      <td>106.0</td>\n",
+       "      <td>11.26</td>\n",
+       "      <td>648</td>\n",
+       "      <td>6550</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>997</th>\n",
+       "      <td>998</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>1</td>\n",
+       "      <td>10.12</td>\n",
+       "      <td>70.0</td>\n",
+       "      <td>10.66</td>\n",
+       "      <td>478</td>\n",
+       "      <td>4511</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>998</th>\n",
+       "      <td>999</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.89</td>\n",
+       "      <td>208.0</td>\n",
+       "      <td>9.79</td>\n",
+       "      <td>1398</td>\n",
+       "      <td>19659</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>999</th>\n",
+       "      <td>1000</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>1</td>\n",
+       "      <td>11.09</td>\n",
+       "      <td>284.0</td>\n",
+       "      <td>12.88</td>\n",
+       "      <td>1653</td>\n",
+       "      <td>24460</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1000 rows × 9 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "     customer_id cust_gender cust_age cust_credit  cust_mean_product_price  \\\n",
+       "0              1           F    young           1                    15.47   \n",
+       "1              2           F    young           1                     8.91   \n",
+       "2              3           F      old           0                    10.16   \n",
+       "3              4           F      old           2                    23.02   \n",
+       "4              5           M      mid           1                    11.81   \n",
+       "..           ...         ...      ...         ...                      ...   \n",
+       "995          996           F      mid           0                     9.52   \n",
+       "996          997           M    young           2                     9.90   \n",
+       "997          998           M      mid           1                    10.12   \n",
+       "998          999           M      mid           2                    10.89   \n",
+       "999         1000           F      mid           1                    11.09   \n",
+       "\n",
+       "     cust_unique_coupons_used  cust_mean_discount  \\\n",
+       "0                        43.0               10.79   \n",
+       "1                       132.0               11.79   \n",
+       "2                        75.0               12.59   \n",
+       "3                        18.0                9.72   \n",
+       "4                       157.0               11.85   \n",
+       "..                        ...                 ...   \n",
+       "995                     377.0               12.60   \n",
+       "996                     106.0               11.26   \n",
+       "997                      70.0               10.66   \n",
+       "998                     208.0                9.79   \n",
+       "999                     284.0               12.88   \n",
+       "\n",
+       "     cust_unique_products_bought  cust_total_products_bougth  \n",
+       "0                            296                        2381  \n",
+       "1                            851                        8470  \n",
+       "2                            486                        4671  \n",
+       "3                            129                        1155  \n",
+       "4                            981                       11171  \n",
+       "..                           ...                         ...  \n",
+       "995                         2194                       41630  \n",
+       "996                          648                        6550  \n",
+       "997                          478                        4511  \n",
+       "998                         1398                       19659  \n",
+       "999                         1653                       24460  \n",
+       "\n",
+       "[1000 rows x 9 columns]"
+      ]
+     },
+     "execution_count": 41,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "customer_data = pd.merge(customer_demo, cust_stats, on='customer_id', how='left')\n",
+    "customer_data"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "interpreted-murder",
+   "metadata": {},
+   "source": [
+    "## Prepare coupon data\n",
+    "- From `coupons` table, take:\n",
+    "    - type\n",
+    "    - days_valid (end_date - start_date)\n",
+    "    - discount\n",
+    "    - how_many (change to `1` where `how_many == 1`)\n",
+    "- From merging `products` and `coupon_product`, `coupons`, take:\n",
+    "    - mean_product_price"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 42,
+   "id": "vocal-milton",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1033 entries, 0 to 1032\n",
+      "Data columns (total 5 columns):\n",
+      " #   Column      Non-Null Count  Dtype \n",
+      "---  ------      --------------  ----- \n",
+      " 0   coupon_id   1033 non-null   int64 \n",
+      " 1   type        1033 non-null   object\n",
+      " 2   discount    1033 non-null   int64 \n",
+      " 3   how_many    1033 non-null   int64 \n",
+      " 4   days_valid  1033 non-null   int64 \n",
+      "dtypes: int64(4), object(1)\n",
+      "memory usage: 40.5+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "coupon_info = coupons.drop(['department'], axis=1)\n",
+    "coupon_info['days_valid'] = (coupon_info.end_date - coupon_info.start_date).dt.days + 1\n",
+    "coupon_info.drop(['start_date', 'end_date'], axis=1, inplace=True)\n",
+    "coupon_info.loc[coupon_info.how_many == -1, 'how_many'] = 1\n",
+    "coupon_info.info()\n",
+    "coupon_info.rename(columns={'type': 'coupon_type', 'how_many': 'coupon_how_many', \n",
+    "                            'days_valid': 'coupon_days_valid', 'discount': 'coupon_discount'},\n",
+    "                  inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 43,
+   "id": "later-technical",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_mean_prod_price</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2.18</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>9.62</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4.25</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>8.02</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>10.27</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1029</th>\n",
+       "      <td>11.44</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1030</th>\n",
+       "      <td>18.22</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1031</th>\n",
+       "      <td>7.11</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1032</th>\n",
+       "      <td>6.41</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1033</th>\n",
+       "      <td>2.76</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1033 rows × 1 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "           coupon_mean_prod_price\n",
+       "coupon_id                        \n",
+       "1                            2.18\n",
+       "2                            9.62\n",
+       "3                            4.25\n",
+       "4                            8.02\n",
+       "5                           10.27\n",
+       "...                           ...\n",
+       "1029                        11.44\n",
+       "1030                        18.22\n",
+       "1031                         7.11\n",
+       "1032                         6.41\n",
+       "1033                         2.76\n",
+       "\n",
+       "[1033 rows x 1 columns]"
+      ]
+     },
+     "execution_count": 43,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "coupon_price = pd.merge(coupon_product, products[['product_id', 'buy_price']], on='product_id')\\\n",
+    "                 .drop(['product_id'], axis=1).drop_duplicates()\n",
+    "\n",
+    "# Coupon_product does not include coupons valid for all products in a department, add those product prices here\n",
+    "department_coupons = coupons.loc[coupons.type == 'department'][['coupon_id', 'department']]\\\n",
+    "                            .merge(products[['department', 'buy_price']], on='department', how='left')\\\n",
+    "                            .drop(['department'], axis=1)\n",
+    "\n",
+    "coupon_price = coupon_price.append(department_coupons).reset_index(drop=True)\n",
+    "\n",
+    "coupon_price = pd.pivot_table(coupon_price, index='coupon_id', values=['buy_price'],\n",
+    "                              aggfunc={\n",
+    "                                  'buy_price': lambda x: np.round(np.mean(x), decimals=2)\n",
+    "                              })\n",
+    "coupon_price.rename(columns={'buy_price': 'coupon_mean_prod_price'}, inplace=True)\n",
+    "coupon_price"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "id": "chronic-statement",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>coupon_type</th>\n",
+       "      <th>coupon_discount</th>\n",
+       "      <th>coupon_how_many</th>\n",
+       "      <th>coupon_days_valid</th>\n",
+       "      <th>coupon_mean_prod_price</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>22</td>\n",
+       "      <td>2</td>\n",
+       "      <td>7</td>\n",
+       "      <td>2.18</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>17</td>\n",
+       "      <td>1</td>\n",
+       "      <td>6</td>\n",
+       "      <td>9.62</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>30</td>\n",
+       "      <td>1</td>\n",
+       "      <td>25</td>\n",
+       "      <td>4.25</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>47</td>\n",
+       "      <td>3</td>\n",
+       "      <td>18</td>\n",
+       "      <td>8.02</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>35</td>\n",
+       "      <td>3</td>\n",
+       "      <td>1</td>\n",
+       "      <td>10.27</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1028</th>\n",
+       "      <td>1029</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>64</td>\n",
+       "      <td>5</td>\n",
+       "      <td>27</td>\n",
+       "      <td>11.44</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1029</th>\n",
+       "      <td>1030</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>14</td>\n",
+       "      <td>4</td>\n",
+       "      <td>27</td>\n",
+       "      <td>18.22</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1030</th>\n",
+       "      <td>1031</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>51</td>\n",
+       "      <td>3</td>\n",
+       "      <td>3</td>\n",
+       "      <td>7.11</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1031</th>\n",
+       "      <td>1032</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>10</td>\n",
+       "      <td>1</td>\n",
+       "      <td>22</td>\n",
+       "      <td>6.41</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1032</th>\n",
+       "      <td>1033</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>28</td>\n",
+       "      <td>4</td>\n",
+       "      <td>17</td>\n",
+       "      <td>2.76</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>1033 rows × 6 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      coupon_id    coupon_type  coupon_discount  coupon_how_many  \\\n",
+       "0             1       buy_more               22                2   \n",
+       "1             2  just_discount               17                1   \n",
+       "2             3  just_discount               30                1   \n",
+       "3             4       buy_more               47                3   \n",
+       "4             5        buy_all               35                3   \n",
+       "...         ...            ...              ...              ...   \n",
+       "1028       1029        buy_all               64                5   \n",
+       "1029       1030       buy_more               14                4   \n",
+       "1030       1031        buy_all               51                3   \n",
+       "1031       1032  just_discount               10                1   \n",
+       "1032       1033       buy_more               28                4   \n",
+       "\n",
+       "      coupon_days_valid  coupon_mean_prod_price  \n",
+       "0                     7                    2.18  \n",
+       "1                     6                    9.62  \n",
+       "2                    25                    4.25  \n",
+       "3                    18                    8.02  \n",
+       "4                     1                   10.27  \n",
+       "...                 ...                     ...  \n",
+       "1028                 27                   11.44  \n",
+       "1029                 27                   18.22  \n",
+       "1030                  3                    7.11  \n",
+       "1031                 22                    6.41  \n",
+       "1032                 17                    2.76  \n",
+       "\n",
+       "[1033 rows x 6 columns]"
+      ]
+     },
+     "execution_count": 44,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "coupon_data = pd.merge(coupon_info, coupon_price, on='coupon_id', how='left')\n",
+    "coupon_data"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "touched-complaint",
+   "metadata": {},
+   "source": [
+    "## Merge everyghing into one dataframe\n",
+    "\n",
+    "1. Add `customer_id` to `order_coupons` (from `orders`) -> `final`\n",
+    "2. Merge `final` with `customer_data`\n",
+    "3. Merge `final` with `coupon_data`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "id": "therapeutic-swift",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>coupon_used</th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>cust_gender</th>\n",
+       "      <th>cust_age</th>\n",
+       "      <th>cust_credit</th>\n",
+       "      <th>cust_mean_product_price</th>\n",
+       "      <th>cust_unique_coupons_used</th>\n",
+       "      <th>cust_mean_discount</th>\n",
+       "      <th>cust_unique_products_bought</th>\n",
+       "      <th>cust_total_products_bougth</th>\n",
+       "      <th>coupon_type</th>\n",
+       "      <th>coupon_discount</th>\n",
+       "      <th>coupon_how_many</th>\n",
+       "      <th>coupon_days_valid</th>\n",
+       "      <th>coupon_mean_prod_price</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>51</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.17</td>\n",
+       "      <td>282.0</td>\n",
+       "      <td>11.94</td>\n",
+       "      <td>1764</td>\n",
+       "      <td>29456</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>17</td>\n",
+       "      <td>1</td>\n",
+       "      <td>6</td>\n",
+       "      <td>9.62</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>3.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>51</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.17</td>\n",
+       "      <td>282.0</td>\n",
+       "      <td>11.94</td>\n",
+       "      <td>1764</td>\n",
+       "      <td>29456</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>30</td>\n",
+       "      <td>1</td>\n",
+       "      <td>25</td>\n",
+       "      <td>4.25</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>6.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>51</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.17</td>\n",
+       "      <td>282.0</td>\n",
+       "      <td>11.94</td>\n",
+       "      <td>1764</td>\n",
+       "      <td>29456</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>41</td>\n",
+       "      <td>3</td>\n",
+       "      <td>5</td>\n",
+       "      <td>12.51</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>12.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>51</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.17</td>\n",
+       "      <td>282.0</td>\n",
+       "      <td>11.94</td>\n",
+       "      <td>1764</td>\n",
+       "      <td>29456</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>5</td>\n",
+       "      <td>1</td>\n",
+       "      <td>13</td>\n",
+       "      <td>2.01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>13.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>51</td>\n",
+       "      <td>M</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.17</td>\n",
+       "      <td>282.0</td>\n",
+       "      <td>11.94</td>\n",
+       "      <td>1764</td>\n",
+       "      <td>29456</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>45</td>\n",
+       "      <td>3</td>\n",
+       "      <td>14</td>\n",
+       "      <td>3.92</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368809</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1015.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>19</td>\n",
+       "      <td>1</td>\n",
+       "      <td>25</td>\n",
+       "      <td>13.01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368810</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1019.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>6</td>\n",
+       "      <td>1</td>\n",
+       "      <td>22</td>\n",
+       "      <td>4.67</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368811</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1028.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>61</td>\n",
+       "      <td>3</td>\n",
+       "      <td>23</td>\n",
+       "      <td>9.24</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368812</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1029.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>64</td>\n",
+       "      <td>5</td>\n",
+       "      <td>27</td>\n",
+       "      <td>11.44</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368813</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1032.0</td>\n",
+       "      <td>False</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>10</td>\n",
+       "      <td>1</td>\n",
+       "      <td>22</td>\n",
+       "      <td>6.41</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>3368814 rows × 17 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  coupon_id  coupon_used  customer_id cust_gender cust_age  \\\n",
+       "0               1        2.0        False           51           M      mid   \n",
+       "1               1        3.0        False           51           M      mid   \n",
+       "2               1        6.0        False           51           M      mid   \n",
+       "3               1       12.0        False           51           M      mid   \n",
+       "4               1       13.0        False           51           M      mid   \n",
+       "...           ...        ...          ...          ...         ...      ...   \n",
+       "3368809    308313     1015.0        False          985           M    young   \n",
+       "3368810    308313     1019.0        False          985           M    young   \n",
+       "3368811    308313     1028.0        False          985           M    young   \n",
+       "3368812    308313     1029.0        False          985           M    young   \n",
+       "3368813    308313     1032.0        False          985           M    young   \n",
+       "\n",
+       "        cust_credit  cust_mean_product_price  cust_unique_coupons_used  \\\n",
+       "0                 2                    10.17                     282.0   \n",
+       "1                 2                    10.17                     282.0   \n",
+       "2                 2                    10.17                     282.0   \n",
+       "3                 2                    10.17                     282.0   \n",
+       "4                 2                    10.17                     282.0   \n",
+       "...             ...                      ...                       ...   \n",
+       "3368809           2                    12.03                      65.0   \n",
+       "3368810           2                    12.03                      65.0   \n",
+       "3368811           2                    12.03                      65.0   \n",
+       "3368812           2                    12.03                      65.0   \n",
+       "3368813           2                    12.03                      65.0   \n",
+       "\n",
+       "         cust_mean_discount  cust_unique_products_bought  \\\n",
+       "0                     11.94                         1764   \n",
+       "1                     11.94                         1764   \n",
+       "2                     11.94                         1764   \n",
+       "3                     11.94                         1764   \n",
+       "4                     11.94                         1764   \n",
+       "...                     ...                          ...   \n",
+       "3368809               11.81                          440   \n",
+       "3368810               11.81                          440   \n",
+       "3368811               11.81                          440   \n",
+       "3368812               11.81                          440   \n",
+       "3368813               11.81                          440   \n",
+       "\n",
+       "         cust_total_products_bougth    coupon_type  coupon_discount  \\\n",
+       "0                             29456  just_discount               17   \n",
+       "1                             29456  just_discount               30   \n",
+       "2                             29456        buy_all               41   \n",
+       "3                             29456  just_discount                5   \n",
+       "4                             29456        buy_all               45   \n",
+       "...                             ...            ...              ...   \n",
+       "3368809                        4420  just_discount               19   \n",
+       "3368810                        4420  just_discount                6   \n",
+       "3368811                        4420        buy_all               61   \n",
+       "3368812                        4420        buy_all               64   \n",
+       "3368813                        4420  just_discount               10   \n",
+       "\n",
+       "         coupon_how_many  coupon_days_valid  coupon_mean_prod_price  \n",
+       "0                      1                  6                    9.62  \n",
+       "1                      1                 25                    4.25  \n",
+       "2                      3                  5                   12.51  \n",
+       "3                      1                 13                    2.01  \n",
+       "4                      3                 14                    3.92  \n",
+       "...                  ...                ...                     ...  \n",
+       "3368809                1                 25                   13.01  \n",
+       "3368810                1                 22                    4.67  \n",
+       "3368811                3                 23                    9.24  \n",
+       "3368812                5                 27                   11.44  \n",
+       "3368813                1                 22                    6.41  \n",
+       "\n",
+       "[3368814 rows x 17 columns]"
+      ]
+     },
+     "execution_count": 45,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "final = pd.merge(order_coupons, orders[['order_id', 'customer_id']], on='order_id', how='left')\n",
+    "final = pd.merge(final, customer_data, on='customer_id', how='left')\n",
+    "final = pd.merge(final, coupon_data, on='coupon_id', how='left')\n",
+    "final"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 46,
+   "id": "theoretical-crack",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>coupon_used</th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>cust_gender</th>\n",
+       "      <th>cust_age</th>\n",
+       "      <th>cust_credit</th>\n",
+       "      <th>cust_mean_product_price</th>\n",
+       "      <th>cust_unique_coupons_used</th>\n",
+       "      <th>cust_mean_discount</th>\n",
+       "      <th>cust_unique_products_bought</th>\n",
+       "      <th>cust_total_products_bougth</th>\n",
+       "      <th>coupon_type</th>\n",
+       "      <th>coupon_discount</th>\n",
+       "      <th>coupon_how_many</th>\n",
+       "      <th>coupon_days_valid</th>\n",
+       "      <th>coupon_mean_prod_price</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>19</th>\n",
+       "      <td>2</td>\n",
+       "      <td>14.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>54</td>\n",
+       "      <td>F</td>\n",
+       "      <td>old</td>\n",
+       "      <td>1</td>\n",
+       "      <td>13.63</td>\n",
+       "      <td>369.0</td>\n",
+       "      <td>11.93</td>\n",
+       "      <td>1934</td>\n",
+       "      <td>42265</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>50</td>\n",
+       "      <td>2</td>\n",
+       "      <td>24</td>\n",
+       "      <td>7.06</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>21</th>\n",
+       "      <td>3</td>\n",
+       "      <td>8.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>62</td>\n",
+       "      <td>F</td>\n",
+       "      <td>old</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>510.0</td>\n",
+       "      <td>12.53</td>\n",
+       "      <td>2346</td>\n",
+       "      <td>77144</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>44</td>\n",
+       "      <td>2</td>\n",
+       "      <td>20</td>\n",
+       "      <td>3.78</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>70</th>\n",
+       "      <td>7</td>\n",
+       "      <td>2.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>167</td>\n",
+       "      <td>F</td>\n",
+       "      <td>old</td>\n",
+       "      <td>0</td>\n",
+       "      <td>10.65</td>\n",
+       "      <td>220.0</td>\n",
+       "      <td>13.90</td>\n",
+       "      <td>1218</td>\n",
+       "      <td>15034</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>17</td>\n",
+       "      <td>1</td>\n",
+       "      <td>6</td>\n",
+       "      <td>9.62</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>94</th>\n",
+       "      <td>9</td>\n",
+       "      <td>13.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>180</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.81</td>\n",
+       "      <td>305.0</td>\n",
+       "      <td>12.79</td>\n",
+       "      <td>1822</td>\n",
+       "      <td>31331</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>45</td>\n",
+       "      <td>3</td>\n",
+       "      <td>14</td>\n",
+       "      <td>3.92</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>101</th>\n",
+       "      <td>9</td>\n",
+       "      <td>7.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>180</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.81</td>\n",
+       "      <td>305.0</td>\n",
+       "      <td>12.79</td>\n",
+       "      <td>1822</td>\n",
+       "      <td>31331</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>25</td>\n",
+       "      <td>1</td>\n",
+       "      <td>4</td>\n",
+       "      <td>10.00</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368781</th>\n",
+       "      <td>308310</td>\n",
+       "      <td>1020.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>792</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>2</td>\n",
+       "      <td>10.69</td>\n",
+       "      <td>269.0</td>\n",
+       "      <td>13.64</td>\n",
+       "      <td>1692</td>\n",
+       "      <td>26774</td>\n",
+       "      <td>department</td>\n",
+       "      <td>32</td>\n",
+       "      <td>1</td>\n",
+       "      <td>30</td>\n",
+       "      <td>14.57</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368787</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>1020.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>858</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>1</td>\n",
+       "      <td>11.18</td>\n",
+       "      <td>395.0</td>\n",
+       "      <td>12.48</td>\n",
+       "      <td>2103</td>\n",
+       "      <td>43475</td>\n",
+       "      <td>department</td>\n",
+       "      <td>32</td>\n",
+       "      <td>1</td>\n",
+       "      <td>30</td>\n",
+       "      <td>14.57</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368793</th>\n",
+       "      <td>308311</td>\n",
+       "      <td>1018.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>858</td>\n",
+       "      <td>F</td>\n",
+       "      <td>mid</td>\n",
+       "      <td>1</td>\n",
+       "      <td>11.18</td>\n",
+       "      <td>395.0</td>\n",
+       "      <td>12.48</td>\n",
+       "      <td>2103</td>\n",
+       "      <td>43475</td>\n",
+       "      <td>department</td>\n",
+       "      <td>5</td>\n",
+       "      <td>1</td>\n",
+       "      <td>28</td>\n",
+       "      <td>17.83</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368804</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1020.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>department</td>\n",
+       "      <td>32</td>\n",
+       "      <td>1</td>\n",
+       "      <td>30</td>\n",
+       "      <td>14.57</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3368805</th>\n",
+       "      <td>308313</td>\n",
+       "      <td>1021.0</td>\n",
+       "      <td>True</td>\n",
+       "      <td>985</td>\n",
+       "      <td>M</td>\n",
+       "      <td>young</td>\n",
+       "      <td>2</td>\n",
+       "      <td>12.03</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>11.81</td>\n",
+       "      <td>440</td>\n",
+       "      <td>4420</td>\n",
+       "      <td>buy_all</td>\n",
+       "      <td>58</td>\n",
+       "      <td>4</td>\n",
+       "      <td>18</td>\n",
+       "      <td>4.74</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>341463 rows × 17 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  coupon_id  coupon_used  customer_id cust_gender cust_age  \\\n",
+       "19              2       14.0         True           54           F      old   \n",
+       "21              3        8.0         True           62           F      old   \n",
+       "70              7        2.0         True          167           F      old   \n",
+       "94              9       13.0         True          180           F      mid   \n",
+       "101             9        7.0         True          180           F      mid   \n",
+       "...           ...        ...          ...          ...         ...      ...   \n",
+       "3368781    308310     1020.0         True          792           F      mid   \n",
+       "3368787    308311     1020.0         True          858           F      mid   \n",
+       "3368793    308311     1018.0         True          858           F      mid   \n",
+       "3368804    308313     1020.0         True          985           M    young   \n",
+       "3368805    308313     1021.0         True          985           M    young   \n",
+       "\n",
+       "        cust_credit  cust_mean_product_price  cust_unique_coupons_used  \\\n",
+       "19                1                    13.63                     369.0   \n",
+       "21                2                    12.03                     510.0   \n",
+       "70                0                    10.65                     220.0   \n",
+       "94                2                    10.81                     305.0   \n",
+       "101               2                    10.81                     305.0   \n",
+       "...             ...                      ...                       ...   \n",
+       "3368781           2                    10.69                     269.0   \n",
+       "3368787           1                    11.18                     395.0   \n",
+       "3368793           1                    11.18                     395.0   \n",
+       "3368804           2                    12.03                      65.0   \n",
+       "3368805           2                    12.03                      65.0   \n",
+       "\n",
+       "         cust_mean_discount  cust_unique_products_bought  \\\n",
+       "19                    11.93                         1934   \n",
+       "21                    12.53                         2346   \n",
+       "70                    13.90                         1218   \n",
+       "94                    12.79                         1822   \n",
+       "101                   12.79                         1822   \n",
+       "...                     ...                          ...   \n",
+       "3368781               13.64                         1692   \n",
+       "3368787               12.48                         2103   \n",
+       "3368793               12.48                         2103   \n",
+       "3368804               11.81                          440   \n",
+       "3368805               11.81                          440   \n",
+       "\n",
+       "         cust_total_products_bougth    coupon_type  coupon_discount  \\\n",
+       "19                            42265       buy_more               50   \n",
+       "21                            77144        buy_all               44   \n",
+       "70                            15034  just_discount               17   \n",
+       "94                            31331        buy_all               45   \n",
+       "101                           31331  just_discount               25   \n",
+       "...                             ...            ...              ...   \n",
+       "3368781                       26774     department               32   \n",
+       "3368787                       43475     department               32   \n",
+       "3368793                       43475     department                5   \n",
+       "3368804                        4420     department               32   \n",
+       "3368805                        4420        buy_all               58   \n",
+       "\n",
+       "         coupon_how_many  coupon_days_valid  coupon_mean_prod_price  \n",
+       "19                     2                 24                    7.06  \n",
+       "21                     2                 20                    3.78  \n",
+       "70                     1                  6                    9.62  \n",
+       "94                     3                 14                    3.92  \n",
+       "101                    1                  4                   10.00  \n",
+       "...                  ...                ...                     ...  \n",
+       "3368781                1                 30                   14.57  \n",
+       "3368787                1                 30                   14.57  \n",
+       "3368793                1                 28                   17.83  \n",
+       "3368804                1                 30                   14.57  \n",
+       "3368805                4                 18                    4.74  \n",
+       "\n",
+       "[341463 rows x 17 columns]"
+      ]
+     },
+     "execution_count": 46,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "final.loc[final['coupon_used'] == True]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "beautiful-latest",
+   "metadata": {},
+   "source": [
+    "## Droping ids, encoding"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "decimal-guard",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train = final.drop(['order_id', 'coupon_id', 'customer_id'], axis=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "capable-brother",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train.coupon_used = train.coupon_used.astype(int)\n",
+    "train = pd.get_dummies(train, columns=['cust_gender', 'cust_age', 'coupon_type'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "pharmaceutical-trial",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train.columns"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "original-decimal",
+   "metadata": {},
+   "source": [
+    "## Save as csv"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "consistent-arrangement",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train.to_csv(os.path.join(data_dir, 'train.csv'), index=False)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/training-with-artificial-data/02_training_automl.ipynb b/training-with-artificial-data/02_training_automl.ipynb
new file mode 100644
index 0000000..232c5bc
--- /dev/null
+++ b/training-with-artificial-data/02_training_automl.ipynb
@@ -0,0 +1,2473 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "sized-waters",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "\n",
+    "import h2o\n",
+    "from h2o.automl import H2OAutoML"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "handy-terry",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "h2o.init()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 108,
+   "id": "ceramic-association",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dir = 'data_0408_0'\n",
+    "train_file_path = os.path.join(data_dir, 'train.csv')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "electronic-gibraltar",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Parse progress: |█████████████████████████████████████████████████████████| 100%\n",
+      "Rows:3368814\n",
+      "Cols:20\n",
+      "\n",
+      "\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<table>\n",
+       "<thead>\n",
+       "<tr><th>       </th><th>coupon_used        </th><th>cust_credit       </th><th>cust_mean_product_price  </th><th>cust_unique_coupons_used  </th><th>cust_mean_discount  </th><th>cust_unique_products_bought  </th><th>cust_total_products_bougth  </th><th>coupon_discount   </th><th>coupon_how_many   </th><th>coupon_days_valid  </th><th>coupon_mean_prod_price  </th><th>cust_gender_F      </th><th>cust_gender_M      </th><th>cust_age_mid      </th><th>cust_age_old      </th><th>cust_age_young     </th><th>coupon_type_buy_all  </th><th>coupon_type_buy_more  </th><th>coupon_type_department  </th><th>coupon_type_just_discount  </th></tr>\n",
+       "</thead>\n",
+       "<tbody>\n",
+       "<tr><td>type   </td><td>int                </td><td>int               </td><td>real                     </td><td>int                       </td><td>real                </td><td>int                          </td><td>int                         </td><td>int               </td><td>int               </td><td>int                </td><td>real                    </td><td>int                </td><td>int                </td><td>int               </td><td>int               </td><td>int                </td><td>int                  </td><td>int                   </td><td>int                     </td><td>int                        </td></tr>\n",
+       "<tr><td>mins   </td><td>0.0                </td><td>0.0               </td><td>6.53                     </td><td>0.0                       </td><td>0.0                 </td><td>5.0                          </td><td>44.0                        </td><td>5.0               </td><td>1.0               </td><td>1.0                </td><td>1.02                    </td><td>0.0                </td><td>0.0                </td><td>0.0               </td><td>0.0               </td><td>0.0                </td><td>0.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>mean   </td><td>0.10136000384705121</td><td>1.1269829678931513</td><td>11.23579317231524        </td><td>288.1248774791359         </td><td>12.157121176770222  </td><td>1637.607053402177            </td><td>31270.10393895302           </td><td>29.912033730565106</td><td>2.6530633629520652</td><td>20.144611723888584 </td><td>11.517169638929307      </td><td>0.6130243462536074 </td><td>0.38697565374639264</td><td>0.3835444758897345</td><td>0.4098581874808167</td><td>0.20659733662944882</td><td>0.39924644103236334  </td><td>0.253279047166154     </td><td>0.05399466993428548     </td><td>0.29347984186719717        </td></tr>\n",
+       "<tr><td>maxs   </td><td>1.0                </td><td>2.0               </td><td>23.02                    </td><td>562.0                     </td><td>17.94               </td><td>2662.0                       </td><td>102076.0                    </td><td>70.0              </td><td>5.0               </td><td>30.0               </td><td>525.05                  </td><td>1.0                </td><td>1.0                </td><td>1.0               </td><td>1.0               </td><td>1.0                </td><td>1.0                  </td><td>1.0                   </td><td>1.0                     </td><td>1.0                        </td></tr>\n",
+       "<tr><td>sigma  </td><td>0.30180487157306235</td><td>0.8008464106226431</td><td>1.2856230666254118       </td><td>120.79882235111731        </td><td>1.000653355630264   </td><td>575.9128426350499            </td><td>19794.72723022965           </td><td>16.552668866313738</td><td>1.5006576890440027</td><td>7.2175978222313555 </td><td>24.35405441156409       </td><td>0.48705807412676194</td><td>0.48705807412676194</td><td>0.4862490936635265</td><td>0.4918074068513817</td><td>0.4048640830990501 </td><td>0.48974359776536064  </td><td>0.43488944293214415   </td><td>0.22600721385690276     </td><td>0.45535643822669347        </td></tr>\n",
+       "<tr><td>zeros  </td><td>3027351            </td><td>893572            </td><td>0                        </td><td>11                        </td><td>11                  </td><td>0                            </td><td>0                           </td><td>0                 </td><td>0                 </td><td>0                  </td><td>0                       </td><td>1303649            </td><td>2065165            </td><td>2076724           </td><td>1988078           </td><td>2672826            </td><td>2023827              </td><td>2515564               </td><td>3186916                 </td><td>2380135                    </td></tr>\n",
+       "<tr><td>missing</td><td>0                  </td><td>0                 </td><td>0                        </td><td>0                         </td><td>0                   </td><td>0                            </td><td>0                           </td><td>0                 </td><td>0                 </td><td>0                  </td><td>0                       </td><td>0                  </td><td>0                  </td><td>0                 </td><td>0                 </td><td>0                  </td><td>0                    </td><td>0                     </td><td>0                       </td><td>0                          </td></tr>\n",
+       "<tr><td>0      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>17.0              </td><td>1.0               </td><td>6.0                </td><td>9.62                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>0.0                  </td><td>0.0                   </td><td>0.0                     </td><td>1.0                        </td></tr>\n",
+       "<tr><td>1      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>30.0              </td><td>1.0               </td><td>25.0               </td><td>4.25                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>0.0                  </td><td>0.0                   </td><td>0.0                     </td><td>1.0                        </td></tr>\n",
+       "<tr><td>2      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>41.0              </td><td>3.0               </td><td>5.0                </td><td>12.51                   </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>1.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>3      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>5.0               </td><td>1.0               </td><td>13.0               </td><td>2.01                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>0.0                  </td><td>0.0                   </td><td>0.0                     </td><td>1.0                        </td></tr>\n",
+       "<tr><td>4      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>45.0              </td><td>3.0               </td><td>14.0               </td><td>3.92                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>1.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>5      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>35.0              </td><td>3.0               </td><td>1.0                </td><td>10.27                   </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>1.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>6      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>22.0              </td><td>2.0               </td><td>7.0                </td><td>2.18                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>0.0                  </td><td>1.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>7      </td><td>0.0                </td><td>2.0               </td><td>10.17                    </td><td>282.0                     </td><td>11.94               </td><td>1764.0                       </td><td>29456.0                     </td><td>44.0              </td><td>2.0               </td><td>20.0               </td><td>3.78                    </td><td>0.0                </td><td>1.0                </td><td>1.0               </td><td>0.0               </td><td>0.0                </td><td>1.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>8      </td><td>0.0                </td><td>1.0               </td><td>13.63                    </td><td>369.0                     </td><td>11.93               </td><td>1934.0                       </td><td>42265.0                     </td><td>22.0              </td><td>2.0               </td><td>7.0                </td><td>2.18                    </td><td>1.0                </td><td>0.0                </td><td>0.0               </td><td>1.0               </td><td>0.0                </td><td>0.0                  </td><td>1.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "<tr><td>9      </td><td>0.0                </td><td>1.0               </td><td>13.63                    </td><td>369.0                     </td><td>11.93               </td><td>1934.0                       </td><td>42265.0                     </td><td>44.0              </td><td>2.0               </td><td>20.0               </td><td>3.78                    </td><td>1.0                </td><td>0.0                </td><td>0.0               </td><td>1.0               </td><td>0.0                </td><td>1.0                  </td><td>0.0                   </td><td>0.0                     </td><td>0.0                        </td></tr>\n",
+       "</tbody>\n",
+       "</table>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "train_h2o = h2o.H2OFrame(train)\n",
+    "train_h2o.describe()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "6b3f3df2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y = 'coupon_used'\n",
+    "x = train_h2o.columns\n",
+    "x.remove(y)\n",
+    "train_h2o['coupon_used'] = train_h2o['coupon_used'].asfactor()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "43852a5f",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "AutoML progress: |████████████████████████████████████████████████████████| 100%\n"
+     ]
+    }
+   ],
+   "source": [
+    "aml = H2OAutoML(max_models=10, seed=1, balance_classes=True)\n",
+    "aml.train(x=x, y=y, training_frame=train_h2o)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "716251d3",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<table>\n",
+       "<thead>\n",
+       "<tr><th>model_id                                           </th><th style=\"text-align: right;\">     auc</th><th style=\"text-align: right;\">  logloss</th><th style=\"text-align: right;\">   aucpr</th><th style=\"text-align: right;\">  mean_per_class_error</th><th style=\"text-align: right;\">    rmse</th><th style=\"text-align: right;\">      mse</th></tr>\n",
+       "</thead>\n",
+       "<tbody>\n",
+       "<tr><td>StackedEnsemble_AllModels_AutoML_20210408_164243   </td><td style=\"text-align: right;\">0.803702</td><td style=\"text-align: right;\"> 0.241951</td><td style=\"text-align: right;\">0.502149</td><td style=\"text-align: right;\">              0.310092</td><td style=\"text-align: right;\">0.254671</td><td style=\"text-align: right;\">0.0648575</td></tr>\n",
+       "<tr><td>StackedEnsemble_BestOfFamily_AutoML_20210408_164243</td><td style=\"text-align: right;\">0.802811</td><td style=\"text-align: right;\"> 0.242233</td><td style=\"text-align: right;\">0.501732</td><td style=\"text-align: right;\">              0.31046 </td><td style=\"text-align: right;\">0.254723</td><td style=\"text-align: right;\">0.0648838</td></tr>\n",
+       "<tr><td>GBM_4_AutoML_20210408_164243                       </td><td style=\"text-align: right;\">0.802158</td><td style=\"text-align: right;\"> 0.242606</td><td style=\"text-align: right;\">0.499533</td><td style=\"text-align: right;\">              0.310809</td><td style=\"text-align: right;\">0.254927</td><td style=\"text-align: right;\">0.064988 </td></tr>\n",
+       "<tr><td>XGBoost_3_AutoML_20210408_164243                   </td><td style=\"text-align: right;\">0.800627</td><td style=\"text-align: right;\"> 0.242678</td><td style=\"text-align: right;\">0.501184</td><td style=\"text-align: right;\">              0.310592</td><td style=\"text-align: right;\">0.254758</td><td style=\"text-align: right;\">0.0649015</td></tr>\n",
+       "<tr><td>XGBoost_1_AutoML_20210408_164243                   </td><td style=\"text-align: right;\">0.799724</td><td style=\"text-align: right;\"> 0.243167</td><td style=\"text-align: right;\">0.497031</td><td style=\"text-align: right;\">              0.311258</td><td style=\"text-align: right;\">0.255159</td><td style=\"text-align: right;\">0.0651059</td></tr>\n",
+       "<tr><td>GBM_5_AutoML_20210408_164243                       </td><td style=\"text-align: right;\">0.798054</td><td style=\"text-align: right;\"> 0.243914</td><td style=\"text-align: right;\">0.493746</td><td style=\"text-align: right;\">              0.311124</td><td style=\"text-align: right;\">0.255478</td><td style=\"text-align: right;\">0.0652692</td></tr>\n",
+       "<tr><td>GBM_3_AutoML_20210408_164243                       </td><td style=\"text-align: right;\">0.79799 </td><td style=\"text-align: right;\"> 0.243694</td><td style=\"text-align: right;\">0.497595</td><td style=\"text-align: right;\">              0.310238</td><td style=\"text-align: right;\">0.255119</td><td style=\"text-align: right;\">0.0650856</td></tr>\n",
+       "<tr><td>GBM_2_AutoML_20210408_164243                       </td><td style=\"text-align: right;\">0.794188</td><td style=\"text-align: right;\"> 0.244514</td><td style=\"text-align: right;\">0.495236</td><td style=\"text-align: right;\">              0.310557</td><td style=\"text-align: right;\">0.255312</td><td style=\"text-align: right;\">0.0651843</td></tr>\n",
+       "<tr><td>GBM_1_AutoML_20210408_164243                       </td><td style=\"text-align: right;\">0.790586</td><td style=\"text-align: right;\"> 0.245279</td><td style=\"text-align: right;\">0.4932  </td><td style=\"text-align: right;\">              0.311157</td><td style=\"text-align: right;\">0.255517</td><td style=\"text-align: right;\">0.0652888</td></tr>\n",
+       "<tr><td>XGBoost_2_AutoML_20210408_164243                   </td><td style=\"text-align: right;\">0.785739</td><td style=\"text-align: right;\"> 0.250421</td><td style=\"text-align: right;\">0.479367</td><td style=\"text-align: right;\">              0.316064</td><td style=\"text-align: right;\">0.258171</td><td style=\"text-align: right;\">0.066652 </td></tr>\n",
+       "</tbody>\n",
+       "</table>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": []
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "aml.leaderboard"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "71abc828",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Model Details\n",
+      "=============\n",
+      "H2OGradientBoostingEstimator :  Gradient Boosting Machine\n",
+      "Model Key:  GBM_4_AutoML_20210408_164243\n",
+      "\n",
+      "\n",
+      "Model Summary: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th>number_of_trees</th>\n",
+       "      <th>number_of_internal_trees</th>\n",
+       "      <th>model_size_in_bytes</th>\n",
+       "      <th>min_depth</th>\n",
+       "      <th>max_depth</th>\n",
+       "      <th>mean_depth</th>\n",
+       "      <th>min_leaves</th>\n",
+       "      <th>max_leaves</th>\n",
+       "      <th>mean_leaves</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td></td>\n",
+       "      <td>70.0</td>\n",
+       "      <td>70.0</td>\n",
+       "      <td>578602.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>10.0</td>\n",
+       "      <td>7.571429</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>994.0</td>\n",
+       "      <td>654.6</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "     number_of_trees  number_of_internal_trees  model_size_in_bytes  \\\n",
+       "0               70.0                      70.0             578602.0   \n",
+       "\n",
+       "   min_depth  max_depth  mean_depth  min_leaves  max_leaves  mean_leaves  \n",
+       "0        0.0       10.0    7.571429         1.0       994.0        654.6  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "ModelMetricsBinomial: gbm\n",
+      "** Reported on train data. **\n",
+      "\n",
+      "MSE: 0.2711710163001577\n",
+      "RMSE: 0.5207408341009544\n",
+      "LogLoss: 0.8622691624238035\n",
+      "Mean Per-Class Error: 0.2690251558453727\n",
+      "AUC: 0.8156996013643698\n",
+      "AUCPR: 0.8394483144715442\n",
+      "Gini: 0.6313992027287396\n",
+      "\n",
+      "Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.05733060682487723: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th>0</th>\n",
+       "      <th>1</th>\n",
+       "      <th>Error</th>\n",
+       "      <th>Rate</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0</td>\n",
+       "      <td>1729723.0</td>\n",
+       "      <td>1297628.0</td>\n",
+       "      <td>0.4286</td>\n",
+       "      <td>(1297628.0/3027351.0)</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>513782.0</td>\n",
+       "      <td>2513447.0</td>\n",
+       "      <td>0.1697</td>\n",
+       "      <td>(513782.0/3027229.0)</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Total</td>\n",
+       "      <td>2243505.0</td>\n",
+       "      <td>3811075.0</td>\n",
+       "      <td>0.2992</td>\n",
+       "      <td>(1811410.0/6054580.0)</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                  0          1   Error                    Rate\n",
+       "0      0  1729723.0  1297628.0  0.4286   (1297628.0/3027351.0)\n",
+       "1      1   513782.0  2513447.0  0.1697    (513782.0/3027229.0)\n",
+       "2  Total  2243505.0  3811075.0  0.2992   (1811410.0/6054580.0)"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Maximum Metrics: Maximum metrics at their respective thresholds\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>metric</th>\n",
+       "      <th>threshold</th>\n",
+       "      <th>value</th>\n",
+       "      <th>idx</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>max f1</td>\n",
+       "      <td>0.057331</td>\n",
+       "      <td>7.351083e-01</td>\n",
+       "      <td>336.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>max f2</td>\n",
+       "      <td>0.032548</td>\n",
+       "      <td>8.452362e-01</td>\n",
+       "      <td>374.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>max f0point5</td>\n",
+       "      <td>0.133512</td>\n",
+       "      <td>7.645287e-01</td>\n",
+       "      <td>255.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>max accuracy</td>\n",
+       "      <td>0.095349</td>\n",
+       "      <td>7.309765e-01</td>\n",
+       "      <td>292.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>max precision</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>max recall</td>\n",
+       "      <td>0.007737</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>max specificity</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>max absolute_mcc</td>\n",
+       "      <td>0.139263</td>\n",
+       "      <td>4.796984e-01</td>\n",
+       "      <td>250.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>max min_per_class_accuracy</td>\n",
+       "      <td>0.078557</td>\n",
+       "      <td>7.252985e-01</td>\n",
+       "      <td>310.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>max mean_per_class_accuracy</td>\n",
+       "      <td>0.095349</td>\n",
+       "      <td>7.309748e-01</td>\n",
+       "      <td>292.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>max tns</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>3.027351e+06</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>max fns</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>3.027017e+06</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>max fps</td>\n",
+       "      <td>0.006237</td>\n",
+       "      <td>3.027351e+06</td>\n",
+       "      <td>399.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>max tps</td>\n",
+       "      <td>0.007737</td>\n",
+       "      <td>3.027229e+06</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>max tnr</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>max fnr</td>\n",
+       "      <td>0.880308</td>\n",
+       "      <td>9.999300e-01</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>16</th>\n",
+       "      <td>max fpr</td>\n",
+       "      <td>0.006237</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>399.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>17</th>\n",
+       "      <td>max tpr</td>\n",
+       "      <td>0.007737</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                         metric  threshold         value    idx\n",
+       "0                        max f1   0.057331  7.351083e-01  336.0\n",
+       "1                        max f2   0.032548  8.452362e-01  374.0\n",
+       "2                  max f0point5   0.133512  7.645287e-01  255.0\n",
+       "3                  max accuracy   0.095349  7.309765e-01  292.0\n",
+       "4                 max precision   0.880308  1.000000e+00    0.0\n",
+       "5                    max recall   0.007737  1.000000e+00  398.0\n",
+       "6               max specificity   0.880308  1.000000e+00    0.0\n",
+       "7              max absolute_mcc   0.139263  4.796984e-01  250.0\n",
+       "8    max min_per_class_accuracy   0.078557  7.252985e-01  310.0\n",
+       "9   max mean_per_class_accuracy   0.095349  7.309748e-01  292.0\n",
+       "10                      max tns   0.880308  3.027351e+06    0.0\n",
+       "11                      max fns   0.880308  3.027017e+06    0.0\n",
+       "12                      max fps   0.006237  3.027351e+06  399.0\n",
+       "13                      max tps   0.007737  3.027229e+06  398.0\n",
+       "14                      max tnr   0.880308  1.000000e+00    0.0\n",
+       "15                      max fnr   0.880308  9.999300e-01    0.0\n",
+       "16                      max fpr   0.006237  1.000000e+00  399.0\n",
+       "17                      max tpr   0.007737  1.000000e+00  398.0"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Gains/Lift Table: Avg response rate: 50.00 %, avg score: 21.13 %\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>group</th>\n",
+       "      <th>cumulative_data_fraction</th>\n",
+       "      <th>lower_threshold</th>\n",
+       "      <th>lift</th>\n",
+       "      <th>cumulative_lift</th>\n",
+       "      <th>response_rate</th>\n",
+       "      <th>score</th>\n",
+       "      <th>cumulative_response_rate</th>\n",
+       "      <th>cumulative_score</th>\n",
+       "      <th>capture_rate</th>\n",
+       "      <th>cumulative_capture_rate</th>\n",
+       "      <th>gain</th>\n",
+       "      <th>cumulative_gain</th>\n",
+       "      <th>kolmogorov_smirnov</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>0.010002</td>\n",
+       "      <td>0.837540</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>0.990753</td>\n",
+       "      <td>0.846832</td>\n",
+       "      <td>0.990753</td>\n",
+       "      <td>0.846832</td>\n",
+       "      <td>0.019819</td>\n",
+       "      <td>0.019819</td>\n",
+       "      <td>98.154496</td>\n",
+       "      <td>98.154496</td>\n",
+       "      <td>0.019634</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>0.020009</td>\n",
+       "      <td>0.828769</td>\n",
+       "      <td>1.975119</td>\n",
+       "      <td>1.978331</td>\n",
+       "      <td>0.987540</td>\n",
+       "      <td>0.832717</td>\n",
+       "      <td>0.989146</td>\n",
+       "      <td>0.839772</td>\n",
+       "      <td>0.019766</td>\n",
+       "      <td>0.039585</td>\n",
+       "      <td>97.511902</td>\n",
+       "      <td>97.833106</td>\n",
+       "      <td>0.039151</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>0.030006</td>\n",
+       "      <td>0.822702</td>\n",
+       "      <td>1.969374</td>\n",
+       "      <td>1.975347</td>\n",
+       "      <td>0.984667</td>\n",
+       "      <td>0.825749</td>\n",
+       "      <td>0.987654</td>\n",
+       "      <td>0.835101</td>\n",
+       "      <td>0.019686</td>\n",
+       "      <td>0.059272</td>\n",
+       "      <td>96.937366</td>\n",
+       "      <td>97.534695</td>\n",
+       "      <td>0.058531</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>0.040002</td>\n",
+       "      <td>0.817128</td>\n",
+       "      <td>1.967028</td>\n",
+       "      <td>1.973268</td>\n",
+       "      <td>0.983494</td>\n",
+       "      <td>0.819833</td>\n",
+       "      <td>0.986614</td>\n",
+       "      <td>0.831285</td>\n",
+       "      <td>0.019663</td>\n",
+       "      <td>0.078935</td>\n",
+       "      <td>96.702794</td>\n",
+       "      <td>97.326806</td>\n",
+       "      <td>0.077864</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>0.050000</td>\n",
+       "      <td>0.812258</td>\n",
+       "      <td>1.963398</td>\n",
+       "      <td>1.971295</td>\n",
+       "      <td>0.981679</td>\n",
+       "      <td>0.814630</td>\n",
+       "      <td>0.985627</td>\n",
+       "      <td>0.827955</td>\n",
+       "      <td>0.019630</td>\n",
+       "      <td>0.098565</td>\n",
+       "      <td>96.339839</td>\n",
+       "      <td>97.129454</td>\n",
+       "      <td>0.097127</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>6</td>\n",
+       "      <td>0.100008</td>\n",
+       "      <td>0.785588</td>\n",
+       "      <td>1.956363</td>\n",
+       "      <td>1.963828</td>\n",
+       "      <td>0.978162</td>\n",
+       "      <td>0.799591</td>\n",
+       "      <td>0.981894</td>\n",
+       "      <td>0.813772</td>\n",
+       "      <td>0.097833</td>\n",
+       "      <td>0.196398</td>\n",
+       "      <td>95.636342</td>\n",
+       "      <td>96.382841</td>\n",
+       "      <td>0.192776</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>7</td>\n",
+       "      <td>0.150011</td>\n",
+       "      <td>0.751169</td>\n",
+       "      <td>1.942474</td>\n",
+       "      <td>1.956710</td>\n",
+       "      <td>0.971217</td>\n",
+       "      <td>0.769543</td>\n",
+       "      <td>0.978335</td>\n",
+       "      <td>0.799029</td>\n",
+       "      <td>0.097131</td>\n",
+       "      <td>0.293529</td>\n",
+       "      <td>94.247368</td>\n",
+       "      <td>95.671019</td>\n",
+       "      <td>0.287029</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>8</td>\n",
+       "      <td>0.200001</td>\n",
+       "      <td>0.346516</td>\n",
+       "      <td>1.862231</td>\n",
+       "      <td>1.933095</td>\n",
+       "      <td>0.931097</td>\n",
+       "      <td>0.641898</td>\n",
+       "      <td>0.966528</td>\n",
+       "      <td>0.759754</td>\n",
+       "      <td>0.093093</td>\n",
+       "      <td>0.386622</td>\n",
+       "      <td>86.223067</td>\n",
+       "      <td>93.309509</td>\n",
+       "      <td>0.373233</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>9</td>\n",
+       "      <td>0.300003</td>\n",
+       "      <td>0.126728</td>\n",
+       "      <td>1.329560</td>\n",
+       "      <td>1.731916</td>\n",
+       "      <td>0.664767</td>\n",
+       "      <td>0.164262</td>\n",
+       "      <td>0.865941</td>\n",
+       "      <td>0.561256</td>\n",
+       "      <td>0.132958</td>\n",
+       "      <td>0.519579</td>\n",
+       "      <td>32.956029</td>\n",
+       "      <td>73.191627</td>\n",
+       "      <td>0.439145</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>10</td>\n",
+       "      <td>0.400000</td>\n",
+       "      <td>0.098188</td>\n",
+       "      <td>1.110646</td>\n",
+       "      <td>1.576603</td>\n",
+       "      <td>0.555312</td>\n",
+       "      <td>0.111048</td>\n",
+       "      <td>0.788285</td>\n",
+       "      <td>0.448707</td>\n",
+       "      <td>0.111062</td>\n",
+       "      <td>0.630641</td>\n",
+       "      <td>11.064569</td>\n",
+       "      <td>57.660273</td>\n",
+       "      <td>0.461273</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>11</td>\n",
+       "      <td>0.500001</td>\n",
+       "      <td>0.078522</td>\n",
+       "      <td>0.959023</td>\n",
+       "      <td>1.453085</td>\n",
+       "      <td>0.479502</td>\n",
+       "      <td>0.088037</td>\n",
+       "      <td>0.726528</td>\n",
+       "      <td>0.376572</td>\n",
+       "      <td>0.095904</td>\n",
+       "      <td>0.726545</td>\n",
+       "      <td>-4.097720</td>\n",
+       "      <td>45.308544</td>\n",
+       "      <td>0.453078</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>12</td>\n",
+       "      <td>0.600000</td>\n",
+       "      <td>0.061271</td>\n",
+       "      <td>0.817325</td>\n",
+       "      <td>1.347127</td>\n",
+       "      <td>0.408654</td>\n",
+       "      <td>0.069658</td>\n",
+       "      <td>0.673550</td>\n",
+       "      <td>0.325421</td>\n",
+       "      <td>0.081732</td>\n",
+       "      <td>0.808276</td>\n",
+       "      <td>-18.267544</td>\n",
+       "      <td>34.712655</td>\n",
+       "      <td>0.416544</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>13</td>\n",
+       "      <td>0.700001</td>\n",
+       "      <td>0.049013</td>\n",
+       "      <td>0.705426</td>\n",
+       "      <td>1.255455</td>\n",
+       "      <td>0.352706</td>\n",
+       "      <td>0.054639</td>\n",
+       "      <td>0.627715</td>\n",
+       "      <td>0.286737</td>\n",
+       "      <td>0.070543</td>\n",
+       "      <td>0.878819</td>\n",
+       "      <td>-29.457405</td>\n",
+       "      <td>25.545453</td>\n",
+       "      <td>0.357630</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>14</td>\n",
+       "      <td>0.800000</td>\n",
+       "      <td>0.040772</td>\n",
+       "      <td>0.577073</td>\n",
+       "      <td>1.170657</td>\n",
+       "      <td>0.288531</td>\n",
+       "      <td>0.044764</td>\n",
+       "      <td>0.585317</td>\n",
+       "      <td>0.256491</td>\n",
+       "      <td>0.057707</td>\n",
+       "      <td>0.936526</td>\n",
+       "      <td>-42.292719</td>\n",
+       "      <td>17.065740</td>\n",
+       "      <td>0.273046</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>15</td>\n",
+       "      <td>0.900000</td>\n",
+       "      <td>0.032451</td>\n",
+       "      <td>0.439379</td>\n",
+       "      <td>1.089404</td>\n",
+       "      <td>0.219685</td>\n",
+       "      <td>0.036779</td>\n",
+       "      <td>0.544691</td>\n",
+       "      <td>0.232078</td>\n",
+       "      <td>0.043938</td>\n",
+       "      <td>0.980464</td>\n",
+       "      <td>-56.062056</td>\n",
+       "      <td>8.940442</td>\n",
+       "      <td>0.160925</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>16</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.002556</td>\n",
+       "      <td>0.195360</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.097678</td>\n",
+       "      <td>0.024191</td>\n",
+       "      <td>0.499990</td>\n",
+       "      <td>0.211290</td>\n",
+       "      <td>0.019536</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>-80.463982</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "    group  cumulative_data_fraction  lower_threshold      lift  \\\n",
+       "0       1                  0.010002         0.837540  1.981545   \n",
+       "1       2                  0.020009         0.828769  1.975119   \n",
+       "2       3                  0.030006         0.822702  1.969374   \n",
+       "3       4                  0.040002         0.817128  1.967028   \n",
+       "4       5                  0.050000         0.812258  1.963398   \n",
+       "5       6                  0.100008         0.785588  1.956363   \n",
+       "6       7                  0.150011         0.751169  1.942474   \n",
+       "7       8                  0.200001         0.346516  1.862231   \n",
+       "8       9                  0.300003         0.126728  1.329560   \n",
+       "9      10                  0.400000         0.098188  1.110646   \n",
+       "10     11                  0.500001         0.078522  0.959023   \n",
+       "11     12                  0.600000         0.061271  0.817325   \n",
+       "12     13                  0.700001         0.049013  0.705426   \n",
+       "13     14                  0.800000         0.040772  0.577073   \n",
+       "14     15                  0.900000         0.032451  0.439379   \n",
+       "15     16                  1.000000         0.002556  0.195360   \n",
+       "\n",
+       "    cumulative_lift  response_rate     score  cumulative_response_rate  \\\n",
+       "0          1.981545       0.990753  0.846832                  0.990753   \n",
+       "1          1.978331       0.987540  0.832717                  0.989146   \n",
+       "2          1.975347       0.984667  0.825749                  0.987654   \n",
+       "3          1.973268       0.983494  0.819833                  0.986614   \n",
+       "4          1.971295       0.981679  0.814630                  0.985627   \n",
+       "5          1.963828       0.978162  0.799591                  0.981894   \n",
+       "6          1.956710       0.971217  0.769543                  0.978335   \n",
+       "7          1.933095       0.931097  0.641898                  0.966528   \n",
+       "8          1.731916       0.664767  0.164262                  0.865941   \n",
+       "9          1.576603       0.555312  0.111048                  0.788285   \n",
+       "10         1.453085       0.479502  0.088037                  0.726528   \n",
+       "11         1.347127       0.408654  0.069658                  0.673550   \n",
+       "12         1.255455       0.352706  0.054639                  0.627715   \n",
+       "13         1.170657       0.288531  0.044764                  0.585317   \n",
+       "14         1.089404       0.219685  0.036779                  0.544691   \n",
+       "15         1.000000       0.097678  0.024191                  0.499990   \n",
+       "\n",
+       "    cumulative_score  capture_rate  cumulative_capture_rate       gain  \\\n",
+       "0           0.846832      0.019819                 0.019819  98.154496   \n",
+       "1           0.839772      0.019766                 0.039585  97.511902   \n",
+       "2           0.835101      0.019686                 0.059272  96.937366   \n",
+       "3           0.831285      0.019663                 0.078935  96.702794   \n",
+       "4           0.827955      0.019630                 0.098565  96.339839   \n",
+       "5           0.813772      0.097833                 0.196398  95.636342   \n",
+       "6           0.799029      0.097131                 0.293529  94.247368   \n",
+       "7           0.759754      0.093093                 0.386622  86.223067   \n",
+       "8           0.561256      0.132958                 0.519579  32.956029   \n",
+       "9           0.448707      0.111062                 0.630641  11.064569   \n",
+       "10          0.376572      0.095904                 0.726545  -4.097720   \n",
+       "11          0.325421      0.081732                 0.808276 -18.267544   \n",
+       "12          0.286737      0.070543                 0.878819 -29.457405   \n",
+       "13          0.256491      0.057707                 0.936526 -42.292719   \n",
+       "14          0.232078      0.043938                 0.980464 -56.062056   \n",
+       "15          0.211290      0.019536                 1.000000 -80.463982   \n",
+       "\n",
+       "    cumulative_gain  kolmogorov_smirnov  \n",
+       "0         98.154496            0.019634  \n",
+       "1         97.833106            0.039151  \n",
+       "2         97.534695            0.058531  \n",
+       "3         97.326806            0.077864  \n",
+       "4         97.129454            0.097127  \n",
+       "5         96.382841            0.192776  \n",
+       "6         95.671019            0.287029  \n",
+       "7         93.309509            0.373233  \n",
+       "8         73.191627            0.439145  \n",
+       "9         57.660273            0.461273  \n",
+       "10        45.308544            0.453078  \n",
+       "11        34.712655            0.416544  \n",
+       "12        25.545453            0.357630  \n",
+       "13        17.065740            0.273046  \n",
+       "14         8.940442            0.160925  \n",
+       "15         0.000000            0.000000  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "ModelMetricsBinomial: gbm\n",
+      "** Reported on cross-validation data. **\n",
+      "\n",
+      "MSE: 0.06498795354987083\n",
+      "RMSE: 0.2549273495525163\n",
+      "LogLoss: 0.24260618450943383\n",
+      "Mean Per-Class Error: 0.2770904651860334\n",
+      "AUC: 0.8021582436506262\n",
+      "AUCPR: 0.4995332242622148\n",
+      "Gini: 0.6043164873012523\n",
+      "\n",
+      "Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.22919100477943258: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th>0</th>\n",
+       "      <th>1</th>\n",
+       "      <th>Error</th>\n",
+       "      <th>Rate</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0</td>\n",
+       "      <td>2977495.0</td>\n",
+       "      <td>49856.0</td>\n",
+       "      <td>0.0165</td>\n",
+       "      <td>(49856.0/3027351.0)</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>206636.0</td>\n",
+       "      <td>134827.0</td>\n",
+       "      <td>0.6051</td>\n",
+       "      <td>(206636.0/341463.0)</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Total</td>\n",
+       "      <td>3184131.0</td>\n",
+       "      <td>184683.0</td>\n",
+       "      <td>0.0761</td>\n",
+       "      <td>(256492.0/3368814.0)</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                  0         1   Error                   Rate\n",
+       "0      0  2977495.0   49856.0  0.0165    (49856.0/3027351.0)\n",
+       "1      1   206636.0  134827.0  0.6051    (206636.0/341463.0)\n",
+       "2  Total  3184131.0  184683.0  0.0761   (256492.0/3368814.0)"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Maximum Metrics: Maximum metrics at their respective thresholds\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>metric</th>\n",
+       "      <th>threshold</th>\n",
+       "      <th>value</th>\n",
+       "      <th>idx</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>max f1</td>\n",
+       "      <td>0.229191</td>\n",
+       "      <td>5.125079e-01</td>\n",
+       "      <td>167.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>max f2</td>\n",
+       "      <td>0.098798</td>\n",
+       "      <td>5.028615e-01</td>\n",
+       "      <td>259.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>max f0point5</td>\n",
+       "      <td>0.541164</td>\n",
+       "      <td>6.446633e-01</td>\n",
+       "      <td>104.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>max accuracy</td>\n",
+       "      <td>0.533739</td>\n",
+       "      <td>9.262245e-01</td>\n",
+       "      <td>105.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>max precision</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>9.166667e-01</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>max recall</td>\n",
+       "      <td>0.007146</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>max specificity</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>9.999993e-01</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>max absolute_mcc</td>\n",
+       "      <td>0.477452</td>\n",
+       "      <td>5.089233e-01</td>\n",
+       "      <td>115.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>max min_per_class_accuracy</td>\n",
+       "      <td>0.077382</td>\n",
+       "      <td>7.161830e-01</td>\n",
+       "      <td>287.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>max mean_per_class_accuracy</td>\n",
+       "      <td>0.098798</td>\n",
+       "      <td>7.229095e-01</td>\n",
+       "      <td>259.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>max tns</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>3.027349e+06</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>max fns</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>3.414410e+05</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>max fps</td>\n",
+       "      <td>0.005775</td>\n",
+       "      <td>3.027351e+06</td>\n",
+       "      <td>399.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>max tps</td>\n",
+       "      <td>0.007146</td>\n",
+       "      <td>3.414630e+05</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>max tnr</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>9.999993e-01</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>max fnr</td>\n",
+       "      <td>0.883700</td>\n",
+       "      <td>9.999356e-01</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>16</th>\n",
+       "      <td>max fpr</td>\n",
+       "      <td>0.005775</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>399.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>17</th>\n",
+       "      <td>max tpr</td>\n",
+       "      <td>0.007146</td>\n",
+       "      <td>1.000000e+00</td>\n",
+       "      <td>398.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                         metric  threshold         value    idx\n",
+       "0                        max f1   0.229191  5.125079e-01  167.0\n",
+       "1                        max f2   0.098798  5.028615e-01  259.0\n",
+       "2                  max f0point5   0.541164  6.446633e-01  104.0\n",
+       "3                  max accuracy   0.533739  9.262245e-01  105.0\n",
+       "4                 max precision   0.883700  9.166667e-01    0.0\n",
+       "5                    max recall   0.007146  1.000000e+00  398.0\n",
+       "6               max specificity   0.883700  9.999993e-01    0.0\n",
+       "7              max absolute_mcc   0.477452  5.089233e-01  115.0\n",
+       "8    max min_per_class_accuracy   0.077382  7.161830e-01  287.0\n",
+       "9   max mean_per_class_accuracy   0.098798  7.229095e-01  259.0\n",
+       "10                      max tns   0.883700  3.027349e+06    0.0\n",
+       "11                      max fns   0.883700  3.414410e+05    0.0\n",
+       "12                      max fps   0.005775  3.027351e+06  399.0\n",
+       "13                      max tps   0.007146  3.414630e+05  398.0\n",
+       "14                      max tnr   0.883700  9.999993e-01    0.0\n",
+       "15                      max fnr   0.883700  9.999356e-01    0.0\n",
+       "16                      max fpr   0.005775  1.000000e+00  399.0\n",
+       "17                      max tpr   0.007146  1.000000e+00  398.0"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Gains/Lift Table: Avg response rate: 10.14 %, avg score:  9.87 %\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>group</th>\n",
+       "      <th>cumulative_data_fraction</th>\n",
+       "      <th>lower_threshold</th>\n",
+       "      <th>lift</th>\n",
+       "      <th>cumulative_lift</th>\n",
+       "      <th>response_rate</th>\n",
+       "      <th>score</th>\n",
+       "      <th>cumulative_response_rate</th>\n",
+       "      <th>cumulative_score</th>\n",
+       "      <th>capture_rate</th>\n",
+       "      <th>cumulative_capture_rate</th>\n",
+       "      <th>gain</th>\n",
+       "      <th>cumulative_gain</th>\n",
+       "      <th>kolmogorov_smirnov</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>0.010001</td>\n",
+       "      <td>0.816049</td>\n",
+       "      <td>8.421531</td>\n",
+       "      <td>8.421531</td>\n",
+       "      <td>0.853606</td>\n",
+       "      <td>0.830672</td>\n",
+       "      <td>0.853606</td>\n",
+       "      <td>0.830672</td>\n",
+       "      <td>0.084220</td>\n",
+       "      <td>0.084220</td>\n",
+       "      <td>742.153097</td>\n",
+       "      <td>742.153097</td>\n",
+       "      <td>0.082591</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>0.020000</td>\n",
+       "      <td>0.794009</td>\n",
+       "      <td>8.171586</td>\n",
+       "      <td>8.296564</td>\n",
+       "      <td>0.828272</td>\n",
+       "      <td>0.805167</td>\n",
+       "      <td>0.840940</td>\n",
+       "      <td>0.817920</td>\n",
+       "      <td>0.081713</td>\n",
+       "      <td>0.165933</td>\n",
+       "      <td>717.158645</td>\n",
+       "      <td>729.656427</td>\n",
+       "      <td>0.162393</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>0.030000</td>\n",
+       "      <td>0.768283</td>\n",
+       "      <td>7.894065</td>\n",
+       "      <td>8.162397</td>\n",
+       "      <td>0.800142</td>\n",
+       "      <td>0.781534</td>\n",
+       "      <td>0.827341</td>\n",
+       "      <td>0.805791</td>\n",
+       "      <td>0.078943</td>\n",
+       "      <td>0.244876</td>\n",
+       "      <td>689.406521</td>\n",
+       "      <td>716.239659</td>\n",
+       "      <td>0.239112</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>0.040000</td>\n",
+       "      <td>0.728944</td>\n",
+       "      <td>7.604888</td>\n",
+       "      <td>8.023025</td>\n",
+       "      <td>0.770831</td>\n",
+       "      <td>0.751144</td>\n",
+       "      <td>0.813214</td>\n",
+       "      <td>0.792130</td>\n",
+       "      <td>0.076046</td>\n",
+       "      <td>0.320922</td>\n",
+       "      <td>660.488801</td>\n",
+       "      <td>702.302462</td>\n",
+       "      <td>0.312608</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>0.050000</td>\n",
+       "      <td>0.376858</td>\n",
+       "      <td>5.845286</td>\n",
+       "      <td>7.587469</td>\n",
+       "      <td>0.592478</td>\n",
+       "      <td>0.588207</td>\n",
+       "      <td>0.769066</td>\n",
+       "      <td>0.751344</td>\n",
+       "      <td>0.058454</td>\n",
+       "      <td>0.379376</td>\n",
+       "      <td>484.528645</td>\n",
+       "      <td>658.746923</td>\n",
+       "      <td>0.366527</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>6</td>\n",
+       "      <td>0.100000</td>\n",
+       "      <td>0.140183</td>\n",
+       "      <td>1.934508</td>\n",
+       "      <td>4.761005</td>\n",
+       "      <td>0.196082</td>\n",
+       "      <td>0.177726</td>\n",
+       "      <td>0.482576</td>\n",
+       "      <td>0.464537</td>\n",
+       "      <td>0.096725</td>\n",
+       "      <td>0.476101</td>\n",
+       "      <td>93.450753</td>\n",
+       "      <td>376.100516</td>\n",
+       "      <td>0.418523</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>7</td>\n",
+       "      <td>0.150000</td>\n",
+       "      <td>0.116331</td>\n",
+       "      <td>1.335611</td>\n",
+       "      <td>3.619212</td>\n",
+       "      <td>0.135378</td>\n",
+       "      <td>0.127054</td>\n",
+       "      <td>0.366843</td>\n",
+       "      <td>0.352043</td>\n",
+       "      <td>0.066780</td>\n",
+       "      <td>0.542882</td>\n",
+       "      <td>33.561146</td>\n",
+       "      <td>261.921178</td>\n",
+       "      <td>0.437196</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>8</td>\n",
+       "      <td>0.200001</td>\n",
+       "      <td>0.101635</td>\n",
+       "      <td>1.145526</td>\n",
+       "      <td>3.000784</td>\n",
+       "      <td>0.116110</td>\n",
+       "      <td>0.108554</td>\n",
+       "      <td>0.304159</td>\n",
+       "      <td>0.291170</td>\n",
+       "      <td>0.057277</td>\n",
+       "      <td>0.600159</td>\n",
+       "      <td>14.552575</td>\n",
+       "      <td>200.078385</td>\n",
+       "      <td>0.445293</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>9</td>\n",
+       "      <td>0.300000</td>\n",
+       "      <td>0.081609</td>\n",
+       "      <td>0.936360</td>\n",
+       "      <td>2.312647</td>\n",
+       "      <td>0.094909</td>\n",
+       "      <td>0.091048</td>\n",
+       "      <td>0.234410</td>\n",
+       "      <td>0.224463</td>\n",
+       "      <td>0.093635</td>\n",
+       "      <td>0.693794</td>\n",
+       "      <td>-6.364005</td>\n",
+       "      <td>131.264731</td>\n",
+       "      <td>0.438211</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>10</td>\n",
+       "      <td>0.400000</td>\n",
+       "      <td>0.065673</td>\n",
+       "      <td>0.765235</td>\n",
+       "      <td>1.925794</td>\n",
+       "      <td>0.077564</td>\n",
+       "      <td>0.073385</td>\n",
+       "      <td>0.195198</td>\n",
+       "      <td>0.186694</td>\n",
+       "      <td>0.076524</td>\n",
+       "      <td>0.770318</td>\n",
+       "      <td>-23.476472</td>\n",
+       "      <td>92.579373</td>\n",
+       "      <td>0.412087</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>11</td>\n",
+       "      <td>0.500000</td>\n",
+       "      <td>0.053659</td>\n",
+       "      <td>0.611310</td>\n",
+       "      <td>1.662897</td>\n",
+       "      <td>0.061962</td>\n",
+       "      <td>0.059244</td>\n",
+       "      <td>0.168551</td>\n",
+       "      <td>0.161204</td>\n",
+       "      <td>0.061131</td>\n",
+       "      <td>0.831449</td>\n",
+       "      <td>-38.869035</td>\n",
+       "      <td>66.289660</td>\n",
+       "      <td>0.368833</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>12</td>\n",
+       "      <td>0.600001</td>\n",
+       "      <td>0.045913</td>\n",
+       "      <td>0.502951</td>\n",
+       "      <td>1.469572</td>\n",
+       "      <td>0.050979</td>\n",
+       "      <td>0.049540</td>\n",
+       "      <td>0.148956</td>\n",
+       "      <td>0.142593</td>\n",
+       "      <td>0.050295</td>\n",
+       "      <td>0.881744</td>\n",
+       "      <td>-49.704892</td>\n",
+       "      <td>46.957168</td>\n",
+       "      <td>0.313522</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>13</td>\n",
+       "      <td>0.700000</td>\n",
+       "      <td>0.040044</td>\n",
+       "      <td>0.417208</td>\n",
+       "      <td>1.319235</td>\n",
+       "      <td>0.042288</td>\n",
+       "      <td>0.042899</td>\n",
+       "      <td>0.133718</td>\n",
+       "      <td>0.128351</td>\n",
+       "      <td>0.041720</td>\n",
+       "      <td>0.923465</td>\n",
+       "      <td>-58.279224</td>\n",
+       "      <td>31.923506</td>\n",
+       "      <td>0.248670</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>14</td>\n",
+       "      <td>0.800000</td>\n",
+       "      <td>0.034543</td>\n",
+       "      <td>0.354182</td>\n",
+       "      <td>1.198604</td>\n",
+       "      <td>0.035900</td>\n",
+       "      <td>0.037320</td>\n",
+       "      <td>0.121490</td>\n",
+       "      <td>0.116972</td>\n",
+       "      <td>0.035418</td>\n",
+       "      <td>0.958883</td>\n",
+       "      <td>-64.581772</td>\n",
+       "      <td>19.860360</td>\n",
+       "      <td>0.176804</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>15</td>\n",
+       "      <td>0.900000</td>\n",
+       "      <td>0.027748</td>\n",
+       "      <td>0.284277</td>\n",
+       "      <td>1.097012</td>\n",
+       "      <td>0.028814</td>\n",
+       "      <td>0.031404</td>\n",
+       "      <td>0.111193</td>\n",
+       "      <td>0.107465</td>\n",
+       "      <td>0.028428</td>\n",
+       "      <td>0.987310</td>\n",
+       "      <td>-71.572289</td>\n",
+       "      <td>9.701187</td>\n",
+       "      <td>0.097159</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>16</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.001796</td>\n",
+       "      <td>0.126895</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.012862</td>\n",
+       "      <td>0.019823</td>\n",
+       "      <td>0.101360</td>\n",
+       "      <td>0.098700</td>\n",
+       "      <td>0.012690</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>-87.310507</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "    group  cumulative_data_fraction  lower_threshold      lift  \\\n",
+       "0       1                  0.010001         0.816049  8.421531   \n",
+       "1       2                  0.020000         0.794009  8.171586   \n",
+       "2       3                  0.030000         0.768283  7.894065   \n",
+       "3       4                  0.040000         0.728944  7.604888   \n",
+       "4       5                  0.050000         0.376858  5.845286   \n",
+       "5       6                  0.100000         0.140183  1.934508   \n",
+       "6       7                  0.150000         0.116331  1.335611   \n",
+       "7       8                  0.200001         0.101635  1.145526   \n",
+       "8       9                  0.300000         0.081609  0.936360   \n",
+       "9      10                  0.400000         0.065673  0.765235   \n",
+       "10     11                  0.500000         0.053659  0.611310   \n",
+       "11     12                  0.600001         0.045913  0.502951   \n",
+       "12     13                  0.700000         0.040044  0.417208   \n",
+       "13     14                  0.800000         0.034543  0.354182   \n",
+       "14     15                  0.900000         0.027748  0.284277   \n",
+       "15     16                  1.000000         0.001796  0.126895   \n",
+       "\n",
+       "    cumulative_lift  response_rate     score  cumulative_response_rate  \\\n",
+       "0          8.421531       0.853606  0.830672                  0.853606   \n",
+       "1          8.296564       0.828272  0.805167                  0.840940   \n",
+       "2          8.162397       0.800142  0.781534                  0.827341   \n",
+       "3          8.023025       0.770831  0.751144                  0.813214   \n",
+       "4          7.587469       0.592478  0.588207                  0.769066   \n",
+       "5          4.761005       0.196082  0.177726                  0.482576   \n",
+       "6          3.619212       0.135378  0.127054                  0.366843   \n",
+       "7          3.000784       0.116110  0.108554                  0.304159   \n",
+       "8          2.312647       0.094909  0.091048                  0.234410   \n",
+       "9          1.925794       0.077564  0.073385                  0.195198   \n",
+       "10         1.662897       0.061962  0.059244                  0.168551   \n",
+       "11         1.469572       0.050979  0.049540                  0.148956   \n",
+       "12         1.319235       0.042288  0.042899                  0.133718   \n",
+       "13         1.198604       0.035900  0.037320                  0.121490   \n",
+       "14         1.097012       0.028814  0.031404                  0.111193   \n",
+       "15         1.000000       0.012862  0.019823                  0.101360   \n",
+       "\n",
+       "    cumulative_score  capture_rate  cumulative_capture_rate        gain  \\\n",
+       "0           0.830672      0.084220                 0.084220  742.153097   \n",
+       "1           0.817920      0.081713                 0.165933  717.158645   \n",
+       "2           0.805791      0.078943                 0.244876  689.406521   \n",
+       "3           0.792130      0.076046                 0.320922  660.488801   \n",
+       "4           0.751344      0.058454                 0.379376  484.528645   \n",
+       "5           0.464537      0.096725                 0.476101   93.450753   \n",
+       "6           0.352043      0.066780                 0.542882   33.561146   \n",
+       "7           0.291170      0.057277                 0.600159   14.552575   \n",
+       "8           0.224463      0.093635                 0.693794   -6.364005   \n",
+       "9           0.186694      0.076524                 0.770318  -23.476472   \n",
+       "10          0.161204      0.061131                 0.831449  -38.869035   \n",
+       "11          0.142593      0.050295                 0.881744  -49.704892   \n",
+       "12          0.128351      0.041720                 0.923465  -58.279224   \n",
+       "13          0.116972      0.035418                 0.958883  -64.581772   \n",
+       "14          0.107465      0.028428                 0.987310  -71.572289   \n",
+       "15          0.098700      0.012690                 1.000000  -87.310507   \n",
+       "\n",
+       "    cumulative_gain  kolmogorov_smirnov  \n",
+       "0        742.153097            0.082591  \n",
+       "1        729.656427            0.162393  \n",
+       "2        716.239659            0.239112  \n",
+       "3        702.302462            0.312608  \n",
+       "4        658.746923            0.366527  \n",
+       "5        376.100516            0.418523  \n",
+       "6        261.921178            0.437196  \n",
+       "7        200.078385            0.445293  \n",
+       "8        131.264731            0.438211  \n",
+       "9         92.579373            0.412087  \n",
+       "10        66.289660            0.368833  \n",
+       "11        46.957168            0.313522  \n",
+       "12        31.923506            0.248670  \n",
+       "13        19.860360            0.176804  \n",
+       "14         9.701187            0.097159  \n",
+       "15         0.000000            0.000000  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "Cross-Validation Metrics Summary: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th>mean</th>\n",
+       "      <th>sd</th>\n",
+       "      <th>cv_1_valid</th>\n",
+       "      <th>cv_2_valid</th>\n",
+       "      <th>cv_3_valid</th>\n",
+       "      <th>cv_4_valid</th>\n",
+       "      <th>cv_5_valid</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>accuracy</td>\n",
+       "      <td>0.9239439</td>\n",
+       "      <td>6.1908754E-4</td>\n",
+       "      <td>0.92440516</td>\n",
+       "      <td>0.92351615</td>\n",
+       "      <td>0.9231228</td>\n",
+       "      <td>0.9240608</td>\n",
+       "      <td>0.9246143</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>auc</td>\n",
+       "      <td>0.8021638</td>\n",
+       "      <td>0.0012602289</td>\n",
+       "      <td>0.8009509</td>\n",
+       "      <td>0.80137473</td>\n",
+       "      <td>0.80144393</td>\n",
+       "      <td>0.8033893</td>\n",
+       "      <td>0.80366004</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>aucpr</td>\n",
+       "      <td>0.4995746</td>\n",
+       "      <td>0.0022548349</td>\n",
+       "      <td>0.49709672</td>\n",
+       "      <td>0.49797425</td>\n",
+       "      <td>0.49965167</td>\n",
+       "      <td>0.5002288</td>\n",
+       "      <td>0.5029216</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>err</td>\n",
+       "      <td>0.076056145</td>\n",
+       "      <td>6.1908754E-4</td>\n",
+       "      <td>0.07559483</td>\n",
+       "      <td>0.07648387</td>\n",
+       "      <td>0.076877184</td>\n",
+       "      <td>0.07593916</td>\n",
+       "      <td>0.07538567</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>err_count</td>\n",
+       "      <td>51243.8</td>\n",
+       "      <td>417.1387</td>\n",
+       "      <td>50933.0</td>\n",
+       "      <td>51532.0</td>\n",
+       "      <td>51797.0</td>\n",
+       "      <td>51165.0</td>\n",
+       "      <td>50792.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>f0point5</td>\n",
+       "      <td>0.6247343</td>\n",
+       "      <td>0.0031962688</td>\n",
+       "      <td>0.624532</td>\n",
+       "      <td>0.6235206</td>\n",
+       "      <td>0.62106586</td>\n",
+       "      <td>0.62473017</td>\n",
+       "      <td>0.6298228</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>f1</td>\n",
+       "      <td>0.51262814</td>\n",
+       "      <td>0.0020927442</td>\n",
+       "      <td>0.51327825</td>\n",
+       "      <td>0.50949</td>\n",
+       "      <td>0.5123657</td>\n",
+       "      <td>0.5127004</td>\n",
+       "      <td>0.51530653</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>f2</td>\n",
+       "      <td>0.43464056</td>\n",
+       "      <td>0.0022552824</td>\n",
+       "      <td>0.4356685</td>\n",
+       "      <td>0.43071926</td>\n",
+       "      <td>0.43604782</td>\n",
+       "      <td>0.43474048</td>\n",
+       "      <td>0.43602678</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>lift_top_group</td>\n",
+       "      <td>8.421473</td>\n",
+       "      <td>0.06115962</td>\n",
+       "      <td>8.466085</td>\n",
+       "      <td>8.333362</td>\n",
+       "      <td>8.395673</td>\n",
+       "      <td>8.4234</td>\n",
+       "      <td>8.488844</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>logloss</td>\n",
+       "      <td>0.24260618</td>\n",
+       "      <td>0.0010718547</td>\n",
+       "      <td>0.24179175</td>\n",
+       "      <td>0.243898</td>\n",
+       "      <td>0.24364957</td>\n",
+       "      <td>0.24194153</td>\n",
+       "      <td>0.24175005</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>max_per_class_error</td>\n",
+       "      <td>0.6053806</td>\n",
+       "      <td>0.0024194021</td>\n",
+       "      <td>0.6042265</td>\n",
+       "      <td>0.6095273</td>\n",
+       "      <td>0.603341</td>\n",
+       "      <td>0.6052736</td>\n",
+       "      <td>0.6045347</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>mcc</td>\n",
+       "      <td>0.5020918</td>\n",
+       "      <td>0.0025971201</td>\n",
+       "      <td>0.5025298</td>\n",
+       "      <td>0.4998116</td>\n",
+       "      <td>0.4998053</td>\n",
+       "      <td>0.5021768</td>\n",
+       "      <td>0.50613576</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>mean_per_class_accuracy</td>\n",
+       "      <td>0.6891337</td>\n",
+       "      <td>0.001123119</td>\n",
+       "      <td>0.68969077</td>\n",
+       "      <td>0.68717736</td>\n",
+       "      <td>0.6897318</td>\n",
+       "      <td>0.68919575</td>\n",
+       "      <td>0.68987286</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>mean_per_class_error</td>\n",
+       "      <td>0.3108663</td>\n",
+       "      <td>0.001123119</td>\n",
+       "      <td>0.31030926</td>\n",
+       "      <td>0.31282264</td>\n",
+       "      <td>0.3102682</td>\n",
+       "      <td>0.31080425</td>\n",
+       "      <td>0.3101271</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>mse</td>\n",
+       "      <td>0.06498795</td>\n",
+       "      <td>3.5334285E-4</td>\n",
+       "      <td>0.064677946</td>\n",
+       "      <td>0.06543439</td>\n",
+       "      <td>0.06530259</td>\n",
+       "      <td>0.064802855</td>\n",
+       "      <td>0.06472199</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>pr_auc</td>\n",
+       "      <td>0.4995746</td>\n",
+       "      <td>0.0022548349</td>\n",
+       "      <td>0.49709672</td>\n",
+       "      <td>0.49797425</td>\n",
+       "      <td>0.49965167</td>\n",
+       "      <td>0.5002288</td>\n",
+       "      <td>0.5029216</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>16</th>\n",
+       "      <td>precision</td>\n",
+       "      <td>0.73137695</td>\n",
+       "      <td>0.005741021</td>\n",
+       "      <td>0.73002064</td>\n",
+       "      <td>0.7328715</td>\n",
+       "      <td>0.7233771</td>\n",
+       "      <td>0.7312541</td>\n",
+       "      <td>0.7393614</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>17</th>\n",
+       "      <td>r2</td>\n",
+       "      <td>0.28652275</td>\n",
+       "      <td>0.0020135231</td>\n",
+       "      <td>0.2858809</td>\n",
+       "      <td>0.28392112</td>\n",
+       "      <td>0.28594545</td>\n",
+       "      <td>0.28759506</td>\n",
+       "      <td>0.28927112</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>18</th>\n",
+       "      <td>recall</td>\n",
+       "      <td>0.3946194</td>\n",
+       "      <td>0.0024194021</td>\n",
+       "      <td>0.39577347</td>\n",
+       "      <td>0.3904727</td>\n",
+       "      <td>0.39665905</td>\n",
+       "      <td>0.39472643</td>\n",
+       "      <td>0.39546534</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>19</th>\n",
+       "      <td>rmse</td>\n",
+       "      <td>0.2549266</td>\n",
+       "      <td>6.9268135E-4</td>\n",
+       "      <td>0.2543186</td>\n",
+       "      <td>0.25580147</td>\n",
+       "      <td>0.2555437</td>\n",
+       "      <td>0.25456405</td>\n",
+       "      <td>0.25440517</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                    mean            sd   cv_1_valid  \\\n",
+       "0                  accuracy    0.9239439  6.1908754E-4   0.92440516   \n",
+       "1                       auc    0.8021638  0.0012602289    0.8009509   \n",
+       "2                     aucpr    0.4995746  0.0022548349   0.49709672   \n",
+       "3                       err  0.076056145  6.1908754E-4   0.07559483   \n",
+       "4                 err_count      51243.8      417.1387      50933.0   \n",
+       "5                  f0point5    0.6247343  0.0031962688     0.624532   \n",
+       "6                        f1   0.51262814  0.0020927442   0.51327825   \n",
+       "7                        f2   0.43464056  0.0022552824    0.4356685   \n",
+       "8            lift_top_group     8.421473    0.06115962     8.466085   \n",
+       "9                   logloss   0.24260618  0.0010718547   0.24179175   \n",
+       "10      max_per_class_error    0.6053806  0.0024194021    0.6042265   \n",
+       "11                      mcc    0.5020918  0.0025971201    0.5025298   \n",
+       "12  mean_per_class_accuracy    0.6891337   0.001123119   0.68969077   \n",
+       "13     mean_per_class_error    0.3108663   0.001123119   0.31030926   \n",
+       "14                      mse   0.06498795  3.5334285E-4  0.064677946   \n",
+       "15                   pr_auc    0.4995746  0.0022548349   0.49709672   \n",
+       "16                precision   0.73137695   0.005741021   0.73002064   \n",
+       "17                       r2   0.28652275  0.0020135231    0.2858809   \n",
+       "18                   recall    0.3946194  0.0024194021   0.39577347   \n",
+       "19                     rmse    0.2549266  6.9268135E-4    0.2543186   \n",
+       "\n",
+       "    cv_2_valid   cv_3_valid   cv_4_valid  cv_5_valid  \n",
+       "0   0.92351615    0.9231228    0.9240608   0.9246143  \n",
+       "1   0.80137473   0.80144393    0.8033893  0.80366004  \n",
+       "2   0.49797425   0.49965167    0.5002288   0.5029216  \n",
+       "3   0.07648387  0.076877184   0.07593916  0.07538567  \n",
+       "4      51532.0      51797.0      51165.0     50792.0  \n",
+       "5    0.6235206   0.62106586   0.62473017   0.6298228  \n",
+       "6      0.50949    0.5123657    0.5127004  0.51530653  \n",
+       "7   0.43071926   0.43604782   0.43474048  0.43602678  \n",
+       "8     8.333362     8.395673       8.4234    8.488844  \n",
+       "9     0.243898   0.24364957   0.24194153  0.24175005  \n",
+       "10   0.6095273     0.603341    0.6052736   0.6045347  \n",
+       "11   0.4998116    0.4998053    0.5021768  0.50613576  \n",
+       "12  0.68717736    0.6897318   0.68919575  0.68987286  \n",
+       "13  0.31282264    0.3102682   0.31080425   0.3101271  \n",
+       "14  0.06543439   0.06530259  0.064802855  0.06472199  \n",
+       "15  0.49797425   0.49965167    0.5002288   0.5029216  \n",
+       "16   0.7328715    0.7233771    0.7312541   0.7393614  \n",
+       "17  0.28392112   0.28594545   0.28759506  0.28927112  \n",
+       "18   0.3904727   0.39665905   0.39472643  0.39546534  \n",
+       "19  0.25580147    0.2555437   0.25456405  0.25440517  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "See the whole table with table.as_data_frame()\n",
+      "\n",
+      "Scoring History: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th>timestamp</th>\n",
+       "      <th>duration</th>\n",
+       "      <th>number_of_trees</th>\n",
+       "      <th>training_rmse</th>\n",
+       "      <th>training_logloss</th>\n",
+       "      <th>training_auc</th>\n",
+       "      <th>training_pr_auc</th>\n",
+       "      <th>training_lift</th>\n",
+       "      <th>training_classification_error</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:34</td>\n",
+       "      <td>4 min 10.858 sec</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.639457</td>\n",
+       "      <td>1.197953</td>\n",
+       "      <td>0.500000</td>\n",
+       "      <td>0.499990</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.500010</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:39</td>\n",
+       "      <td>4 min 15.279 sec</td>\n",
+       "      <td>5.0</td>\n",
+       "      <td>0.616837</td>\n",
+       "      <td>1.091785</td>\n",
+       "      <td>0.792265</td>\n",
+       "      <td>0.824312</td>\n",
+       "      <td>1.976849</td>\n",
+       "      <td>0.322680</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:43</td>\n",
+       "      <td>4 min 19.044 sec</td>\n",
+       "      <td>10.0</td>\n",
+       "      <td>0.593316</td>\n",
+       "      <td>1.024019</td>\n",
+       "      <td>0.799346</td>\n",
+       "      <td>0.827751</td>\n",
+       "      <td>1.969823</td>\n",
+       "      <td>0.313274</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:46</td>\n",
+       "      <td>4 min 22.648 sec</td>\n",
+       "      <td>15.0</td>\n",
+       "      <td>0.569878</td>\n",
+       "      <td>0.971272</td>\n",
+       "      <td>0.802795</td>\n",
+       "      <td>0.830288</td>\n",
+       "      <td>1.972070</td>\n",
+       "      <td>0.311855</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:50</td>\n",
+       "      <td>4 min 26.494 sec</td>\n",
+       "      <td>20.0</td>\n",
+       "      <td>0.551765</td>\n",
+       "      <td>0.934923</td>\n",
+       "      <td>0.804866</td>\n",
+       "      <td>0.831898</td>\n",
+       "      <td>1.973781</td>\n",
+       "      <td>0.310356</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:54</td>\n",
+       "      <td>4 min 30.256 sec</td>\n",
+       "      <td>25.0</td>\n",
+       "      <td>0.539484</td>\n",
+       "      <td>0.910068</td>\n",
+       "      <td>0.806598</td>\n",
+       "      <td>0.833337</td>\n",
+       "      <td>1.976033</td>\n",
+       "      <td>0.304712</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:33:58</td>\n",
+       "      <td>4 min 34.008 sec</td>\n",
+       "      <td>30.0</td>\n",
+       "      <td>0.532110</td>\n",
+       "      <td>0.893191</td>\n",
+       "      <td>0.808993</td>\n",
+       "      <td>0.834841</td>\n",
+       "      <td>1.976818</td>\n",
+       "      <td>0.303763</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:01</td>\n",
+       "      <td>4 min 37.697 sec</td>\n",
+       "      <td>35.0</td>\n",
+       "      <td>0.527083</td>\n",
+       "      <td>0.881022</td>\n",
+       "      <td>0.810823</td>\n",
+       "      <td>0.836076</td>\n",
+       "      <td>1.979631</td>\n",
+       "      <td>0.304047</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:05</td>\n",
+       "      <td>4 min 41.397 sec</td>\n",
+       "      <td>40.0</td>\n",
+       "      <td>0.524119</td>\n",
+       "      <td>0.873296</td>\n",
+       "      <td>0.812040</td>\n",
+       "      <td>0.836947</td>\n",
+       "      <td>1.980430</td>\n",
+       "      <td>0.299840</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:09</td>\n",
+       "      <td>4 min 45.047 sec</td>\n",
+       "      <td>45.0</td>\n",
+       "      <td>0.522347</td>\n",
+       "      <td>0.868030</td>\n",
+       "      <td>0.813448</td>\n",
+       "      <td>0.837918</td>\n",
+       "      <td>1.980489</td>\n",
+       "      <td>0.298310</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:12</td>\n",
+       "      <td>4 min 48.648 sec</td>\n",
+       "      <td>50.0</td>\n",
+       "      <td>0.521098</td>\n",
+       "      <td>0.863513</td>\n",
+       "      <td>0.815260</td>\n",
+       "      <td>0.839107</td>\n",
+       "      <td>1.981774</td>\n",
+       "      <td>0.301199</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:14</td>\n",
+       "      <td>4 min 50.656 sec</td>\n",
+       "      <td>55.0</td>\n",
+       "      <td>0.520741</td>\n",
+       "      <td>0.862269</td>\n",
+       "      <td>0.815700</td>\n",
+       "      <td>0.839448</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>0.299180</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:15</td>\n",
+       "      <td>4 min 51.297 sec</td>\n",
+       "      <td>60.0</td>\n",
+       "      <td>0.520741</td>\n",
+       "      <td>0.862269</td>\n",
+       "      <td>0.815700</td>\n",
+       "      <td>0.839448</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>0.299180</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:16</td>\n",
+       "      <td>4 min 51.956 sec</td>\n",
+       "      <td>65.0</td>\n",
+       "      <td>0.520741</td>\n",
+       "      <td>0.862269</td>\n",
+       "      <td>0.815700</td>\n",
+       "      <td>0.839448</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>0.299180</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td></td>\n",
+       "      <td>2021-04-08 17:34:16</td>\n",
+       "      <td>4 min 52.613 sec</td>\n",
+       "      <td>70.0</td>\n",
+       "      <td>0.520741</td>\n",
+       "      <td>0.862269</td>\n",
+       "      <td>0.815700</td>\n",
+       "      <td>0.839448</td>\n",
+       "      <td>1.981545</td>\n",
+       "      <td>0.299180</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                timestamp           duration  number_of_trees  training_rmse  \\\n",
+       "0     2021-04-08 17:33:34   4 min 10.858 sec              0.0       0.639457   \n",
+       "1     2021-04-08 17:33:39   4 min 15.279 sec              5.0       0.616837   \n",
+       "2     2021-04-08 17:33:43   4 min 19.044 sec             10.0       0.593316   \n",
+       "3     2021-04-08 17:33:46   4 min 22.648 sec             15.0       0.569878   \n",
+       "4     2021-04-08 17:33:50   4 min 26.494 sec             20.0       0.551765   \n",
+       "5     2021-04-08 17:33:54   4 min 30.256 sec             25.0       0.539484   \n",
+       "6     2021-04-08 17:33:58   4 min 34.008 sec             30.0       0.532110   \n",
+       "7     2021-04-08 17:34:01   4 min 37.697 sec             35.0       0.527083   \n",
+       "8     2021-04-08 17:34:05   4 min 41.397 sec             40.0       0.524119   \n",
+       "9     2021-04-08 17:34:09   4 min 45.047 sec             45.0       0.522347   \n",
+       "10    2021-04-08 17:34:12   4 min 48.648 sec             50.0       0.521098   \n",
+       "11    2021-04-08 17:34:14   4 min 50.656 sec             55.0       0.520741   \n",
+       "12    2021-04-08 17:34:15   4 min 51.297 sec             60.0       0.520741   \n",
+       "13    2021-04-08 17:34:16   4 min 51.956 sec             65.0       0.520741   \n",
+       "14    2021-04-08 17:34:16   4 min 52.613 sec             70.0       0.520741   \n",
+       "\n",
+       "    training_logloss  training_auc  training_pr_auc  training_lift  \\\n",
+       "0           1.197953      0.500000         0.499990       1.000000   \n",
+       "1           1.091785      0.792265         0.824312       1.976849   \n",
+       "2           1.024019      0.799346         0.827751       1.969823   \n",
+       "3           0.971272      0.802795         0.830288       1.972070   \n",
+       "4           0.934923      0.804866         0.831898       1.973781   \n",
+       "5           0.910068      0.806598         0.833337       1.976033   \n",
+       "6           0.893191      0.808993         0.834841       1.976818   \n",
+       "7           0.881022      0.810823         0.836076       1.979631   \n",
+       "8           0.873296      0.812040         0.836947       1.980430   \n",
+       "9           0.868030      0.813448         0.837918       1.980489   \n",
+       "10          0.863513      0.815260         0.839107       1.981774   \n",
+       "11          0.862269      0.815700         0.839448       1.981545   \n",
+       "12          0.862269      0.815700         0.839448       1.981545   \n",
+       "13          0.862269      0.815700         0.839448       1.981545   \n",
+       "14          0.862269      0.815700         0.839448       1.981545   \n",
+       "\n",
+       "    training_classification_error  \n",
+       "0                        0.500010  \n",
+       "1                        0.322680  \n",
+       "2                        0.313274  \n",
+       "3                        0.311855  \n",
+       "4                        0.310356  \n",
+       "5                        0.304712  \n",
+       "6                        0.303763  \n",
+       "7                        0.304047  \n",
+       "8                        0.299840  \n",
+       "9                        0.298310  \n",
+       "10                       0.301199  \n",
+       "11                       0.299180  \n",
+       "12                       0.299180  \n",
+       "13                       0.299180  \n",
+       "14                       0.299180  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Variable Importances: \n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>variable</th>\n",
+       "      <th>relative_importance</th>\n",
+       "      <th>scaled_importance</th>\n",
+       "      <th>percentage</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>coupon_type_department</td>\n",
+       "      <td>1.227769e+06</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.534741</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>coupon_mean_prod_price</td>\n",
+       "      <td>2.428440e+05</td>\n",
+       "      <td>0.197793</td>\n",
+       "      <td>0.105768</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>coupon_how_many</td>\n",
+       "      <td>1.721953e+05</td>\n",
+       "      <td>0.140251</td>\n",
+       "      <td>0.074998</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>coupon_type_just_discount</td>\n",
+       "      <td>1.354846e+05</td>\n",
+       "      <td>0.110350</td>\n",
+       "      <td>0.059009</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>coupon_type_buy_all</td>\n",
+       "      <td>1.111026e+05</td>\n",
+       "      <td>0.090492</td>\n",
+       "      <td>0.048390</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>coupon_discount</td>\n",
+       "      <td>1.098288e+05</td>\n",
+       "      <td>0.089454</td>\n",
+       "      <td>0.047835</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>coupon_type_buy_more</td>\n",
+       "      <td>8.621582e+04</td>\n",
+       "      <td>0.070222</td>\n",
+       "      <td>0.037550</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>coupon_days_valid</td>\n",
+       "      <td>7.937698e+04</td>\n",
+       "      <td>0.064651</td>\n",
+       "      <td>0.034572</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>cust_unique_coupons_used</td>\n",
+       "      <td>2.874513e+04</td>\n",
+       "      <td>0.023412</td>\n",
+       "      <td>0.012520</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>cust_mean_discount</td>\n",
+       "      <td>2.412111e+04</td>\n",
+       "      <td>0.019646</td>\n",
+       "      <td>0.010506</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>cust_total_products_bougth</td>\n",
+       "      <td>2.372456e+04</td>\n",
+       "      <td>0.019323</td>\n",
+       "      <td>0.010333</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>cust_mean_product_price</td>\n",
+       "      <td>1.860420e+04</td>\n",
+       "      <td>0.015153</td>\n",
+       "      <td>0.008103</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>cust_unique_products_bought</td>\n",
+       "      <td>1.549414e+04</td>\n",
+       "      <td>0.012620</td>\n",
+       "      <td>0.006748</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>cust_credit</td>\n",
+       "      <td>4.810139e+03</td>\n",
+       "      <td>0.003918</td>\n",
+       "      <td>0.002095</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>cust_age_mid</td>\n",
+       "      <td>4.308483e+03</td>\n",
+       "      <td>0.003509</td>\n",
+       "      <td>0.001877</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>15</th>\n",
+       "      <td>cust_age_old</td>\n",
+       "      <td>3.579277e+03</td>\n",
+       "      <td>0.002915</td>\n",
+       "      <td>0.001559</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>16</th>\n",
+       "      <td>cust_gender_F</td>\n",
+       "      <td>3.050989e+03</td>\n",
+       "      <td>0.002485</td>\n",
+       "      <td>0.001329</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>17</th>\n",
+       "      <td>cust_age_young</td>\n",
+       "      <td>2.436828e+03</td>\n",
+       "      <td>0.001985</td>\n",
+       "      <td>0.001061</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>18</th>\n",
+       "      <td>cust_gender_M</td>\n",
+       "      <td>2.314412e+03</td>\n",
+       "      <td>0.001885</td>\n",
+       "      <td>0.001008</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       variable  relative_importance  scaled_importance  \\\n",
+       "0        coupon_type_department         1.227769e+06           1.000000   \n",
+       "1        coupon_mean_prod_price         2.428440e+05           0.197793   \n",
+       "2               coupon_how_many         1.721953e+05           0.140251   \n",
+       "3     coupon_type_just_discount         1.354846e+05           0.110350   \n",
+       "4           coupon_type_buy_all         1.111026e+05           0.090492   \n",
+       "5               coupon_discount         1.098288e+05           0.089454   \n",
+       "6          coupon_type_buy_more         8.621582e+04           0.070222   \n",
+       "7             coupon_days_valid         7.937698e+04           0.064651   \n",
+       "8      cust_unique_coupons_used         2.874513e+04           0.023412   \n",
+       "9            cust_mean_discount         2.412111e+04           0.019646   \n",
+       "10   cust_total_products_bougth         2.372456e+04           0.019323   \n",
+       "11      cust_mean_product_price         1.860420e+04           0.015153   \n",
+       "12  cust_unique_products_bought         1.549414e+04           0.012620   \n",
+       "13                  cust_credit         4.810139e+03           0.003918   \n",
+       "14                 cust_age_mid         4.308483e+03           0.003509   \n",
+       "15                 cust_age_old         3.579277e+03           0.002915   \n",
+       "16                cust_gender_F         3.050989e+03           0.002485   \n",
+       "17               cust_age_young         2.436828e+03           0.001985   \n",
+       "18                cust_gender_M         2.314412e+03           0.001885   \n",
+       "\n",
+       "    percentage  \n",
+       "0     0.534741  \n",
+       "1     0.105768  \n",
+       "2     0.074998  \n",
+       "3     0.059009  \n",
+       "4     0.048390  \n",
+       "5     0.047835  \n",
+       "6     0.037550  \n",
+       "7     0.034572  \n",
+       "8     0.012520  \n",
+       "9     0.010506  \n",
+       "10    0.010333  \n",
+       "11    0.008103  \n",
+       "12    0.006748  \n",
+       "13    0.002095  \n",
+       "14    0.001877  \n",
+       "15    0.001559  \n",
+       "16    0.001329  \n",
+       "17    0.001061  \n",
+       "18    0.001008  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": []
+     },
+     "execution_count": 16,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "h2o.get_model('GBM_4_AutoML_20210408_164243')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/training-with-artificial-data/03_training.ipynb b/training-with-artificial-data/03_training.ipynb
new file mode 100644
index 0000000..547aa17
--- /dev/null
+++ b/training-with-artificial-data/03_training.ipynb
@@ -0,0 +1,643 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "unusual-blade",
+   "metadata": {},
+   "source": [
+    "# Training with scikit-learn"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "impressive-poison",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import pickle\n",
+    "\n",
+    "from imblearn.combine import SMOTETomek\n",
+    "from imblearn.over_sampling import SMOTE\n",
+    "from matplotlib import pyplot\n",
+    "import pandas as pd\n",
+    "from sklearn.ensemble import GradientBoostingClassifier\n",
+    "from sklearn.metrics import classification_report, auc, precision_recall_curve, f1_score\n",
+    "from sklearn.model_selection import train_test_split"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "determined-radius",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dir = 'data_0408_0'\n",
+    "train_file_path = os.path.join(data_dir, 'train.csv')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "technical-mapping",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train = pd.read_csv(train_file_path)\n",
+    "train.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "federal-character",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train.sample(10)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "natural-argentina",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "train.coupon_used.value_counts(normalize=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "congressional-hartford",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_train = train['coupon_used']\n",
+    "X_train = train.drop(['coupon_used'], axis=1)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(X_train, y_train, test_size=0.2, random_state=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "published-retrieval",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "X_all = X_train.append(X_test)\n",
+    "y_all = y_train.append(y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "incredible-secretariat",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def plot_pr_curve(probs, preds, y, legend=''):\n",
+    "    precision, recall, _ = precision_recall_curve(y, probs)\n",
+    "    f1_, auc_ = f1_score(y, preds), auc(recall, precision)\n",
+    "    # summarize scores\n",
+    "    print(f'{legend}:\\nf1={round(f1_, 3)} auc={round(auc_, 3)}')\n",
+    "    # plot the precision-recall curves\n",
+    "    no_skill = len(y[y==1]) / len(y)\n",
+    "    pyplot.plot([0, 1], [no_skill, no_skill], linestyle='--', label='No Skill')\n",
+    "    pyplot.plot(recall, precision, marker='.', label='GBM')\n",
+    "    # axis labels\n",
+    "    pyplot.xlabel('Recall')\n",
+    "    pyplot.ylabel('Precision')\n",
+    "    # show the legend\n",
+    "    pyplot.legend()\n",
+    "    # show the plot\n",
+    "    pyplot.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cleared-national",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "gbm_params = {\n",
+    "    'n_estimators': 70,\n",
+    "    'max_depth': 10,\n",
+    "    'max_leaf_nodes': 994\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "exclusive-enlargement",
+   "metadata": {},
+   "source": [
+    "## 1. Training with no balancing"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "modern-fleece",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "gbm = GradientBoostingClassifier(**gbm_params)\n",
+    "gbm.fit(X_train, y_train)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "documented-county",
+   "metadata": {},
+   "source": [
+    "#### 1.1. Evaluating on the test dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "oriental-backing",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm.predict_proba(X_test)[:, 1]\n",
+    "preds = gbm.predict(X_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "alleged-modern",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_test, legend='GBM trained on an unbalanced dataset, evaluated on the test dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "executed-seattle",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_test, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "extraordinary-fisher",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_test, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "regulated-imperial",
+   "metadata": {},
+   "source": [
+    "#### 1.2 Evaluating on the whole dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "declared-tyler",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm.predict_proba(X_all)[:, 1]\n",
+    "preds = gbm.predict(X_all)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "married-garlic",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_all, legend='GBM trained on an unbalanced dataset, evaluated on the whole dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "funky-integral",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_all, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "separated-sugar",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_all, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "hollow-complex",
+   "metadata": {},
+   "source": [
+    "#### 1.3. Pickle the model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "sensitive-reggae",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open(os.path.join(data_dir, 'pickled_model_gbm_no_balancing'), 'wb') as f:\n",
+    "    pickle.dump(gbm, f)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "intense-injury",
+   "metadata": {},
+   "source": [
+    "## 2. Training with balancing - SMOTE + Tomek"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "right-three",
+   "metadata": {},
+   "source": [
+    "#### 2.1. Balancing"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "sharp-routine",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# NOTE! This takes very long\n",
+    "smt = SMOTETomek()\n",
+    "X_train_smt, y_train_smt = smt.fit_resample(X_train, y_train)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "recognized-stone",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_train_smt.coupon_used.value_counts(normalize=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "emotional-valley",
+   "metadata": {},
+   "source": [
+    "#### 2.2. Training"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "applicable-taiwan",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "gbm_smt = GradientBoostingClassifier(**gbm_params)\n",
+    "gbm_smt.fit(X_train_smt, y_train_smt)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "progressive-phenomenon",
+   "metadata": {},
+   "source": [
+    "#### 2.3. Evaluating on the test dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "swedish-soldier",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm_smt.predict_proba(X_test)[:, 1]\n",
+    "preds = gbm_smt.predict(X_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "acceptable-bradford",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_test,\n",
+    "              legend='GBM trained on a balanced dataset (SMOTE+Tomek), evaluated on the test dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "lyric-scroll",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_test, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "lesbian-hybrid",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_test, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "deadly-medline",
+   "metadata": {},
+   "source": [
+    "#### 2.4. Evaluating on the entire dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "tested-tower",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm_smt.predict_proba(X_all)[:, 1]\n",
+    "preds = gbm_smt.predict(X_all)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "japanese-complaint",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_all,\n",
+    "              legend='GBM trained on a balanced dataset (SMOTE+Tomek), evaluated on the whole dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "middle-history",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_all, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "removable-causing",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_all, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "loved-present",
+   "metadata": {},
+   "source": [
+    "#### 2.5. Pickle the model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "vocal-twist",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open(os.path.join(data_dir, 'pickled_model_gbm_smote_tomek'), 'wb') as f:\n",
+    "    pickle.dump(gbm_smt, f)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "uniform-malaysia",
+   "metadata": {},
+   "source": [
+    "## 3. Training with balancing (SMOTE)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "republican-foundation",
+   "metadata": {},
+   "source": [
+    "#### 3.1. Balancing"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "congressional-suffering",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "smote = SMOTE(sampling_strategy=0.5)\n",
+    "X_train_sm, y_train_sm = smote.fit_resample(X_train, y_train)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "framed-objective",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_train_sm.value_counts(normalize=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "voluntary-orlando",
+   "metadata": {},
+   "source": [
+    "#### 3.2 Training"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "lightweight-controversy",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "gbm_sm = GradientBoostingClassifier(**gbm_params)\n",
+    "gbm_sm.fit(X_train_sm, y_train_sm)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "equal-words",
+   "metadata": {},
+   "source": [
+    "#### 3.3 Evaluating on the test dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "miniature-contribution",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm_sm.predict_proba(X_test)[:, 1]\n",
+    "preds = gbm_sm.predict(X_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "improved-discretion",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_test,\n",
+    "              legend='GBM trained on a balanced dataset (SMOTE), evaluated on the test dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "lesbian-charity",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_test, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "remarkable-disclaimer",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_test, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "gothic-wales",
+   "metadata": {},
+   "source": [
+    "#### 3.4 Evaluating on the entire dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ambient-carol",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probs = gbm_sm.predict_proba(X_all)[:, 1]\n",
+    "preds = gbm_sm.predict(X_all)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fatal-terrace",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot_pr_curve(probs, preds, y=y_all,\n",
+    "              legend='GBM trained on a balanced dataset (SMOTE), evaluated on the whole dataset')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "incoming-raleigh",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.crosstab(y_all, preds, rownames=['Actual'], colnames=['Predicted'], margins=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "executive-festival",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(classification_report(y_all, preds))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "preliminary-large",
+   "metadata": {},
+   "source": [
+    "#### 3.5. Pickle the model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "enabling-cyprus",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open(os.path.join(data_dir, 'pickled_model_gbm_smote'), 'wb') as f:\n",
+    "    pickle.dump(gbm_sm, f)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/training-with-artificial-data/99_data_prep_v0.ipynb b/training-with-artificial-data/99_data_prep_v0.ipynb
new file mode 100644
index 0000000..f97e7d5
--- /dev/null
+++ b/training-with-artificial-data/99_data_prep_v0.ipynb
@@ -0,0 +1,1103 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "abstract-belize",
+   "metadata": {},
+   "source": [
+    "# Prepare data for training dataframe\n",
+    "\n",
+    "NOTE: This notebook is not finished. The work was aborted when it turned out that the resulting dataset is highly imbalanced. For adjusted and finished version, see the other data_prep notebook.\n",
+    "\n",
+    "The aim is to produce a table where each `order_id` is mapped to all coupons available on the date the order was made, along with information (`True`/`False`) if the coupon was used in that order.\n",
+    "\n",
+    "The steps are as follows:\n",
+    "1. Based on the `coupons` table, create a `coupon_dates` dataframe which maps a date to all coupons available on that date.\n",
+    "2. Merge `coupon_dates` with `orders` in order to create a dataframe mapping `order_id` to all coupons available on the date that order was made -> `order_coupons_available`\n",
+    "3. From the `order_details` table, select only `order_id` - `coupon_id` pairs, resulting in `order_coupons_used` - a dataframe mapping an order to the coupons used in that order, if any.\n",
+    "4. Combine `order_coupons_available` and `order_coupons_used`, add column `coupon_used` which stores `True` for coupons used in an order, and `False` for coupons not used in an order."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "coral-advertiser",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import datetime\n",
+    "import os\n",
+    "\n",
+    "from IPython.display import Image\n",
+    "import numpy as np\n",
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "orange-ministry",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwYAAAIeCAYAAAABNBiWAAAKmXRFWHRteGZpbGUAJTNDbXhmaWxlJTIwaG9zdCUzRCUyMmFwcC5kaWFncmFtcy5uZXQlMjIlMjBtb2RpZmllZCUzRCUyMjIwMjEtMDQtMDhUMDklM0E0NyUzQTM1LjI4M1olMjIlMjBhZ2VudCUzRCUyMjUuMCUyMChXaW5kb3dzJTIwTlQlMjAxMC4wJTNCJTIwV2luNjQlM0IlMjB4NjQpJTIwQXBwbGVXZWJLaXQlMkY1MzcuMzYlMjAoS0hUTUwlMkMlMjBsaWtlJTIwR2Vja28pJTIwQ2hyb21lJTJGODkuMC40Mzg5LjExNCUyMFNhZmFyaSUyRjUzNy4zNiUyMiUyMGV0YWclM0QlMjJoMS1ZdEpGb0ctaEZXUG5ZNDJhLSUyMiUyMHZlcnNpb24lM0QlMjIxMy4xMC42JTIyJTIwdHlwZSUzRCUyMmRldmljZSUyMiUzRSUzQ2RpYWdyYW0lMjBuYW1lJTNEJTIyUGFnZS0xJTIyJTIwaWQlM0QlMjJlZmE3YTBhMS1iZjliLWEzMGUtZTZkZi05NGE3NzkxYzA5ZTklMjIlM0U3VnR0YzlvNEVQNDFUSm9QeVJnYkhQZ1lDRWx2THVuMW11dGNlMThZWVF1alZyWmNXU2JRWDM5NnhhOGs1QVZNRWpLZEdxOVcwa3BhN1Q2N2tsdk9NRnhjVVJEUGJvZ1BjY3UyJTJGRVhMdVdqWmRydmQ2JTJGS0hvQ3dWeFhVMElhREkxMHdaNFJiOWhwcG9hV3FLZkpnVUdCa2htS0c0U1BSSUZFR1BGV2lBVW5KWFpKc1NYT3cxQmdHc0VHNDlnS3ZVZjVIUFpvcmFzOTJNJTJGaEdpWUdaNmJydDlWVElCM3MlMkJBa2pUUyUyRmJWc1p5ciUyRlZIRUlURnQ2b01rTSUyQk9RdVIzSkdMV2RJQ1dIcVY3Z1lRaXptMWt5YnFuZTVwblFsTjRVUjI2U0NyZGRsRG5BS2pjZ3U1blVIc1JDUExmV1V1TDlTSWRNZ0JEUkFVY3M1NTZWV3ZPRCUyRmM2SWNtYUNmTUJLcnNrNnVqTUVGT3dFWUJicWV4NFdEdEZqc1E0OVF3QkRSUEh3R0ljVW9nbG5YJTJGRmVnbjFMQWhGRVNCWWJxcFFraklhU0pLZWFETG5Ma0N1SXliVWJMbENjTkg4TXBVNFU5VVZpV25NJTJCNmJYMUFFVHZtejYlMkJmJTJGdmo3NiUyQmclMkJvYlloUWdSQ0tJVGdVM084Njc0REtCWlY5RTdFdWxwa0tocm9EbzV1amxyMmtQOCUyQnVqeHFkUzkyTHBjd0JxdDEyVzNYOFV4TlJTUHJBWHlmd2lScHFuc1BzV1ZUZmNja1lRQVB1ZU5xYXVVOTdpUVliV3dDUEFwOXhLNVJpTmo5TTJBWE9yZm5rRExFSGVXNU11Y1gwdUFQdEhHJTJGVUowT0NPZWFZdW5YcG9qN0cyY3dKUkhUYnI1dDYlMkZkTEVDSXNBTUpIaU9kUXRDb01NUXV4WU5MMiUyRlNjY0VreW83TnV4TE1keFhWRml2S1pnRXoza21DNDZvJTJGWmxWM2RSclZ6eGo5cGxpbkhCUlk2ayUyRmVVVjVFNUZyWk11N1dqWHJhR04zVDUxTzRweWwwR0Z0cXU1Wm5tWTBPdWRPbG9Fb0JGS3NHbyUyRjg5TDhoM2JVOVU3YlolMkYlMkZkJTJGZlhuSnhCJTJGdmsyJTJGb0t2SjJYSVluYlNkVG1XeG9NOVJqWDRsbE0xSVFDS0FSeGwxSUtFS0ZPMEs3Y2w0cm9sWVZ6bSUyRlB5QmpTNzEySUdXa3VFaHdnZGkzM08lMkZ2b3FuVHJuNjdXT2lXNWN2U3ZBakYlMkY1WXhpdGZ2JTJCYktzbW53ejlmaVdwZXhjd0R0T0dIMzVEU241aDl5QWFHbktMcVd5U1ZuVVhJZ0pXSXVNTkNraEtmWGdmVFBiMW12R3V3Z2d1NDl6bmVwUWlEbTJtUmRsZVhrVk1JSyUyQlJqUTN5Um5HbUVSNUdEY3A4NzVyQk1lV01hekJVRDg0JTJGaDM3S0pGJTJCeGVDcFNib2NoNFJDOCUyQjVEdmdGWUNJc2NBT05HY0ZjbVRWTmUwTXhYemdXQ1VEaTBhSkxFcWlHeEpjVSUyRks0WlVxTG9DaTNMYXM5cTdsWHBHN3NhaE1Ic2JTTTE1WmJ1U082YkVUejBtRUYlMkZJMVlVJTJGSm1Jd0U1SUtIeVZpYkxGWEUwRUREWTV2RmVERklESTBhZCUyRkhQbUI1dEx4MjFHd20ySlFsRVZZelVVWlI3dDRFUlI1YzhhZ1dzNGd4MSUyQlUlMkJESnRIVEM4eWFMNktDSmNHTFpUSHc2bVBvdUQ0VVZQd0ZpRmhCZiUyRlZZSVcxa05EdXVCdGd3azY3RGhOMnRvb0ozU1l4WVlZRFYlMkJqdTBaaXclMkZSaE02R0dRSk1qYk5odzgyeGdPV3ZWcXN5czRlUFlHNENDSzVyeEZJdVpqSHdHaGRxbmpQUUNHdjFMQXJUekx3NEtEOFg3WWVMdE9yMkM4T3oxcnMzRCUyQmJKdVcyJTJCNDNhcmxiVDRqbVY2WTZWMnZ6YUQ2c3h2SFdTeHZ1M3FhR3U5MSUyRnB1SFdWVDhUSklNTGd4TmNwNkJxam1PZFdybSUyRlVvTktUdDFHU2FGV1FqMEhIVmhONnRqVE1rYlAwckhkWll3MjFyU0dJWUpqdnl1QXVKY3FvTEoyemFIRTNodEFpWVJ5M3JFUEdVRDRrRHBjSjRLYXBid2dPejUxcXdITHpXRGtzWndLMkpnY2hDTE9CZkE0cGtobW96NXdmQTBlU0VCcTFnbWNFbXB5ZENiZHMxdnhSZGI0Q1pLRHFkcjB6UW11ZXMwVWtEJTJCSWtBaUdNVnNlSXFiTklpYkhWREhwTHJjYU1kbTFCNkRtN0hRTFhxeiUyRlZyell3WDJ0M2J6NmlsZUQlMkZrUGpqR0pHJTJGbUF6TnNpeWRJc213OXdRelNkWjZ2TGolMkZhMVpETWM1aEQ3YkNYM001ZGE5dnk5aEJIM1ZUaU03MUQyNGpiMjc4dXB4VHhHUTVxNzVKZHdPTlhiQmN3NGpYNkxyWnU1MndNU2pLSmFicWlFUjZzT2t2YnhqY3dBc1pjRFNMeUdXYnMyeFVPMkp2cnMxeUxLNnJQcWElMkZaVU93YlhiMmslMkJ2WmRyTzM0VXBwdzRldWpNenFVamFUSnJ0c09jZnNlZUxkN3RYSjJUNVBXJTJGWDdQbXo3VzE1dTdMbHI4RkVmRzVYamwwaXYzb041cDBGSGlZSmxBODgxQmR1TDM2VVc5SVVwNXpaVXNKV3ptNnJSOExkQnhwYWN3ak1aeGNzYzJ5eFlFald5N3RTMjFMT0xsTk8xZUxMbmpEYlRrVjdDMG1jbXZoYmFUZVB0Q0hIcm1BaWk2eWlLaHVUcEQyVEFKcEZ5eFVpMzVmaHV1cEJmeVdxRzJ1dFByNTh3RmJjdHlGZlBtVGxyOWxubjJvQnNtOXJuZEglMkYlM0MlMkZkaWFncmFtJTNFJTNDJTJGbXhmaWxlJTNFezAxywAAIABJREFUeF7svXv8VlWVP74dU0JEB9EcrzHq10SZVETNlEFowvD2GZgJLwxf6IVRpMJP0EZUpinyUgqmaRhGaYwW1g/DazoTXpBSQ5DCQEYdisQL4g2UQZ3h+1pb1+P6rM/Z5+zznPM85+xz3s8/8Hmec/blvddee7/3WnutbcwR47d2DB5g8AECRSCw4IElZuuS7xdRdaY6txnwJYN5kwlCvAwEgAAQAAIxCIS6PmJQw0ZgGyIGK+ZfHXYv0PpgEeg34rxgiQHmTbBih4YDASAABEqPQKjrY+mBRQNjEQAxgIAUigApvhfv/oZtw+67715oW9JUThYDEIM0iOFZIAAEgAAQSINAqOtjmj7i2fIhAGJQvjGpVYtI8S36zmjzP//zP2bQoEHB9B3EIJihQkOBABAAAkEiEOr6GCTYaHQDARADCEOhCJDie3DGmbYNIAaFDgUqBwJAAAgAgRIhEOr6WCII0ZQmEAAxaAI0vJIfAqEqPlgM8pMBlAQEgAAQAAJdEQh1fcRYho0AiEHM+P33lnfMrNvuM2M7hpheO/UIe6RL2vpQFR+IQUkFCs0CAkAACFQEgVDXx4rAX9tugBjEDP3Vc+80v33qWXP9RV8EMWjRFAlV8YEYtEggUCwQAAJAAAhYBEJdHzF8YSNQKWKw5KlnzNhp19sR2XO3XcwN08abPXfrZb72vXn2u69/5TSzecu75uzLbjRHHrK/OW/0KfZ7IgBzbl9o/3/SwP72uXsXLzPTrvtpp7L223v3yGc/2m17c/vCx8ysefebMacOMpfPud2+N3XccPO71X80dy9aaj554McbBOO1N9+ybaDf6HPT9LPNgEMOMM/9+SXz5emzzcAj+ppFT6w0hx/Ux7aFrBa6fVRnFT6hKj4QgypIH/ogEWD9s279q530Ev3h+o2/n3DaUDN8yNGGdTDptH4H7Gt178a3NtvyHl66sqGXSZf6lPv5zx5jHliywurKccOHNHR2XFsxqkCgKgiEuj5WBf+69qMyxIAXissmntFYkN7fnI8wl8+Z7yQGciHbZeeedmPOi5y2GMi/qUBJMIgYEJGYfs7pZtixh9sFkQgBLZBc7rDjDjMTRp7QiagQASFCQSSGPlQ/EwLa/Me1rwpCG6riAzGogvShD4wAH1Z8fugxdoMvdd1rb27qpBf5EETqrDhisGzVGqvfeu20o9WZ++ze2x54rFv/WmK5rAulntSHPXRwAssuZLmKCIS6PlZxLOrUp8oQA1qsfnb/b7q4/dA9gTiLgbQy8Mk9C0AcEaBn5AK5fPWaxgafLQu8WEkS0TH4qE6LoSQ0mpjQe3Htq4Kghqr4QAyqIH3oAyNAeuaia39iN/B8ms+/ST1HvzGJIKur1mdRFgMqh4iAtKxSPVpnRpVLhylk2Y0qlw5e2MJbFQsqJBIISARCXR8ximEjUBli4LoPkEQM+HdaZPjDBCHu1KxZYjCwf9+Gu5MUHbI0HHpgn06kgX6Pa1/Yovd+60NVfCAGVZA+9MG1+ZfIaN2aFzFY8MDjnU76fcslt0t5YEJtBUGALFcRgVDXxyqORZ36VBlioE+1eBCTiIEcbHYH4vsANy1Y2Fi46Dl9NyEPi4GsX/vrakHU7atCpKRQFR+IQZ3UZPX7Kk/kaeOt9SK7O7bbYhDloiTbJ+9r0eEKuUHhAwSqgkCo62NV8K9rPypDDPSmWp5yyQ3+s2tfsCf2fJFNbu6TfFeT7hjIxdP1rM8dA14MtVVCt68K5vNQFR+IQV1VZjX7LU/ryXUn7h6B/I3vDXAwBw7kIC8f812r/ffZI/UdgyhiwOVwnXGkppqjhV7VBYFQ18e6jE9V+1kZYkADFBWViE64ZAQLMjmvfWmDMyoRRzOi96L8+2UEIxklQ1ss4khEUlQiSQyoX7JO2b4qCGWoig/EoArShz5EWSyTohJpHcSWTCqLdCJFUJPEgL7fsUd3M++XiztFZ6PvpW6W5cZFO5IR3LitUhdjVIFAVRAIdX2sCv517UeliEFdBzHkfoeq+EAMQpY6tL0dCGg3zipYONuBG+oAAoxAqOsjRjBsBCpBDO5dtNRccPXcsEciZeuvPG+0GTawf8q3yvd4qIoPxKB8soQWNY8AzcO8P0t+8q1OEeEGnPHPeVfRqbxtt/0Ls/xnM1paBwoHAu1EINT1sZ0Yoa78EagEMcgfFpTYLgRCVXwgBu2SENQDBIAAEKgnAqGuj/Ucrer0GsSgOmMZZE9CVXwgBkGKGxoNBIAAEAgGgVDXx2AARkMjEQAxgGAUikCoig/EoFCxQeVAAAgAgcojEOr6WPmBqXgHQQwqPsBl716oig/EoOyShfYBASAABMJGINT1MWzU0XoQA8hAoQiEqvhADAoVG1QOBIAAEKg8AqGuj5UfmIp3EMSg4gNc9u6FqvhADMouWWgfEAACQCBsBEJdH8NGHa0HMYAMFIpAqIoPxKBQsUHlQAAIAIHKIxDq+lj5gal4B7eZPXv21or3Ed0rKQIHH3ywOW7Sj82DM860LRw0aFBJW9q1WTfeeGMwbUVDgQAQAAJAICwEQl4fw0IardUIbPPII49sJQHEBwgUgcAun70wSGKwePFig3lThMSgTiAABIBAPRAIdX2sx+hUt5dwJaru2AbRs1BNpXAlCkK80EggAASAQLAIhLo+Bgs4Gm4RADGAIBSKQKiKD8SgULFB5UAACACByiMQ6vpY+YGpeAdBDCo+wGXvXqiKD8Sg7JKF9gEBIAAEwkYg1PUxbNTRehADyEChCISq+EAMChUbVA4EgAAQqDwCoa6PlR+YincQxKDiA1z27oWq+EAMyi5ZaB8QAAJAIGwEQl0fw0YdrQcxgAwUikCoig/EoFCxQeVAAAgAgcojEOr6WPmBqXgHQQwqPsBl716oig/EoOyShfYBASAABMJGINT1MWzU0XoQA8hAoQiEqvhADAoVG1QOBIAAEKg8AqGuj5UfmIp3MJEYPPfnl8xVNy0wl04cZXrt1KMTHLcvfMw8uny1+fpXTjMf7bZ9F6iunnunGdi/r/3+Z/f/xvkc/f7am2+Zi6+9xZw/tsPsuVsvc+XNd5hRJw40++29e5dyqU0Xf/dWc+m5Z3b6nerrs9fHzPAhRxv6/2+fetZcf9EXG+2WfaFCuT6ug9pw9mU3mt+t/qOt85MHfrzL+z71zrl9YZc2Tz/ndNsufDojEKriSyIGmDed5x3mDWY+EAACQCAdAqGuj+l6iafLhkAiMYhrcBwxWPLUM2bR0pXmvNGnePVZEgPaqNPG6pZ7FpkLxpzahXT4EgPaoMsNeRwxoN++PH22mXDa0MYGnvo3a9795oZp4y0B8a2XOuzbby9wKvxQqIoviRhg3mDeVHjaomtAAAi0AYFQ18c2QIMqWohAIjHQJ5+04R877Xqz5267mIFH9DWb3trcxRLw31ve6XTiT++QxWDquBHm8jnzzV/t+peGT9XHDR9iJow8wXzte/PM3YuW2nJ5I84WhwGHHNAJAt8N+qbNW8wfnl3bsCzEEQOqK2pDL7/3rRfEwF9iQ1V8ScQA8+bD+YR54z8f8CQQAAJAgBEIdX3ECIaNQCpi8Nqbm+yp+mUTzzD9DtjXbubpo12J9Gm/Jgb8zopn/mQuuvYnlgj02mnHLq49LquD70aD3Iros+b5l+0JvosYkOsS9eXzQ48xmoRQG2bOvcu6FFH/fVwiQAz8J0Woii8NMcC8wbzxnxF4EggAASDwPgKhro8Yv7ARSEUMnl37Qqe7Arzh18RAb+g1MfjUoQdadx3pPhRFDFzuRGmIwfED+tl7A5NHn2x22bln474EDRvfMaC6+Zko6wTfsfAlBvqOgbSChC0u+bc+VMWXhhhg3vgRA8yb/OcXSgQCQCBcBEJdH8NFHC0nBFIRgweXrOh02dhFDMg3n0/pqRJNDPhkvh3EgAgI1/9/TzneXP/Te+1FakkMYDEobjKEqvjSEAPMGz9iAEtbcfMQNQMBIFA+BEJdH8uHJFqUBoFUxCCvk8+sxEBfVKYO070Gcgdia4SMUMS/7diju3nh5Ve7EAO6WKzvGNz36yfNoCMONrNuu8/iSa5IvvVig+MvgqEqvjTEAPPmw4hjHAEsar5i3vjPGzwJBIBA9REIdX2s/shUu4epiAFBwS43We4Y+BID1x0D3lTQJWaO/kPP8n0F3uhz6FJqN0cd2rVXT3tfQFoMOOIQRyUaduzh5qEn/mCmXHVzp5ClvvVig+M/aUJVfGmIAebN+6Q9ab5i3vjPGzwJBIBA9REIdX2s/shUu4epiAHlMeCoRATL1HHDzXPPv9wlpGhSVKIoYsDuPMtWrUmMSiQtBBTJiD7aj19aDHgIycWJoiNFEQN6RucxOO1zx9qoS7JNTA7i6o3KY0DRlxDCtOtkClXxpSEGmDe7Nyx6mDfVXlDQOyAABPJDINT1MT8EUFIRCCQSg2YblTaPga4nLo9Bs23Ce+VDIFTFl0QMmkUa86ZZ5PAeEAACQKBaCIS6PlZrFOrXm5YRA4LSlYcgCWZtcUh6Hr+Hi0Coiq9VxADzJlxZRsuBABAAAnkiEOr6mCcGKKv9CLSUGLS/O6gxNARCVXytJAahjSHaCwSAABAAAvkjEOr6mD8SKLGdCIAYtBNt1NUFgVAVH4gBhBkIAAEgAARaiUCo62MrMUHZrUcAxKD1GKOGGARCVXwgBhBrIAAEgAAQaCUCoa6PrcQEZbcegaCIgc+F5KhoRAxj1oudrR+O+tUQquJLIgZROS9aOboUcYs+lNCvqA/NPYrIddLA/kZmQ6e2Tbvup/b7Iw7ezzz/8qtmwsgTbAhTjlCWts2kCzgjOUV98v1QWx5dvtpMHTfCXD5nftP1+9bn81xSokddhuy7zpHhU1/Zn9EJMqPa63qGI8tRpnudxT7vfrdzjsv+psEn6zzLAzMprzctWGgojPihB/ax83f8P37WfPXqfzMcRpzkedHSlYjg9wHwoa6PecgNyigOgWCIge+F5DhiQDA3eyG6uCGqds2hKr4yEQMfwtxqKYqbn2k3NT5tzUoMJHHxqa+Vz4AYdEY3zcZXh4AGMXgfS8aw7MTg0omjbHuvmDPfXDhuhCGSjzX6w/kQ6vrYSn2JsluPQCIx4AW4Z4/uhmOQ3zT97MZpDJ8GUlM5lwDnJPjkgR83N9/xkFm3/lUz/ZzTzZrnX7YnivQ95RIgJaDzAsiyZff1ab/Mp8DlPbhkhT2ZpA/Vt8/uvc2MH99pNry+yRx+UB97irlu/WvmlnsWdcm90HqoUUMUAqEqvjhiIGWa5sQ5Z3zOLF62qnGKzhv5c08fZk+t9TzhU39OykfzR84ZjaNeSF3v0Zz54e0L7esPL13ZKfeH7zyMeo6THUbpB7lJoU0ctWHtSxsMJRFki8EuO/fsdPpPz1C+EZqv9KHnuGya18cP6GcTLf5u9R9jcWGcWFfQWAw8oq/NTaItBmztoHdkzhGZ20SOQRzG3PaPdtu+ywaNkrxxnhOqhzdu1D9q3zX//AXzy8XLzNiOIVY/Rn2oTXTySu+ueOZPFivK6j7vl4sbeHTvtl0ni4wkHwseeNye2rKc+VibeExkPXQqT0klWb9zeVI3s/WIx/GNt942jyxdZUjPu+SGZSTO+iWfkTJJ9ZF8scXANU403jt2724eWLLCypFrzKU8UFkXf/dWOyRbjTF79O5l/v3R5Y1xu2nBA13mcpo1z7XWUn1jp11v66V2Duzf1/bRBx+eZ1E6Rs4zKasdg4/qNB9JduTG3bWS6TxAjGmUxYDmMMvw5i3vdiq/DAcdZVmtQ10fy4If2tEcAl7EgDMCkyIihfriK6/bRZsWpZlz72ps8uk3+vBiR//n50ix0aLOiupThx5oFRu/w5sGmb1YdklufrQJVy5s0mKgsyFTeb6Wh+bgxFtpEQhV8aWxGPTaaUdz8bW3mPPHdhjKss3yynOBFnkiyq+9ucluPC4990xD73CWcXKJkPOOFnH+6EWbF2dy0YmarzQP5aaMsxGnmYdy/vN85cOAKNegqBNg3szR83HE4N7Fy+yBAukH2VfCyseViDeGl008o7ERJewkMaC/JRG58uY7zKgTBxrdJ+4HbdppbFwYxxEDqRMZOy0faeYQb8JJt0p9qk+Kpc589Y2Nkf0l2XR9uB4pOyy35P7B6wCNC60XEm+SMW6PzH7tK3NJeOg1iWV8/332cM6hWbfdZ+595EmbTFPONSYrvD4xGaA5SR/uG81JiSnLSpY1j8ufcNrQLnNXzoMkPOTvPM+idIyUgygSy3PZxwWX63HhxnOVXYkkqdE6DGv0hyMY6vqYRkbxbPkQ8CIGvFmhhUOfMsguafOlVBJyEefNO5/88emO3CxI/1CtKOJ8OzUxkMSF2wpTZXkEMVTFl4YY0LxhmaONh9548jyhUWH5JWuXlF2X64w+XdPPyY0NbQRkmXqj6zMP5b0AuRmQFgDt250XMZBS6+tKpPUV/+0iBpJ0xWEu9ZnGOI4Y8FindR+K27BHycm0L48035l7Z+MOha6PT4BpI+9jQSXc9CEQWx3i7jzwezPOH2vbI/svia9L9ydpKv2e/Jvedc0h2qDSh12R4u4sMKmn5+VaGEUMsqx5RJBda21WYuDSMVGySpjowzayUqS5syGxIdzSEAPWg2nrTJKVEH8PdX0MEWu0+UMEvIiBXATjzPxUrDSP84mDXmA1MSBTrvzw6Rd/F0UEpIlYXnbUxEAqPhCD8ol+qIovLTHgUzcy0/NGjEZDX8CVxIDdB3jU2FVPnuxGbXzlZsh1UixPCPkE3Gceys2cJDKtIgbUTunmw66GvsSALxvznYIoYkAbHukSyfpHb4Z5HPT3PhjHneBntRhIHce4xBEDSVTpJJk+SZfWtZxJPSvHgtw56XI34x3XHnYHi9P9SRpL3ynQxMA1h6LcqdgyJV2hqH6ed3KDS25eUcQgy5rHF3LJ757Kl5hnJQbSkid1jIsY8IHDWSM+Y74z966G/3/cePjg5mMxADH4EOVQ18ekeYvfy41AJmJAykouAq4LT3HEQLpYuKBKMi3K0x4Qg3ILnG5dqIovLTFgc/kB++xhdt2lp92IafO7/JssBlGkVuOX1mIQtREgYuA7D9thMdCb+agDAr1Jc0m9j8VAWydddx+4jiSrjMRYu1fyBq1VFgN9Qi83qvo0+t7FT9oukdtUnBsRPeNLDHSUJFd78orok2QxcM0h6cbEG1H6l0kyW89cJ99ZiIFrrmm5ypMYsMUgTsdITHjN3W+vj5mXX30jMUqQJmiwGOSzDoe6PubTe5RSFAK5EQO6QEQnQEcesn+XUIQuYqDvGEh/YG22lO4/ury4OwZ6YUgiGUUNRF3rDVXxpSUGvPlgv2baiPEiTb/xxXjXHQOScZJlvrQvN8vyYmDSHQPXCaHcFCTNw1bcMdB+6YwJ+YKzy0ozdwzkhoX9x6ls6UpEp+Z8Wiz1g77nwYTlvNEnm8lX3ey8YyDvDrh0Yp7EgH3q+S4K9U/79FPbZ8273/rUk+wxLkRAfSI0+RKDpDsG+uSa2kquK3Eyl6QbJZGke2+uOwZyDtHJ9W+fetbOJ/qwJUzfS5C4aTLajMUgbs3zvWuThIf8PU7HkGthlKxK9yoK5uEKCCLr0cTAhZuPxQBr9IfIhro+ppFRPFs+BDIRAyYDHB1kzKnHm4WP/b5LxI84YqCjnGg3IoZMX4CSpn/pSsTfc1QiTQwQ8aBcQhiq4ksiBizXy1ataWzG9OaKn5GRauQiLN3lotyIeCTTRCVyEQPfeeiKXhTnI550x0C68lA/x5w6yEaKoQ2r1DHUX8aHNyL0nSZLWsKli8PUccPNc8+/bDgiFG1UdXQcqYPkGPhEJWICyNHXBg/oZzZt3tzlsCRqUyllhcrxcZdimeJoQVIXyrZPOvMks2zVc40L8NxOGZ0oTjP4EgN2gWEXHh2VSBIDH5mTFmBX+3RUIsJi2LGHWZ941xyicjdt3mIWPbGyS1QlubaQvJAsRl2Sl3OcoklRVKIkVyJpKZSRtuj7OIsBEx4ZPclnDLmNMiqR1DHspke/s6wyMfCRPzkmPrj5EAOs0SAG5dql1K81icSgLJDkdYqAi8dlGdH321FVYhCFst7kNHvhUpdd9oXUhxiUSyrL0xq6j3Licf2doUubbWlerjzN1u/7Hsn2f/7pBXPCpw/zfcXrOR/C4VVQwQ+1Ch/qlk8+iTy6r6MSYY0GMchDrlBG8wgEQwz4BM0ngoYLDp+wa81DiTebQaAOxECb8znyTV7EgBdx+jfpEmkzY5T1HT5J5NNjyiVCrkOcW0RGAkpblzwR1u/qDMxpyy76edow3fPIUnsHIM8PY8ZhMVm30phQToIy4Xjfr580/2ffPRLvQKTFpyrEoJX4sKsV59OQQQA03j7uRlFjxLKIzMfREhzq+ph2PuL5ciEQFDEoF3RoTR4IhKr4klyJ8sAGZQABIAAEgEB9EQh1fazviFWj5yAG1RjHYHsRquIDMQhW5NBwIAAEgEAQCIS6PgYBLhrpRADEAMJRKAKhKj4Qg0LFBpUDASAABCqPQKjrY+UHpuIdLJwYuJIINYN70ZeWZHQMGT1Cfk8RTyg0IkUDoYQ2Ub7WOqLN8tVrGuEUXbjoXA4UFUV+KNrLtyePNrN/dr+NTEIJfugZag8lp+JsvEnxzJsZl7h3QlV8aYlBnnKeNAZpo4kklRf3e5KvtgwlHFVO3CXYvAIOZOkf3gUCQAAIFIVAqOtjUXih3nwQqAwxKMPFYh1dgYdIXzLlzRQRA4pbv3PPHcxXx/5944Idbey+fdMvzBsb3zaXnnumaYYYUN0cdo7boTdhctNWVGSbUBUfiMH7UhVHDHxkKik6jk8Z+ahClAIEgAAQKBcCoa6P5UIRrUmLgBcxkHHA6SScExxRNBH5m45ZLeOzyxN0GUmEvpfRD+LKe+Ott80jS1d1SbgSdbKYtV0yZrkGNSq2OT1DSXIo5rWOOa+JAW3IKbEQJ7Q5/KD9GtlwqRz6/ZVXNzbijlMiGkrCFBdxhvrLz+isnpKgUMIoyu5JkSb0aW4RFpdQFZ8PMWhWzmXMcRlTX8o0jSlHAqHvZ/z4TrPh9U3mkP33Ni+88rqhuOdUzlfHdpjZP/93QzJ2za13W9m8bOIZZubcu6ysuuallH8q/4cfWKAeXrqyId9EWCkBEn2i8o9oeZJRTaheTsJF8dypXRwLXs/zIuQyrSLF80AACACBvBEIdX3MGweU114EEomBK3OoztS65269zNe+N88QGeAFn7pCz9EmRWdYpEQwnAWSiQFlzeTMr1HlUdn6FJzq0KeKMosmJy/ybdfk0SfbxDi0GZEESJ66EwGQ7efnKBlTVLp7V1hKdvnoGHKUTQxHWNGH3HooQQ9txsjtJ617j4sYaNHSxKAIq0uoii+JGERlIPaVcyJ4lLRLzgcaO54bJA862yvPL/pNuhJxJloOTUmywdmXZZk6y6+Uf5lNVs4nmosui4G2nskEWSzjFIaT6uU5I+e8nOdFyGV71TBqAwJAAAh0RSDU9RFjGTYCicRAZ7yUf9+7eJl5dPlqu6Fl6wFlVp06boS5fM5886lDD7Sbf+kuQHBdddMCc+nEUfbUWm5iHlyywqs8DbneOES1mU5IZ5w/1nxn7p2R7aJTeXqGs6i6/LT19/Q3b9jkJkdu5pOIwQVf+Htz5Y9+YUkAfShXw1kjPmOm3/CzpomBvmMQFWdaE4Mi3DZCVXxJxCBKTlju08i5a+MtZZw27i7ZleSCCQXdcaGNt5yXcfL/7NoXOpWv77NEZc/VsqTnpCTamhiw3uBnipDLsNU6Wg8EgEAVEAh1fawC9nXuQyIxkCeTcvNPZICIAbsSMIjkgsAbcE4P77sBoQ2TT3lRJ9+82aHfdJt5kzbtyyMtMXC1a+y06zsVrV2C6Ed9iVT2rVliQCSJUsUP7N/XugNRX8jdJ8r64COszVoMitiAhar4koiBlhNNgH3lXBIDnWCIXfeIGBAhZ4KuLQaSiMtNvZ6XLvkn0iDL9yEGUUSAE52RDLPrkWwDWwx4foIY+Mx2PAMEgEBVEQh1fazqeNSlX4nEIMliIDfkDJo+IU9jMfApL6vFwEUM5ObHJQCtsBgQMaBT2XsXP2k2vbXZEpf999kDxKDEszCJGCRZDOLknE/MeR7R33QnRVoFtMUgD2Lgkn+tA3yIQRzJlPpByjmIQYkFHk0DAkCg7QiAGLQdclRojEkkBr53DMhNgf2S2ZUoagPOiz9vfuidKN/ruPL0yKW9YxDVLu1jTZsf2iixaxHXGeU7nvWOARED9gXn1PBUX7stBkX4coeq+JKIgdzUx92liZJzGnt9h0e6+nTvtp29z8PP5WExiJN/Iq1pLQb6joEkEzJYQNQdA20xKEIusToAASAABIpGINT1sWjcUH82BBKJARXP0VDItWbMqYNsNBN5r4BdEDiSCW9cojbgfDmS4vevW/+qmXTmSeaZtS+YC8eNsHcOZOQVV3m6y2mjEvm0K8qNiOuNikpEbXeFXky6Y0DEgDHjC9auspLixlMbm3UlKiL6S6iKL4kY0DhIOUkj5zKaF98NkbkwSDanjDnFLFj4uL2rozfuTF6pDRyViO/0uFyJ9LyU8h9nMWD3oKSoRLL91C5+Xubs4KhEmhgUIZfZ1CreBgJAAAhkR4DWx6M/styMHDnSTJ48ObbAjRs3mv/6r/8yf/3Xf2169uyZvXKUUFsEvIiBRMd309luRMt8qugiBs1gRJvN//zTC+aETx/WzOud3kEeg+Yh9CEGaUvPU07S1t2K5/O4s5JHGa3oG8oEAkAACLQaASIGFwzZ2dx2223m4IMPtuTg7/7u77pU+9WvftXMmDHD9OvXz6xYscJMmTLFfPvb325181B+RRFIJAby1JMw4AuPdBG5bJ+ynizqzMdx+QiSML3v10+a/7OVY+0yAAAgAElEQVTvHqlDmOpy+SIrMh8nIR79O4iBH25JmY/jSkHmYz+M8RQQAALVREBa1P/zP//TzJw50/Tp06cTQZg+fbp57LHHzM0332x69+5tNmzYYMaMGWOOPvpoM23atGoCg161FIFEYtDS2lF47RGositR7QcXAAABIAAEgEDTCND6OOcrR9v3BwwYYP+dP3++mTt3rtlzzz3NsGHDzHXXXWd+9atfmb59+zbqWblypfnMZz5j7rvvvqbrLvrFbbbZptAmkPWlrh8Qg7qOfEn6DWJQkoFAM4AAEAACQKBUCND6uP/rv7Jt6tGjR6NtmzZtMuvWrTPbbrutefvtt817771n/uIv/qLx+//+7/+aj3zkI+Zv/uZvGt9t3bq1VH3zbUwR7Sb8nnrqKd8mVu45EIPKDWlYHQIxCGu80FogAASAABBoDwJ6ffyP//gP6060Zs0a60501llnmU9/+tPm/PPPNyNGjGg0iqwKV111lfn1r3/dnoZWrBayVhRBSMoCYymIAV8wPPf0YTZjso5KosHSCZwoU/AFY0612ZfxCQuBuhADmfBMRxEqcsSSolzxHYFDD+zTKWO5q82yvLQBAeTzodwvcEUPa/eYyrtgUVnOuT1llUMZLavd2KE+IFBWBHh9fOKJJwyRAkkIuM333HOPjVp09dVX23sFdN/gvPPOsxeWTzzxxLJ2rdTtAjE4YvzWFfOvLmyQmtkA6ORRWS44FtZxVGwRqCMxoNC2ZfnEEYNmIgLp8tIEBJDPNlN3EZiWhRj4kjCdkbsIzKLqBDEoy0igHWVCgMOVvvjii+aSSy6xFoKoz8MPP2zvGjz99NPmE5/4hDnnnHPM3/7t35apK0G1BcTAgxjI3AI0uvJEiuOYU9zzgUf0NTt272bOG32K0XHLXadY+pSQEjeRxWCXnXvaE8qePbqbuxcttUJFZVCm1LMvu9HmUuA8B/TbFXPmN3IhBCWBNW9slYmBPMUdN3xII5GftBhQcjKdB4SJA0eOknPONa80WdaZkX94+0IraQ8vXWk4R8Hy1WvMtOt+ar9PykMgy6f2+5bnu8HXBwRRG13uO837m+94yOZBke125Rehsmb8+E6z4fVN5vCD+jSSx3EuFRlpTeo6+T2NxabNW8yiJ1baeuk3TuRI+okx5QSOrLNcZZM8cGJEsnS66tXqIeo5KUNRUeN85HDd+tcM46FzuLCO13Ii5ZPHoRk5pOR9PAcIF1o/8AECQOD9gzMKV3rSSSeZQYMGAZI2IQBikEAMSNFf/N1bzaXnnmlDZNIi8ejy1V0ys3Lm1CMP2d8qdpnvgBazi679iblh2vguYTblKaGM407EgBaqCacNNZw5lhdSWsSINHDSJpKVNCeTbZItVOOBQFWJQVSGbM7wzcTgvNEnm2/c8DNz/tiOxtwiyEje5Qmq3GzdtGBhY0NJG0KeV/SenBOaGNDGi4h1vwP2tVmTOZGey2KgMxdrYuBbnq9FUBOIqPnM+mHtSxtsRnLKFs66SWdu5izslIhR4kQ6TJ/yMwbkLsXl8QZf4iQztJNuumziGfagQmYol3pPZo3nA43Jo09ujAGNmc5wreuVU4g3+FSvHkedhI7f85FDnalekjJZ7uYt7zb6SqSSM8Nz1nZqFx/osG72lUNYDDyUJR6pHQKhro+hDxSIgYfFQA6yVPT3Ll5m1jz/cuOEh5X72I4h9lSfFsEBhxzQsB7ouwN606CJgSQkst4oYgB3ojCnYqiKLymPgT451Rtr2lRpYsAj6Ep0pr/nvz916IFG3wHQG7KZc++ym2myRshNmIsY6I26bn+a8nxIu5y/mpRoXKi/nAuE27/P7r2NbFNce12baHnowaf4NE60eZ912/th/6Q1lPSZJgZyFsrxou+5LF026VE+bNG/yXtTut0+9wV85FATA9kHl3xEuYzRex2Dj4olqC65ATEIU3+j1a1FINT1sbWotL50EAMPYiBNxjQkbK6WiyV9r4kBufvIj3ZX0Cd3mhi4TkBBDFo/MdpVQ6iKL4kYaF/uKGLAp8XapYVOZiWx1qe/TLjpe3naHWcxkJtSH2KgN6Gu9tPGNak8H2Lgc78gijBJYsDuKIwXu8SQq4ruv9yI8/PSZYa/Y3dFstT02etjlpDIdmhioBNCUjlkqSErh2vzT8SAXbp0vfI+iiYucWPC5fjKIcucdNHs3m07a12SRIzKlYSUCRrLQBIxcMkhiEG7NC7qCQmBUNfHkDCOaiuIQQIx0AuLr8VAmtddQpJkMQAxCH16Jbc/VMWXRAx8TmqJGMgTYd4cTRh5gt2QRVnY5PdxFgO5iSRXmrTEIMlikKa8JGKgLQQu65/ekMq/yWIg2yQlL+qkPerZuM2pPCF3EQN2A+KNdBqLgbS8umZNqywGWg5lPXT4w4QozpLALlSaGPjKIYhBsq7EE/VDINT1MfSRAjFIQQz4BIkGXfvGxt0xkL6x5FqkF5iB/ft2cTmK81XFHYPQp92H7Q9V8SURA72JpY2TvmMwtmOw+e6t9zTuysgNsfRVl/d8FjzwuPOOAfu9s/85z9NmiEHSHQNfYuBzx8A3TCljGqd/2JpCWLIPvA4Pq+9NyU2tdF+U9xTkBtmXGPBFYRk0IemOAd2BkPVK4tjMHQMfOdQubXI85MZeWrLIAuK6Y9CMHIIYVEenoyf5IRDq+pgfAsWUBGKQQAxkFBQyzU8Zc4pZsPDxTpsZMoOTyX3wgH5m0+bNkVGJoqKe0JAnRSWKusTGCxS9T37T9EFUomImUNZaQ1V8ScSAcJFuJZPOPMk8s/YFGzlLblSlG4mMKOOKPhQX7UtGCBtz6iAbuYsv38a5cND8TRuVyLc8n6hEPvcLCE/uO10InvNBlCUZ7UziLSPrRN0pcEUBkt+zGxG587gsBkzClq1aY4Mr6Ag71G4+ceeyqW1yfHRUIlmvnl9x7XZZTHzk0Dc6lpSTqKhE1N5m5JD7hahEWTUq3q8SAqGuj6GPAYiBxx0Dn0GO8jv1fe/Km+8wo04c2CVikc/7vBDRv+zv6vsenisegVAVnw8xKB7dbC3w2dQn1ZDkRpT0vvzddSk7TRlleVZahMrSJrQDCACBciEQ6vpYLhTTtwbEIAMxkKdXBH2zpz1ZNiBZ3k0vLngjbwRCVXx1IAZZSbe0BuYhNyETA30pOSrfQB4YoQwgAASqg0Co62PoIwBikIEYhD74aH/xCISq+OpCDIqXELQACAABIFBPBEJdH0MfLRADEIPQZTjo9oeq+EAMghY7NB4IAAEgUHoEQl0fSw9sQgNBDEAMQpfhoNsfquIDMQha7NB4IAAEgEDpEQh1fSw9sCAGsQhsYzIQA1cG0XYJRdLFRh1LXrYLdxPaNUrx9YSq+JKIQZzslQP51rWC59a5pw8zl8+Z3yUfg65ZYvXam5vMLfcsMheMObVTjgdXa3U41CidULSeah3SrS9Z57FJW2OZwpDGzUlXhmff/rZzvsfJs28uDt9+le05CqPMyR+pba5IXO1ud6vkPNT1sd34510fLAaBEgOfi41JytqVSClvIUN5bgRCVXwgBtFj6pO3II4YUGjQNPNSEnwKYxwVthjEoHkNBGLgh13SWuNXit9TIAY3Gpn93Q+11j4FYtBafNtdOohBAjFghXf4QfuZa26920TFB9+xR3cz75eLbS4DyitAi7sr3jqX17NHd3P3oqV2vGUsclecbikYrozJXB7F2j5+QD97skCx3KldXx3bYb71o1/YYmjQkf+g3VMtur6qE4OoeUNIRMk55TTgDLjyZIySAvpsbvkZOR9pAb3o2p+Ydetf7ZSrgGPNU1t4TnPmXpkngCON6RNV1+Y9KS+Jnvf777NHp3madl7K+lxWwChcWE9R/3U8/mHHHt4p87TeHEedLkfpycsmnmFmzr3L6iAZsU1Hc2P9R9//8IP8DA8vXdkYF2qjzAKvk8/xzIrrJ7X56T+uM48sXdWQA1ceAhlBidqtE/NxpmS9GZIyRTqYslGPnXa9bR6Vwxm9pZ7WIaZ9cdQbcTk/qD7KEB63HsjxZxnYsXt388CSFV3GS+LhyjERt0665jvlr3Dlx6A5zzqA5IeiWG18a7M5f2yHzZVBJ+V63aX8KBLv80af0lC6emwId9f4S00d1z5Xzo+4XCuybNk/lhHZZn5Wlkc4UHI90mv0YYsBJT2lxHqk5+QehZ6J6nvUvCdMdOQzauPF197SwD1qfurcJVF9aHa1DXV9bLa/ZXkPxMCDGNCEG3bcYTZxmU5xT4qIFjZO9EObCnpOZu+k5Dm0OaEEQPSh8iacNtTmHZDP0eTmzKNyk6Inmt4AyAVKLprklsALKv2fM3LK7MtJ7khlEdSqtiNUxedjMYiTc50dluZNx+CjGm40NBe+fdMv7Bw54dOHeZ2iy0y7PB9pEaVNEG0aaJOq/8/Ju0i+eONG/+fEaDxveTNC39PHlXtEzie5yFIm8zg85MaXF23OiB4n+7K+OLLi0lMyQ7LUEYQbfQh/1nlf+vxQQ3opqu+8eZR67d5HnmzoPNZrVKbMruyjT3lcPj/0GJsh3mUt5fHnJGQyV4LOpuzqNxM1qov1sw8xkBtzstzwZmr56jUNsuvS0ySD/EmDo5QZWb8m2GxFkuuBrJPljcer1047NtxVGA8+oU7KSh21TvKmlYhi1DrJCfD0plSOH2E3a979nZLoRa27vq5ErvGXayPjEtU+woXHmLJ1u7K207iwDqHn9Gb/U4ceaOVMZyPX5OTFV15v6CSey0wMpo4b0cllUc6POLmMytzN48NzTRODOD3Chzp5rtWhro95YlBEWSAGHsRALmR6ovBmg10AaHLohUwmPzv0wD6dFkat0B9dvtoqAM4GGuVDqBdGlzLUvsuyHyxsadwWihDQqtcZquLzIQZS3qScE1GWcs0n0pedO8p8+0e3N06nNryxyfzh2bVmbMcQc91P701MAhh3sh3n6sDzh+ctL9ZyrtNmSW6wou4BuCx5tMgSMXDhQRsnTQx85qU8BOjebTsnWdG4SH2hT/95M0akhMfo///VY2bXv+xp3v7vLYb0V1Tf9cZGb4LlJkrOaS0XUfqUD2SYqLgOM3Q/5ZjftGChrZYPWVz9JnIqx0KWITN2k36Ow1HqV94w+bhb+OJI5fsQA4l10h0DiQ+3leRAjomrDN1uvU5GzXci6TQuPhtvn3WX5cS1QZVj7hp/fQjnshhoYsA4a0unb+4R2b8oAsEbdVmeixhoUsH4xn3P817vXfLAPcsaHur6mKXPZXgXxMCDGGgztjStSYXHypQ2MnxBiE8gWMHQwhqn0Kdd99NOchFluo1aYKRZlE3zmhjozQdV5LMBKYOgVrUNoSo+H2IQJ+eSAOsNHG1Glq1cYz592CfM/b950vz9kKPNDbfdZy4cN8K66bk+2t1ILuqyDtpES1cLKk+6ekSdltFizRtSeZou26IXd20xcOHRLDHwuV9A7dO4uIgQ6wPaWJEOIyL05ZEnmF8sfMwMPeYw8+snnzaH9+1jFi1d2dhgc//1hjGOGEhdRe9zsjNNGGUZ3NezRnzGfGfuXZGyoPup5UpvQJkAyn7HbYRdxEATSikTWldH6em4zbsLxzhiwC46cz5wy4paD6IsBnIDKYkBu+dwO7WrCn2vx1/OBbKaxM33KGKgiXQc0ZAY+VgM2FUuavx9iQGt69LFimWYrEXswivHla1Y8jvtUheFaxzRYGJAB4myXt4zsJ6T/aR35EElu7JpneCyGETtd5IIWZa1O9T1MUufy/AuiIEHMZAnffJkRLoXyBOkJItB2pMeLSguUzo9JxWoXDxcZmQQg2KnYaiKLwsxcFkM2M3n6TXrzMuvvmE3pz+//zdmh+7b27+TfFd9icGDS1Z02qj4LIpEDKj8exc/aQVm1IkDjTzZ4wVXutm0mhjIuRunE1zEgF0e5YZQum/c+PP/MPvuuav507pXzD8OPcae7m7avMUMO/Yw69LTzIaWdKY8gY6zJMlNHltj9tvrY05Z0BYD+bc8maZ252Ex0K5KUSezro2q64TYl2BpYkD1yM03j41rPYgiBvSOtKjQ39JyRGuc66MtBnHrpGtc5CZYn8jnSQzYRcwl97KPUk70Jj2KAJLOclnH5PO6rCwWA/Yw4PLlnJp1230Ni4yrXzwf6F+9d/HZ74AYFLuHaEXtIAYexED6BksFHHfCFXfHIO7kUJKQOH9Oac7XyivKp9RFDHDHoBXTyr/MOhKDOJ9jWojobsFeH+ttQ3bSHPvB7b8yZw3/TJfNaBRhlidaLouBJAZ80nbkIft3WRT1Ys2LOV0q1YsxtyXujkEai0HSvPQJUxq1UdAuMHG+1vTbfb9+0t7xoM0UEYVn1r4QeVrvu6GVxIBPNKmdfKfDdSJJz1B7yKIqgzVIGZB3TIi4xG3cXf1mH2s+ZaUy5B0D9hdnP3ySG333jGWK/PLJusSuLS49re8YSDmJsxjoezqMo9wMuu6cRRED7ieVwxZvfcdA4hZ3N0Kuk0l3DKL85/X4UXnyjkEzJ9cS/zi51xvoqPaRRUOPE73HpINJFlsV6H6FJNOaGMj+6QMHvefQdwzOG32y+cYNP7MumHyAwVY9+a6Wy7g7BnxXMg/c/VfDrk+Guj5m6XMZ3gUx8CAGpAA4moh07Yk7iUuKSnTpxFHWLUKXIc2LrggQelOgoxvwwsnfk6BRVKLZP/93w/XSd67oHmUQzLq0IVTFl8ViwPdn2EWBzfD0vTZzx/lFN0sMeOPDEbvGnHq8WfjY742+xBd1iqdPmqPawIuyr8WAF2wqyzcqke/9AiozTk/xaSG7nUiXhzh/d91vX2LAQRUoYg65TkwZc4pZsPBxq5fifPipviRZoH7KqClSrqLGTbr16H5zhJdJZ57UiQzxO6SbBw/oZzZt3tzplF3jyPqcXNXYxZTkjj5RBMcXR77wSkSJcBxz6iAbTUi7lch65HqQFJXIhUeUu4scG45Cptcuua7JcZFr12mfO9ZsemtzI++H/I3GYdmq5zpFJYqKDiXx1hZGJpbycnqU3EvZjmsfl0fPR+mwuOhTkujS/6eOG27Hj114ZBt0VCKKxkRWO/owQaIDFNanGnuXnPvKf1bcs6zloa6PWfpchndBDDyJgdxQl2Hg4lwHfNsHNyJfpFr3XKiKL4kYtA6x4kp2mfv1Iu6KWOTbcszLaKTi/MejCJAv3niu/AhoN7HytxgtzAOBUNfHPPpeZBkgBoESAxKaJHeDOMGSFxfjfEeLFM461B2q4iuCGMjLflo25IldK+SG6+ZwnK2aW5iX0chKlx7XBXRtGWmFHKDM9iCgLe4uS0V7WoNaikIg1PWxKLyarXfGjBnmkksuMVdccYWZNGmSzXW1detWc80115gLL7zQfPOb3zRTpkxptvjg3tvGZMh8HFxv0eDSIRCq4iuCGJRu8NAgIAAEgAAQaBkCoa6PLQOkRQVv3LjR9O7d23zkIx8xO+ywg9mwYYP9++233zbvvfee/btnz54tqr18xYIYlG9MatWiUBUfiEGtxBSdBQJAAAi0HYFQ18e2A5VDhVOnTjUzZ84077zzTqO07bff3kyePNlcfvnlOdQQThEgBuGMVSVbGqriAzGopDiiU0AACACB0iAQ6vpYGgBTNISsBrvttpvZsmVL461u3bqZ9evX18paQJ3PlRjExZLWCc9SjFemR5MuMsZdqNTRjzI1BC9HIhCq4muWGPhc4I0Cqtn3fC8tJkUbihPfpLs+cdF00t4p8AlT6pPYsKrTMcs4+mAisdVRlPj9qARSSWW3ut1Uf6vlwrX+6b63o69JeOP3MBAIdX0MA92urZRWg7paCypPDHw2HUkbLp8yQp0EZWh3qIoPxOB96fGJDpYUZjOJvEs59cl23OoNYBnmTVFtSBpLalczxKAd/Wm1XIAYtGMU61VHqOtjqKMkrQZ1tRZ4EwOdulzGn+bfKGrCwCP62jjMFF+ZPl/73jxDsYwpYgoluqGEN5zq3Va+zTY2bjklH+O42TIGsY7MwPGXXd9HncxQ5kpOrCLjFlNcbc5yyDHFr/nnL5ibFjxg3njrbfPI0lWNONtJJ6KhToIytDtUxedDDJLk7YZp4w0niuLY7iSXFH+cY+jTGG01xuzRu5f590eX25jt9J5OAqQ3zzyfqDxO3ETRbGTscS5r+eo1NnEWfXiOueKuy3r06X3UvDx+QD+bLIpzJlA+kW/96Be2GJ7/9H9OSuiKuMP1SiISRdpl3HXSJZy/hOPLa/x8+kl1y4hQMgKU6315KhyVzyFte6g8yrq86ImVZt36VzvFjZdjIuvV8qfj2utNvjwkkfkWqHzuMyXr4sRW2mLA9f1t/762SZ/99KE24ZVLNuVYcLsPPbCPLZ/z5tC7cr1JOw5cd6vkwmf9k/jdu3hZl7nmiqdfBv2MNhSLQKjrY7GoZaudrAYUpYiiENXtbgEjl+hKFJXkh1PPR2V0pIKJGFD2SVfGRM5YSRt2nQGRlCS/R0qUs2bKhEaUuTXqe7mp0MnLZCg/aiPHWqeNGadw58WQsx5GbUayiRze1giEqviSiEEaeeMss3KuEU56nrCcxpECnk+UKIizkMrMtTPn3mXJOM0VV2Zc2Q7XnODNssxArrPUujKQy36xPPiSb/mcy9KgT4Zl5napX0h/cab1uH5qq6LcxOrsu6w74ogBvTPsuMO6ZAxOypDL48gHKTqbLGHJ9VJ2ak7+JPWdlB1JWEgXS5kl/U0fIhNSR8uMt5IYkK7WmWR1aNuo+jgZnsaU35XjRYnw5ByQ7/iMY95yIbP6csZouf5F4ccZjfvs9TE7P10ZkmWWYKwa9UUg1PUx5BEjq8EXvvAF86Mf/ah2dwu8iYEeYLl4ULZBmZqdf9MZVPWJGStxTl8uNytSeUsCINuRlOgnatPiivEddUrGGzWuE+5ErZvmoSq+NMRA5smIc12TvxHicp4kubxJWeUTXdr8x7l+yHkkN7LaJcI1d7QbkWte6g2Z7Be328edyDfbcVx9si+0meVDDs5GLfWZPBjw+V7e57hpwULDm784/SfH9dU3NnbRp6wbqTzeaOoNtpydUcQgLk+LixzKMl0ZrCUxICLB/ZUERVoMfNpNFgO9PsjstlHj4CuvrZYL1zzR/eYxGnbs4daqLtcb3D9o3VoTYsllWR9pvcOn2ghsXfL9RgcTLQas5Dl1Ov3NZmXXwnre6JPN5Ktutq5DdPLhWlho46LdlKh8bWLmuqVJWZpf5fdxi7l0o2CXiShioNOygxi0bkKURfGl7WESMaDykuSNifHYadc3qmfZpy/kBt+XGOjLxnIz1L3bdg33Pq6QXZc0MWDXIn5OuvjJzTxb7vi7qHnp49vtQwx87hfwoQBjR6frEkdNDHz7KQmE7L/8XvYzjhi4xpVcutKWp0+WXePI+k7LOePx/40+2Uy/4Wfm/LEd1k1NuuzwO6RnoywGfBDks8HVLj3SekXEgl2JLp04ylq14oicHAefcYyTwzzkQpbB1h9y+5L4SYsBEwNyZZUfnpNpdRKerx4CZVkffda76qFfnx6RnKUiBnqjkYfFwLVQx51suTZGru/jNvOSqOy/zx5dXIlADNo3Icqi+NL2OI2idMkb3y9gAq0tBs0QA20h0BY4ufGMsxjoDX8UPnEXj1190Rt1ubmj/2ufdFmvJA9xdafZAPr2M4vFQLviyBNx6baVZDGIskDEEQPGLu6knsepY8hR5vEVz5gLxpxqX5Mn2XlZDFykhF1Pk4iBtlDHEdQoeW21XGiLOROlJIuBXm/S6iI8X10EyrI+plnvqjsa1e1ZJmLAJ44EDylz8vnkMKTax1JaE0ih06moPnGiEyF9x4AWf/ZVladu0o3A9X3cHQO5CZKXJqPuGGhF7RN5pboi09qelUXxpe1lkqL0kTdNDOidWfPut5eL6dMMMdARYegEmX3TyTWPiQHP3SMP2d/6kcuTZn2vSPp5S/KuybcsQ85XSQZcxCDpjoFPmFIeQ98NoLxjQKfkcf2Um3l2v+kYfFTjHgjrP3nHIO6OFfvQu3zMo8pLQwwICyY9SWGX2cqjgzvwxpatui6LAa0FpOPZ7YnvQOg7BvoQR+pVffk4ymIQNV7UTxoHPT6MvZZXH0uSr1zItUuuf9qCIvGLu2PAc5LvB6XVSXi+egiUZX1MWu+qh3y9eqTlLNGVSEYaITeHKWNOMQsWPm6k4mZXiKnjhpvnnn+506kTRyXasUd3M+zYwzqZonkjL83W0o1Imp1pmNhlyPW9Hkq52UiKcLRs1RrDUYk0MUjatNRLhPLtbVkUX9peJSlKH3kjAiAjAtH8oeg9JH/SZYPmCZdHcpomKtGkM08yz6x9wVw4boTtoowQNObU483Cx35vST5HS4mKShTlRkRl6Q1n0nyldzhKEOsP+k4HCogaC9/7BVwe9TOqPu0HLl0ZXf2kclzRh1zfSyxO+9yxNlqbHFeOuqPrdJXnuszsshhoNxWXKxH37aJrf9JJrmQ7yLWFPvpE3xWViPo0eEA/s+suPbtYgFx4+xADvgfC640rOpRrHHlMWiUXcv3jwzCqS+LHF47J9YnHRLrfwY0orSau9vNlWR+T1rtqj0L1e5eaGIQMSR53A/IoI2QMW932sii+tP2EonwfsTysaT73C9KOT1mfj7sMXtY2o11AAAgUg0BZ1kesd8WMf7tqrRUxIFCzbDqSTPDtGrQq11MWxZcW46IUpbZEyHb75DhI20+f57NY1OpGvEEMfCQKzwABIEAIlGV9LGq9q7oUaDfDZvtL+wKKDDe2Y4gN2pD2UztikBYgPN9eBMqi+NL2GooyLWJ4HggAASAABNIgUJb1EetdmlHzfzYvYiDvEoIY+OOPJ0uKQFkUX1p4oCjTIobngQAQAAJAIA0CZVkfm13v5J0ibdGWd2ui7gvxnVIdlIMSRA48oq8zC3xSuey9oZAAACAASURBVHQX6OY7HrIZ5OPuXvE4cR/o/g+Hzue7QHw/lttz+EF97J09Ha5YhtSXYcy5TPpdB/DRAUKoPVF4ynuKzXoNwGKQZlbi2ZYjUBbFl7ajzSrKtPXgeSAABIAAEKgnAmVZH5tZ75Iyg3O0PBpZChbBEfL0KXoUMeANOF/ylxf5k8qVuXv42bhTdkkMKIIfb+w5ShsRFW4PBUjgfsvIczLaID3Pme+ZxPgQA188C7UY+CQtSprKPv7FcZkg87j4mNRG/N5aBMqi+NL2shlFmbYOHcrX932dCTbqPd+Eaa53r5gz30Y6cimhuOzkaeet1BPUnitvvsOMOnGgTcTFH1mfjjzki1uIz7XjzoLEU2c4lqdqMoN9Vixd4+mqv5Xjr0MAu/qWZSxYxs89fZi5fM58G8FKR5zywbSZzMlSX7jw9ambnsly38i3jjo9V5b1sZn1Toae16HkJRGg8ZSbf8rnwuHtaQ5EEQPedPPcpDI4sS4TDFe5TCJkuXIt0fKliQqvy1QPh6yWoZl1ufJ5iuzGJEEmOPUhBi48ed75kBzX3GmJxaAZheh7sTdJ0UERha0my6L40qLYjKJMW0dZiYHPnIsjBmk3EDKAgOswoZUbw7TjVrXnfYiWToSZFQOX/LjWg1aOf6uJge9a6INp0noZVYbPQYJP3fSMT+hh37LwXNiXj11+7/pEPS9iMLZjsJn0rR8Z1yY9jnDkSQx0v/MiBnH3CAq5YxCVxp4TFX31C8PNRd+9xcZfp/jRf3PAvuag/fZqxK92RQXSp4ZR8a0pIROnuieWt8/uvc2MH99pNry+qWG6IV+uW+5ZZHMnxGVOhpIpJwJVJgYyP4eMrU6y/sPbF9oBeXjpSiP9AmXUIfK7XPvSBsOZkeNGkOcPlUX+jhQ3XyYhpPlJHzKjThh5gs1sSzlGZN3SN9Ple6k35lE5SFgBc30D+/fNNG8lEYmyNuiY+1QfJUmk3CnzfrnY6qXrL/piw7rh009erKT+4azMUe/rwxG5meaMvWnaw+VxvgNqj/RTZVmQ9XICShpX+kSNod446g015wiQ9bksBnLsSa6iTqyScImaB1Hyw5mno060WzX+PM5/27+vxfOznz7Urms6Xwf1naKByBwhJG96PKLGj8qVMi2zJNNvLj3hkk8eX85jwZaHqORyNNZaXzC+/D4l7NN+1bJufn/H7t1sokT+jf6Ny2JeztWofK0qy/rYzEGY60RebpSlzPBJeqstBlF3F/IkBkkWA1pTuA3SGpF0xyDOwtF2YqDTufNCQiaUqCySMukNTbMosz99Lxd7rbQkmZALGYGoE/HkedpSPrVQ/RaVRfGlRTpJUerTfpLjqEy4vhlz49wKXL6HOgOrzGgsM36TUnRl4NX1yrmpdYPc4OgNZ7PzVp9AuqwVURtcaZ7ljMS+/ZSbYboEdvG1t5jzx3bYhHScmZ0z/F428YwuCek0MaBNmPSDJXljf1VXeeSLyqdfUn5c2XzpIIWzHbtObmW7qA1EEDnxmsweLHFiYkNEkzeOxw/oZzfCnKU37nRQZvCOwsU1TtwXH1emvMdfmu1lJmfeMHNWaDmnCE/ZV8KExzlq7WKdI2VaEwN2qdB6wiWfCx543CaiiyMGVC/JFsmt3oxoYkDPcmZrnsP0HcsKZ2+X7hs+LsJp9W1dny/L+pi03kWNj7YMSB1x04KFjYMEele6Fsn3WI45sSfL7q69etrDHtpvat2adMegWWLAF6Q5Gai8YyCtFM3eMeCs6Nx+7pu+6CzXBH5W4tmWOwYuNyHXHQMaOPY/JoUadZoflTGVF17N3DQxiPJj9XFtqKtiKXu/y6L40uKUpCi1a4WcLzThpRzzpoZP8vmUT2+8XW3Urh4u1w9JwCUx2HO3XnaDyJsdJu60wdCnfq5NTBSBkBu7ZuetJBtxbgp6YxiHr28/Xf2X3/PmTx6UkGLWG2DZHikLpNBd5cmNumtMZVmSGLhkJW4jK9+JsnhIYnDogX06bYJ91okkXOQYulyDfO8Y5D3+LhcdOackMeDNDlv7XHNZr4WaGET1gwilqz0+FgM6kWULDGeT1hYZTYBkP4kc89ymfmq3L7gTpV1N3M+XZX1MWu/i1ia2QsZFJdIZv9laxxapRU+stJnZmRh8/rPHmAeWrLCeKq536Vn5m74rkPaOQVQ0oyi3KKo3ypLOe1tt4SSLHJMV+R57DEjSLd+VeMrvXZbJOKlMfcfA5TvqIgak/HnzQG4Q9NGbi6iLjxoQWoRIaWliIBVa1IlLflMSJbUDgbIovrR9TVKUcqJy2TyR9cLMCyu7IyRtJnRbtY9w1KmsbkMUMWAXFH5WK9yozU2UmyHpAB+fbx9C73O/QG9O9CZaEy/ffkoCQXVE+ZpHWVCjNsBSb7HunPblkeY7c+/sRMh8ynNZDFj3sutH1AIhx5D0s9zgSRcp6i+fkEVZDMit00V25IlVkiuRxCVPYhBVrnShk3PIJecuAqnnNs/rKGLALnxcn3bv0muhJgZx/dDyKQl9nMWANvaPLl9tLQFJxCDKFYmsEvSRbiBSjkAM0q4m1SUG+SHxfkmujXje9cjyNKFoZV1FlZ2aGPicBPF9g0snjrJ+vATkvYuftH3U0UN4gXW5GOlFHsSgKFFpT71VJgZRJJZQTdq45mUx4CgNTDSSLAY+kVDiNvOyX2RujXIFkZvaJGKgT1Nd95XSEgPffmaxGGhXHLmJloctcRYDlwtOHDHgWRsXdYo34PQs3ccga48+AGqVxcDlokR9ahcxaGb8eR1iF6qoOaWJgcsKLjVrksXARZzSWgykpSiNxSCKGMBi0J61kWopy/qYdBCmEbn/N0+ayVfenDtQd1x7oXWDY3cawqeVHzoc6Rh8pBn/je/bU315B6uV9WYpe+YFY8zQYw5LVURqYqB9pVmx0w3w7956jyEyoIkBv0OnSnwqoVspNwWafMTdMdCbLdwxSDX+pXu4LIovLTBJijJq3rAvubyHE7Uh4tM8jtGcZBqUdUmfYU0MpOk07o4BJ1Zh/3GJjZybURcaFy1d2fCdjyMGPvNW+ipTG3wPE1zEi336Gd+kfurnaDNIp+xxdwKi/Lb1OErf8yhf9qQ7Cy5iIElG3Kktn7qxny4f5jB54UuzhDn7lye5msTdMXD5s0tLRDuIQZrxl0RJ3jHQxEDOKUkM2HpD31G98h6QdruLu2MQZ1GJkk+af/KOgbxbw5c7+V4AyXPSHYMoYkB9wh2DtCtGc8+XZX1MWu+a6x3eKgsCqYkBNVy6+XCED0kG6Bm6PEIfjv7hOtFgIKTvMH1HCpYjgMgsePw9RyXSxAAXncoiWs21oyyKL23rfRSly88wbuOqoxJRJJthxx6WGNNcujeQL+Rzz79sI3XxRSnqH31Prg202POGgC91kf+jdCXR7hWMj55vsl4deYlOV6gcjhIkDwl85q3v/QJqG7cjqj7tA+3TTypTPiddQFzfs64i15Ixpw6yWPPGWkaXkfrNVU+cC46LGLAeZvcVF6FkGeNNI70n5Y7aP2XMKWbBwsftwY8kstLHX8r3pDNPMs+sfSEyt0UcLq6Nr2s84/IouOQt6/iTXA8e0M/suktP6xYr1yo5p/bfZ49O66BPlCiWXSbUPq5E7MITJYdy3dXjs2zVc/YCvYyfznqB9YW+fOyKasQYMDabNm9GVKK0i4jH82VZH33WO4/u4JGSItAUMUjbF5/kST4nhj71Jrkj+JSBZ4pDoCyKLy0CdVaUecy5PMpIO2ZFPe+6OFxUe1BvuRDIay0solf63g3uF+Q7CmVZH+u83uU7ouUsreXEIM3lEJ9TwzgYtdWhnJCjVXEIlEXxpR2ldipKefKn26lPn9P2o5nnsy7+dZu3IAbNSFm93sm6FrYTLX35WloX60T424F5WdbHdq537cAVdXRGoOXEAIADgTQIlEXxpWkzPQtFmRYxPA8EgAAQAAJpECjL+oj1Ls2ohfcsiEF4Y1bpFpdF8aUFGYoyLWJ4HggAASAABNIgUJb1EetdmlEL71kQg/DGrNItLoviSwsyFGVaxPA8EAACQAAIpEGgLOsj1rs0oxbes4UTg6RoRQRpVCIl7deoE8X4DIUOIenzDj/DYRp1xk9XGTrBE0edSFNnHZ4ti+JLizUUZVrE8DwQAAJAAAikQaAs6+ONN74fdRKf6iFw8MEHm+Mm/dg8OONM27lBgwaZbcwR47eumH9123rbDDEgUnDRtT+xKbEp1FpUBlKfDjRLDJq5GBYVIo+TCfm0tS7PlEXxpcUbxCAtYngeCAABIAAE0iBQlvVx8eLFhjaQ+FQTgV0+e2E6YsDxtA8/aD9zza13G07/Tht0+siTfBkhJSq++YNLVjRyFfCJv4wJzWXvuVsv87XvzesUb12nf4+K881Z6WQ7dFx4SlDEWSuj8jNQYhr9cSVjo/jeMj45t58yS3JbOGJDM+SimiLYuVdlUXxpsQYxSIsYngcCQAAIAIE0CJRlfcR6l2bUwns2tSsRb56HHXdYI6MpZ1xct/41m55aZvukpDljO4YYmQ4+LpMxZ9uUmSInjDyhQQx22blnI8sikxENu8wqyYmbOHkPbepffOX1RqIh2rBT4h9ORsMkQT4nEwjp0IySkBAx4PJ0vdpiEHKs6laKeVkUX9o+QlGmRQzPAwEgAASAQBoEyrI+Yr1LM2rhPdsUMeD057Qxl8nL6GRcZq7kNPKXnTvKfPtHtzeyLEqY4lyJeDOticFVNy2wGTijTvPZahHVjhnnjzXfmXuntTxQGnqdVVKSEm2B4Dbrk35NDGQZkgxoYkDlIcZz1wlTFsWXdipDUaZFDM8DASAABIBAGgTKsj5ivUszauE92xQxkBtzSQyWr15j2HpAp+xy0/zam5usNWHd+leNdO2RxEC6+TCU5HqT1mJAm/Codnz1C8PNRd+9peE6pIkBu/tw3dpNKop0aGIgCQmIQfoJURbFl7blUJRpEcPzQAAIAAEgkAaBsqyPWO/SjFp4zzZFDKTFgDbG/LfLYnD9RV/sdLovN8ySGOgNfZTFgF109B0D6eJDLj3NWAzkO66hTLIYgBhkmwRlUXxpewFFmRYxPA8EgAAQAAJpECjL+oj1Ls2ohfdsU8SATv4nnDbUDB9ytJGbedcdg47BRxlpZXDdMZBlbd7yrjn7shvNkYfs38liQC5ArqhEfI8g7o6BrGPFM39q3AnQdwzoOdrka1KTdMfAlxjgjkH0ZCmL4ks7laEo0yKG54EAEAACQCANAmVZH7HepRm18J5tihjQJr9nj+7m7kVLzScP/HinzbMrKpGMNiRdifh7ikp0/IB+lgz8bvUfbbljTj3eLHzs92bquBHm8jnzG3cDCGadx4Cj/fAQuNqhoxLt2KO7GXbsYfbOgYxKFOVGxGXHRSVyEQNuD6ISxU+Ssii+tFMZijItYngeCAABIAAE0iBQlvUR612aUQvv2aaJQdzl3/BgSNfiPEKN4uIxLAbppA5PAwEgAASAQJ0RADGo8+i3r+8gBk1iLd2h0hZB1gNkPgYxSCs3eB4IAAEgAATqiwCIQWvH3hWNkmrVd2B1S/iwl75Puq8qg/ZQdM+yHRSnJgatHRaUXncEyqL40o4DTKtpEcPzQAAIAAEgkAaBsqyPdVzv4ohB2sNeTQz03dU0MtGKZ0EMWoEqymwagbIovrQdqKOiTIsRngcCQAAIAIHmESjL+uiz3tFmV94Z5UAu8i6nvKNKm+uoO5ocrp6evfmOh2zIe7qTSsFv6BNX3g9vX2ifeXjpShN3b5RHRFsM+G4ovTvwiL5m01ubbXJcmfRWB5KR/Zh1231m0+YtZtETKxuh+vnOLN3RlW3K4oXSvERFvwlikDeiKC8TAmVRfGk74aMo05aJ54EAEAACQAAIMAJlWR+T1juZI4oCu3Do+bEdQyxZoCSztLEnF5oXX3ndbrYpSmQcMVj70gYb6IZyYnGI/F477RhbHuWmumn62YbD3HPkSpdEReXeumziGY336T1NDPSdU00MfvvUs412U0RPKo+iYF587S2dkv7mcXc1r5lSWmKgTS1ZOpzkvxXnV1amwcqCQSjvlkXxpcUrSVGmLQ/PAwEgAASAABCQCJRlfUxa71x7Kv19XB4sncdK5q7i/FeHHtinUyh8Xd7MuXc1ombK/Fk+xCAqH1bU3QHtRqSJAdV13uhTjCRLUcSgTO5ElScGPr5fccSABrVMJp6qq8myKL60OCcpyrTl4XkgAASAABAAAiESA9p3yU0590F/Lw+AdYLcqAS3ZH2gDxODfXbv3aken/Jok+5DDB5cssI8unx1w0KgXZ24DE04NDHos9fHrHWk8sQgb98xMu/M+cAXjOL8s08Z+2Bd889fMDcteMC88dbb5pGlq8yF4/7e/H71nxoD5jrV175fMocBDWpU7oSvju0w3/rRL+yYb7PNNpZt0ueKOfPNheNGdMrgDJWVPwIhE4P80UCJQAAIAAEgAAQ+RODBGWfaPwYNGlQYLEkHYXlYDGjzTx/eD7LFgPdx9HeSxcCVV8qHGORlMagFMWiF7xgNEvuYXXTtT8wN08Yb8h1jH6w9d+tlvva9eYb9w7SbketEXxMGyeyk2YZ81jgzM/2f/cCYnTJDHdi/r02Ehk/rEAiVGDAiDz30UOvAQclAoCIIvP322+aKK64wF154odlhhx0q0it0Awi0D4EyEwM+PJ48+mS7Z+KIPueNPtlMvupm5x0Duf+juwhHHrJ/gxjwPnHd+te87xhkIQZUH7WB+sB3FLgN8vJx0h0DX2JQJrf11K5EeTBBl++Y3PBHEQPtY0YbdRqwK2++w4w6caCheLDyo92IXD5m+sIJX2yR5cGdqD0KD8SgPTijFiBQJAKzZ882t912mxk5cqQZP358kU1B3UAgSATKTAwIUFe0INf3fABL3iMUgWjwgH5m0+bNDWIgPUvoQjEf0qaNcuTrStRrpx6GoxJR26aOG26ee/5lc8GYU1NFJYoiBkw0lq1aYw/Caa9Zpj1mamLQSt+xJGJAN9lZGHjT3zH4KHPLPYu6DBYNZBQRIPMUuy2xcGliwNYDEgz+lGnQgtRino0GMfAECo8BgUARIGtBR0eHeffdd812221nFixYAKtBoGOJZheHQNmJQV7IaC+VvMrNsxyfu6xx9ZXp4jG1MzUxyMNioH3HeMOfhhgwkAfss4fZdZeejZi2cRYD+Zusi76XrkQgBnlOmXRlhU4M0vUWTwOB+iEwdepUM3PmTPPOO++Y7bff3kyePNlcfvnl9QMCPQYCgSKQdMcgz27lSQz0PVPZTp88B3H9Sop+2ap388Say0pNDFrhO9YMMaAO0EDc+8iTDVOMBkj7bPFNdrohHnfHIIoYZBn0VgxcVcsEMajqyKJfQMCYjRs3mt12281s2bKlAUe3bt3M+vXrTc+ePQEREAACASDQTmIQAByVa2JqYkAI5O07FkUM+MIx+WBxVCLpSkTtcIWP4lHSUYlkNCV6hl2J+Hv6jqISzf75v5tLJ45qRCAqm5mnclIoOgRiUOXRRd/qjoC0FjAWsBrUXSrQ/9AQADEIbcTStbcpYpCuitY9LS0Arlqy+n5Rubhf0Lox1CWDGLQPa9QEBNqJAFkLevfubbbddlvTo0cPs2HDBvs33Tl477337N+wGrRzRFAXEGgOARCD5nAL5a0giQH7iRHIOj11FPBZ3IDKFEIqFKHK0k4Qgyzo4V0gUF4EZsyYYS655BIbpnTSpEk2T8zWrVvNNddcY8OWfvOb3zRTpkwpbwfQMiAABCwCRRAD7caediiSEtlyeS5PlKT6ffaKcYfZeRxip8XE9XyQxCCvzqOc8iEAYlC+MUGLgEArEGBi0IqyUSYQAAKtQ6DKxMCFWhwx0G7rrjKSvFyyHGLnOdogBnmiibIyIwBikBlCFAAEgkAAxCCIYUIjgUAXBJKIAZ/OH37QfuaaW+82MuIPnYzP+PGdZsPrm8zhB/VpJLcdO+16W89JA/s3PEFkFCH6fu1LG2zCsV127tmIJMn5BmQyM3kPlt7jxGq/W/1HmyPh+ou+2LhDqjsnLQb0GyXXvXvRUtsurl8nutWn/TL/Adf34JIVZtp1P7XVTT/ndLPP7r274EDJ21zh99sphiAG7UQbdSUiAGKQCBEeAAKVQADEoBLDiE7UEAEfYvDl6bPNsOMOM5RQjDMfk+v3imf+ZDjDMSX24k38ZRPPaGQYpmRm9B6doL/4yuudyAMFjYkjBpu3vGsuvvYWc/7YDps4jE/pDz2wTycy4Ro2SQxm3XZfZP2aGMiTfhkKXycukxYDqkfiQO3xtTy0WuRKQQxYaKaOG2EunzPfpsvef589ugwugaGz1vn6jUkg9cBzZroosPnisa9Q6YFftHRllza3elBDLh/EIOTRQ9uBgD8CIAb+WOFJIFAmBHyIwcXfvdVceu6ZdnMuN8uvvrHRzJx7V+PUXvv0cxLdGeePNd+Ze6fdD9JGXOYziCMGRDyk9YBx890rcnvkflTXL4lBVPRLSUzkuOn9ocSBnyuDO1GpiIG8SKxZFydFi0tn3czEifP58rlMouvU5ZVhkJvBpah3QAyKQh71AoH2IgBi0F68URsQyAsBH2Ig80FpYiA37tKa8NFu21sLAr371S8MNxd99xbrOpSGGNy7eJl5dPnqLoFp0hIDdj+Kql8SA71XJYy1KxPvbTUxiCIwZdgzNkUMZD4A9p967c1NhhgifUjhkw8XfUfmpHXrX+3i18U+WOR7NvCIvmbTW5sNM7SOwUeaBQ/81vp1sW/aggcet2XHWQyeXfuCZYo79uhu5v1ysa2TBpXMNdQG8uui5GbMCI84eD/zje//3JbLv2l2N7B/XyuUUqionh/evtA++vDSlY02Ll+9ppMPGdXVDLnIa/KGWA6IQYijhjYDgfQIgBikxwxvAIEyIOBDDKTFgPZB/DdZDOSGOA+LgXZVKtJioMeH2rbm+ZcbrlHsoeKKfhQkMdDpqbnTHYOPsiSA/MRoI61vcEtfMbpgwc/2O2Bfe7mDPlldiWjDThdYyAeNy6XLIkRS6Dc22zCBIBZHPmRRrkQ6qZkmBroe6RMnyyuLz1gZlIlPG0AMfFDCM0AgfARADMIfQ/Sgngj4EAPa4004bag9jI3buMfdMdDv8b6LXImi9pC0p6P9pSQl7G1Ce1RpxXCNnNywS+sDuShx/XF3DLRlQubBSrIYlGW/mNpi4DLHSEZIPmXsJ8a3v/XGOoox5kEMpM+WHARX/S5ioE/69fuyHhcjZMErAwMMRX2BGIQyUmgnEMiGAIhBNvzwNhAoCgEfYkCb8J49ulvPDxkJKOqkXEbxiYtKRN4gw449zB4+076LovyQV8mYUwcZijjELjtR5dGl5LMvu9FC1mxUIlm/xF5HJeK20TOyP/w9RyXSlo2yeJikJgZ6w8/gaMIgB4afYbcgcrmRPmBRlz2auXwsLQHkq5aFGGjhjSM2IAb5qScQg/ywRElAoMwIgBiUeXTQNiDgRsCXGFw6cZQzLGiV8M3rpL8sh8ipiUGcxUCaaVz+UyQMUT5lxJzysBhIBpaFGCRZDPTlmSgfMlgM0k99EIP0mOENIBAiAiAGIY4a2gwEkjMf+170LQJLeTFY1y9P99O2Letpf9CZj/XdAfYBG9sx2Hz31nsMM8So52gzTSYc+pBJhy4G533HIC9ikHTHwJcY5MUk0wppqM+DGIQ6cmg3EEiHAIhBOrzwNBAoCwJJFoOytBPtaA6B1BYDqkYyLhmVSF/skM/JzHdsNeBMd1PHDTfPPf+yOff0YY08BkwYlq1aY26YNt5QVKI5H0QC4q6OGz7EyAslzbgS0eUS8lNLG5XIRQykDxmiEqUXShCD9JjhDSAQIgIgBiGOGtoMBJItBsAobASaIgZhd9m/9VlNQ1RTWXzG/Htd7JMgBsXij9qBQLsQADFoF9KoBwjkiwAsBvniWbbSQAwSRkSGmko7eGXyGUvb9qKeBzEoCnnUCwTaiwCIQXvxRm1AIC8EQAzyQrKc5YAYlHNcatsqEIPaDj06XjMEQAxqNuDobmUQADGozFBGdgTEoNrjG1zvQAyCGzI0GAg0hQCIQVOw4SUgUDgCIAaFD0FLGwBi0FJ4UXhaBEAM0iKG54FAmAiAGIQ5bmg1EAAxqLYMgBhUe3yD6x2IQXBDhgYDgaYQADFoCja8BAQKRwDEoPAhaGkDQAxaCi8KT4sAiEFaxPA8EAgTARCDMMcNrQYCIAbVlgEQg2qPb3C9AzEIbsjQYCDQFAIgBk3BhpeAQOEIgBgUPgQtbQCIQUvhReFpEQAxSIsYngcCYSIAYhDmuKHVQADEoNoyAGJQ7fENrncgBsENGRoMBJpCAMSgKdjwEhAoHAEiBvhUG4EHZ5xpOzho0CCzjTli/NYV86+udo/Ru9IiAGJQ2qFBw4BArgiAGOQKJwoDAm1H4KGHHmp7ne2u8O233zZXXHGFufDCC80OO+zQ7uoLrw/EoPAhQANADCADQKAeCIAY1GOc0cvqIlAHYjB79mxz2223mZEjR5rx48dXdzAdPQMxqN2Ql6/DIAblGxO0CAi0AgEQg1agijKBQPsQqDoxIGtBR0eHeffdd812221nFixYUDurAYhB++YTanIgAGIA0QAC9UAAxKAe44xeAoFQEZg6daqZOXOmeeedd8z2229vJk+ebC6//PJQu9N0u3HHoGno8GIeCIAY5IEiygAC5UcAxKD8Y4QWAoG6IrBx40az2267mS1btjQg6Natm1m/fr3p2bNnrWABMajVcJevsyAG5RsTtAgItAIBEINWoIoygQAQyAMBaS3g8upqNQAxyEOiUEbTCIAYNA0dXgQCQSEAYhDUcKGxQKA2CJC1oHfv3mbbbbc1PXr0MBs2bLB/052D9957z/5dJ6sBiEFtRL+cHQUxKOe4oFVAIG8EYZONhgAAIABJREFUQAzyRhTlAQEgkAcCM2bMMJdccokNUzpp0iTDuuqaa66xYUu/+c1vmilTpuRRVRBlgBgEMUzVbSSIQXXHFj0DAhIBEAPIAxAAAiEgUHddZYlBCAOFNlYXAZlxr7q9RM+AQL0RqPtiW+/RR++BQDgI1F1XbbN161ZLDKoen7YVIln3DHl5Y0rxc/EBAkCgmgjUfbGt5qiiV0CgegjUXVeBGGSQ6bpnyMsAXeSrIAZ5I4rygEB5EKj7YluekUBLgAAQiEOg7roKxKDJ+YEMeU0CF/MaiEH+mKJEIFAWBOq+2JZlHNAOIAAE4hGou65qEAMISjoEkCEvHV54GggAgXojUPfFtt6jj94DgXAQqLuuAjFoQlaRIa8J0PAKEAACtUag7ottrQcfnQcCASFQd10FYtCEsCJDXhOg4RUgAARqjUDdF9taDz46DwQCQqDuugrEIKWwIkNeSsDwOBAAAkDAmEbSIIABBIAAECgzAiAGH4QrLfMglaltyJBXptFAW4AAEAgFgbovtqGME9oJBOqOQN11FSwGGWdA3QUoI3x4HQgAgZogAF1Zk4FGN4FA4AjUXVeBGGQU4LoLUEb48DoQAAI1QQC6siYDjW4CgcARqLuuAjHIKMB1F6CM8OF1IAAEaoIAdGVNBhrdBAKBI1B3XQVikFGA6y5AGeHD60AACNQEAejKmgw0ugkEAkeg7roKxCCjANddgDLCh9eBABCoCQLQlTUZaHQTCASOQN11FYhBRgGuuwBlhA+vAwEgUBMEoCtrMtDoJhAIHIG66yoQg4wCXHcByggfXgcCQKAmCEBX1mSg0U0gEDgCdddVIAYZBbjuApQRPrwOBIBATRCArqzJQKObQCBwBOquq0AMMgpw3QUoI3x4HQgAgZogAF1Zk4FGN4FA4AjUXVeBGGQU4LoLUEb48DoQAAI1QQC6siYDjW4CgcARqLuuAjHIKMB1F6CM8OF1IAAEaoIAdGVNBhrdBAKBI1B3XQVikFGA6y5AGeHD60AACNQEAejKmgw0ugkEAkeg7roKxCCjANddgDLCh9eBABCoCQLQlTUZaHQTCASOQN11FYhBRgGuuwBlhA+vAwEgUBMEoCtrMtDoJhAIHIG66yoQg4wCXHcByggfXgcCQKAmCEBX1mSg0U0gEDgCdddVIAYZBbjuApQRPrwOBIBATRCArqzJQKObQCBwBOquq0AMMgpw3QUoI3x4HQgAgZogAF1Zk4FGN4FAoAj84Ac/MDNnzjQrV640ffv2NZMnTzZnnXVWoL1pvtkgBs1jZ9/EYpcRQLwOBIBALRCArqzFMKOTQCBIBIgUXHPNNWbWrFnmuOOOM4888oiZMGGCmTRpUu3IAYhBRhHGYpcRQLwOBIBALRCArqzFMKOTQCBIBA4++GAze/ZsSwr4Q+Rg/Pjx5g9/+EOQfWq20SAGzSL3wXtY7DICiNeBABCoBQLQlbUYZnQSCASJgEs/1VFvgRhkFOE6Ck1GyPA6EAACNUQAurKGg44uA4FAEIDF4MOBAjHIKLRY7DICiNeBABCoBQLQlbUYZnQSCASJAO4YgBjkJrhY7HKDEgUBASBQYQSgKys8uOgaEKgAAohK9P4gwmJQAWFGF4AAEAACZUcAxKDsI4T2AQEg8PTTT5uDDjrIrFq1ynziE5+oJSAgBrUcdnQaCAABINBeBEAM2os3agMCQCA9AlOmTLG5DCiHwYwZM9IXUIE3QAzUIBJb/Nd//Vdz3XXXmd69exc2xIsXLzZz5841V199tenevXujHbfccot57rnnzLRp0wprGyoGAkAACKRFAMQgLWJ4HggAgXYi8Pzzz5u99967UeWf//xns9dee7WzCaWoC8RADQNtyKdPn25oA14kMXBJB4hBKeYNGgEEgIAnAvDb9QQKjwEBIFAoAmQtoA9bDOj/dbQaVIoY0Gn/aaedZpYvX26+9KUv2dP2pUuXNjb6O+ywgznvvPPM6NGjzbHHHmu//5d/+RcrCN/4xjfMV77yFTNq1Chz33332b/pVJ424v/0T/9kn/m3f/s3+zt9R3W8+eab5vvf/759dr/99rPPnXDCCQ1SId/l8oh4kDWCPhQey3XyLy0G1AdKunHooYeaoUOHmp49e8JiUKj6QOVAAAj4IIBIHz4o4RkgAASKRoCtBWQlIKuB/LduVoPKEIPNmzfbTf/AgQPt5p02/bRZ79OnTyQx2HXXXRsuQySQ55xzjv37lVdeaTxPl0/YekDPULm0kV+zZo258sorzbx586wsExm54IILzOc+97nGM/Q9uwLR/5mQ0P/PPvts+27cxRYmBlTfuHHjbL39+/e35ZCQwpWoaDWC+oEAEEhCALHBkxDC70AACJQBAbYWkIWA3R7ld2VoY7vaUBlisGHDhsbmXm64pWuQtBjwJptP/HmjLZ//5S9/2cmfn8kGDc6iRYusReLtt99u1Lvvvvs2CACRB7Y08GCS1WDIkCFerkpMDEaOHGm+/e1vd7JC4I5Bu6YH6gECQCALAsgmmgU9vAsEgEA7EJDWAjp4Zb2lv29HW8pQR22JAbkS0YcIBbsPPfLII/Y7thJkJQZRG3jfOwwgBmWYHmgDEAACWRCAxSALengXCACBdiCgLQPyQKOOVoPKEIMoVyISKDpxnzhxorn22mutfJHbz/XXX2/IlWjOnDnm61//uv1euvowMYhzJUqyGFCZ7DLElgRyc5KuTXGXm+FK1A51gDqAABBoJQK4Y9BKdFE2ECgOgW0GfKm4ylFzJgS2Lvl+7PuVIQb69J8vAbP7ELkMnX766RYMuk9AFoOoy8F8gfkf/uEfYi8fJxEDV/lpLQZ8gRqXjzPNA7wMBIBAQQggKlFBwKNaINBCBIgYrJh/dQtrQNGtQKDfiPNMrYhBK0BEmUAACAABIJANAWQTzYYf3gYCZUMAxKBsI+LXHhADP5wKe4qsB2QJ0B8KS5oUtaiwRqNiIAAEgEBKBJBNNCVgeBwIlBwBEIOSD5CjeUQMHpxxpv110KBBkU9VypUozGFCq4EAEAAC1UUA2USrO7boWX0RADEIc+xBDNowbq5wfG2oGlUAASAABEqPALKJln6I0EAgkBoBEIPUkJXiBRCDNgwDiEEbQEYVQAAIBIkAsokGOWxoNBBIRKCKxOD2hY+ZWfPuNzdMG2/223v3RAxcDzz355fMrx79vfniP/5d02W06kUQg1YhK8oFMWgDyKgCCACBIBFANtEghw2NBgKJCIAYREP02ptvmbMvu9Ececj+5rzRpyTi2O4HQAzagDiIQRtARhVAAAgEhwCyiQY3ZGgwEPBGoAhi8N9b3jFf+948s/GtzbadDy9dafbcbZfGCf+Sp54xY6ddb8YNH2Lm3L7Q/kub86vn3mn/ps9JA/ubr3/lNPPRbtvbv/k3KmfgEX3NoidW2vLo8+Xps82E04aa4UOONlz2TdPPNgMOOcBwW+5etNQ+S3VNGHmCbR9/x3Xdu3iZmXbdT+1znzzw4+b6i75oeu3UwxvrPB8EMcgTTUdZIAZtABlVAAEgEBwCyCYa3JChwUDAG4EiicGyVWvs5r3XTjva0/l9du9tN/srnvlTgxjwaT1t/H/71LN2M04feZrPm/3p55xuhh17uN3Uc9lJxIDKffGV1zvVS6Rh/3326FQHuRUxwTh+QL/CrQkgBt4i3vyDgwcPNg888EDzBeBNIFBxBJAhs+IDjO4BgZQIJCVYSlkcHi8AgSKJAXWXT/3lvYBX39hoiQGf6ke59cjnFzzweIM00Am+/C2OGPDm//NDj7HWBPnRdTIxWLf+VUMERD/f7qEDMWgD4rAYtAFkVBE0AkUsIEEDhsYDgQoj4JNgqcLdr0zXitDr7L7jSwzkaT1vyPMgBrvs3LOTm1EcMaDfpCsT/V0kQQAxaMMUBDFoA8ioImgEilhAggYMjQcCFUYAxKAag1uEXk9LDHwsBvc+8mTjjkJai0HUBeO4y8fsuiTvRbRbGkAM2oA4iEEbQEYVQSNQxAISNGBoPBCoMAI+G5MKd78yXStCr8sLv9KfX98xYFciPq1v5o4B31/gzT+f+nPZ8u7Ca29uct4j0JeW5XtFXED2mX/IfJxxmoIYZAQQr1cegSIWkMqDig4CgUAR8NmYBNq1WjW7CL0uLQY79uhu5v1ycacoP3oTzgMiXXk4UhH/RlYCihikoxJRHgP+jZ7lSEdxUYnkhWeKgsQRiB5csqIRlYjKksSl3ULjM/9ADDKOCohBRgDxeuURKGIBqTyo6CAQCBQBn41JoF2rVbOL0OtRrkS1Aj2HzvrMPxCDjECDGGQEEK9XHoEiFpDKg9rCDrYrQU+aLKN0Erj2pQ2xET30poFDFxZ5OtfCYQq2aJ+NSbCdq1HDi9Dreo4POOOfg0V82HH9zZWTR7e9/T7zD8Qg47CAGGQEEK9XHoEiFpDKg9rCDpaNGERFFvHpvsutwOddPNM6BHw2Jq2rHSXnhQD0el5Itrccn/kHYpBxTEAMMgKI1yuPABaQYobYle0zKjtox+Cj7OU5irVN2TrpdF5G3JC+tpzNk3ols5A+s/alRnQPV49dWUbJn5fbRe+yby79nxIS/W71H22RUZlM+dnu3baz7aGPTHbEFgPZh6KzjxYjEeWo1WdjUo6WohVxCECvhykfPvMPxCDj2IIYZAQQr1ceASwg7R9in2yfvMlm8zyRAcoO+uzaFzplD5Un7zKr54SRJ3TKFEqb+7iPb5ZRzkDKG/x161/rFDOcyvnZ/b+xm3/+bdhxhxluTxQxkHHHy5B9tP0SUZ4afTYm5WktWuJCAHo9TNnwmX8gBk2O7Q9+8AMzc+ZMs3LlStO3b18zefJkc9ZZZzVZGl4DAtVFAAtIe8c2KXa3zg6qXXX0+zq8Hv394iuvm6njRpjL58xvnNB/tNv2sR3V5cg7BstXrzGz5t3fKZ44bf6JqMhQgDJrqI404kMMypJ9tL0SUa7afDYm5WoxWhOFAPR6mHLhM/9ADJoYWyIF11xzjZk1a5Y57rjjzCOPPGImTJhgJk2aBHLQBJ54pdoIYAFp7/gmZftMQwx4s333oqWdOkGuODPOH2u+M/fOXIjBggceNxTeT344CRB9R25OE04bai8fs0sQWTzGdgyxrkbk9hRHDAYcckCpso+2VyLKVZvPxqRcLUZrQAyqIwM+8w/EoInxPvjgg83s2bMtKeAPkYPx48ebP/zhD02UiFeAQHURADFo79g2azEgdxyKw51kMeDepA0dSCf8riyj2mIgEZNER7sZbd7yrjcx4DLLkH20vRJRrtp8NiblajFaA2JQHRnwmX8gBk2Mt+teAe4bNAEmXqk8AiAG7R9inzsGOlFPs3cMqHfk75/kSpTXHQPZN04cRNaDOIsBtXHstOsbiYWKzj7afokoT40+G5PytBYtcSEAvR6mbPjMPxCDJsYWFoMmQMMrtUUAC0gxQ+/K9hkVxpOtBBT9J21UIl9iQM/FZRmNikrUa6cehi0T5M5EbRvbMdhM+taPbAQlcjfq/Zc7mn3/atcudx50HgMZlYjagvwGxcilz8akmJah1jQI3HjjjWkex7MlQID2rsdN+rF5cMaZtjWDBg2KbBWIQRODhTsGTYCGV2qLAIhBbYceHQcCXRAAMaiGUCxevNjQRhOfsBDY5bMXghi0asgQlahVyKLcqiEAYtDeEZ30rR+aXz32+/ZWGmhtQ4851My8YGygrQ+z2SAGYY6bbjX0epjj6DP/YDHIMLZPP/20Oeigg8yqVavMJz7xiQwl4VUgUF0EsIBUd2zRMyCQFgGfjUnaMvF8+xGAXm8/5nnU6DP/QAwyID1lyhSby4ByGMyYMSNDSXgVCFQXASwg1R1b9AwIpEXAZ2OStkw8334EoNfbj3keNfrMPxCDJpF+/vnnzd577914+89//rPZa6+9miwNrwGB6iKABaS6Y4ueAYG0CPhsTNKWiefbjwD0evsxz6NGn/kHYtAk0mQtoA9bDOj/sBo0CSZeqzQCWEAqPbzoHBBIhYDPxiRVgXi4EASg1wuBPXOlPvMPxKAJmNlaQFYCshrIf2E1aAJQvFJpBLCAVHp40TkgkAoBn41JqgLxcCEIQK8XAnvmSn3mH4hBEzCztYAsBJzUTH7XRJF4BQhUFgEsIJUdWnQMCKRGwGdjkrpQvNB2BKDX2w55LhX6zD8Qg5RQS2sBWQeYGOjvUxaLx4FAZRHAAlLZoUXHgEBqBHw2JqkLxQttRwB6ve2Q51Khz/wDMUgJtbYMMDGgYmA1SAkmHq8FAsiQWYthRieBQCICvplXEwvCA4UjkEQMnvvzS+aqmxaYSyeOMpTBXH4oC/mjy1ebr3/lNPPRbtt36QtlbR/Yv6/9/mf3/8b5HP1OWdsvvvYWc/7YDrPnbr3MlTffYUadONDst/fuXcqlNl383VvNpeee2el3qq/PXh8zw4ccbej/v33qWXP9RV9stFv2hQrl+rgOmTmefv/kgR/v8r5PvXNuX9ilzdPPOd22K69PaYgBCVAlPv/zrjEv/c6Y3Q4yZrsPBP2FJ4zZ44j3u/fuW8asX2XM7p80ZtvtKtHlrUu+X4l+oBPFIYAMmcVhj5qBQNkQ8Mm8WrY2oz1dEUgiBnGYxRGDJU89YxYtXWnOG32KF+ySGNBGnTbxt9yzyFww5tQupMOXGNAGXW7I44gB/fbl6bPNhNOGNjbw1L9Z8+43N0wbbwmIb73UYd9+e4ET8VCpiMGK+Vc32w+8VxACJEAgBgWBX6FqsywgFYIBXQECQMAY47MxAVDlRyBJr2uLAW34x0673uy52y5m4BF9zaa3NnexBPz3lnc6nfjTO2QxmDpuhLl8znzzV7v+peFT9XHDh5gJI08wX/vePHP3oqW2XN6Is8VhwCEHdALSd4O+afMW84dn1zYsC3HEgOqK2tDL733rBTEov9zXvoUgBrUXgVwASFpAcqkEhQABIBAEAiAGQQxTYiOT9LrcTL/25iZ7qn7ZxDNMvwP2tZt5+mhXIn3ar4kBv7PimT+Zi679iSUCvXbasYtrj8vq4LtBJ7ci+qx5/mV7gu8iBuS6RH35/NBjjCYh1IaZc++yLkXUfx9XIhCDRLHDA0UjAAVe9AhUo/6kBaQavUQvgAAQ8EEA64oPSuV/Jkmvy830s2tf6HRXgDf8mhjoDb0mBp869EDrriPdh6KIgcudKA0xOH5AP3P2ZTeayaNPNrvs3LNxX4JGhu8YUN38TJR1gu9Y+BIDfcdAWkHykgif+deWy8dJApRXh1FOvgj4CFC+NaK0KiKA+V/FUUWfgEBzCGBdaQ63sr2VpNclMXhwyYpOl41dxIB88/mUnvqriQGfzLeDGBAB4fr/7ynHm+t/eq+9SC2JASwGGaQySYAyFI1XW4gAFHgLwa1R0Zj/NRpsdBUIJCCAdaUaIpKk11thMchKDPRFZRoJutdA7kBsjZARivi3HXt0Ny+8/GoXYkAXi/Udg/t+/aQZdMTBZtZt99mBJlck33r5+VZKiM/8g8WgiRHg0FRHHrJ/y2+QN9G83F7xEaDcKkNBlUUgaQGpbMfRMSAABLoggHWlGkKRpNe1Xz673GS5Y+BLDFx3DHijT5eYOfoPPcv3FXijz6FLaaQ46tCuvXra+wLSYsARhzgq0bBjDzcPPfEHM+WqmzuFLPWtF8Qg4LkBYhDw4KHpbUcgaQFpe4NQIRAAAoUhAGJQGPS5Vpyk111RiagRU8cNN889/3KXkKJJUYmiiAG78yxbtSYxKhHVzZt0imREH+3HLy0GDBi5OFF0pChiQM/oPAanfe5YG3VJtsmn3qg8BhR9Kc8Qpj7zr9YWAxIAHoiTBvZv3JDnsFo0IPQ7/dsx+Ch7q37d+lcNPbv2pQ1GWgxIcKZd91MrR1wW/Z9MVBvf2my/f2btS1Zwl69e03hWJ8LIdeZmLMxHgDJWgddrgEDSAlIDCNBFIAAEPkAA60o1RKFVej1tHgONZlweg2ogn60XPvOvtsRAZrcjmMnMxRt9SQyIqTHTIzJAjJFu2FM8XmZy/PxN0882+++zR6MsjrErWaNMhsG33svqkuQjQNlEFG/XAYGkBQQZMjtn2PQJa9eODJl1kE30sf0IYF1pP+atqDFJr2ep05WHIKlMbXFIer6Ov/vMv1oSgyhXIJmp7tU3NtqNP230KQSVzmyn39cptOnvF195vZGUg4SPw3JxWWR5yDvVdd5C7iNAedeJ8qqHQJYFBBky35cHad52JdSpnuSgR1VEAOtKNUY1i16vBgJh9sJn/tWSGCSlsE5DDGTmPSkm5CI04/yx5jsfZMWT8XqlCxO9U1aC4CNAYU4NtLqdCCQtIMiQ+WHmTN842zR+efqdtlMeUFe9EcC6Uo3xT9Lr1ehl9XrhM/9qSQyatRgMO+6wRugp6XqkLQYsSuyCJC0GUsxkinBO5V0mMfQRoDK1F20pJwJJCwgyZJYzQ2Y5pQmtCh0BrCuhj+D77U/S69XoZfV64TP/akkMaKh97hiwK1HWOwaSGMj7COSm5CIVZRBHHwEqQzvRhnIjkLSAtCLeNTJkllsm0Lr6IoB1pRpjn6TXy9RLnwvJUdGIuA9ZL0SXCQuf+VdbYsDkgC/xyZBQevNOz8pwVGmjEmmLgYxgRL8xASmT8FBbfASobG1Ge8qHQNICggyZsBiUT2rRolYhgHWlVci2t9wkvd7e1rhr872QHEcMeL84sH9fe+805I/P/Ks1MQh5cNvRdh8Bakc7UEfYCCQtIK2wGPgmwnGdJPlmquREOFXMkBm21KH1ZUUA60pZRyZdu3z1es8e3Q3nDJCHoPKAlHMJcE4CuqN58x0P2fDwdAdzzfMv29DxMry7zgvgOmDVp/188Eu95fIeXLKiEUKe6ttn995mxo/vNBte32QOP6iPDR6zbv1r5pZ7FnXJvZAOteKf9pl/tSQGk6+8ydz/m+XFj5BowdBjDjUzLxhbqjb5CFCpGozGlBIB3wXk0omjbPuRIXOeScrMSTjh8nEpxR2NSkAA60o1RMRHr3NG4OFDjrZu0xStkTbZK575k5k59y4b/r3XTj3sb/ThYC70f36OIkTSZp2yClNeKHYTldHZdPZiibAMfaoPfIic0Ifbxwc9UeX5Wh7KPro+86+WxKDsA1eW9vkIUFnainaUFwGfBeSqmxYYIga0SMgTHWTIjM7M2Y4MmeWVKLQsZASwroQ8eh+23Uevy5wspNcpe7CM0Mil0QadrAJMDHjzryPWsbsP54CaPPpk69rD1gO2FHO5ejMfZQnmZ6UrEbVVEhf5TOjuRD7zD8SgGnO0Jb3wEaCWVIxCK4VA0gLSbGezXgjzuZDWbNvwHhAAAtEIYF2phmQk6fWoMNRMDAgBOv1nFyP6m+55MjHgDX4SMfjd6j92AlOHfo8iAjKXFN0XZaKiiUEUiWk28VqZRtxn/oEYlGnEStYWHwEqWZPRnBIikLSAZGlys4q6KmbhLNjhXSBQBAJYV4pAPf86k/R6HDG4d/Ey8+jy1Y1NubYY+BCDi6+9xZw/tsPst/fuzs4l6Xmul9wyQQw+hBHEIP/5UpkSocArM5SFdiRpASm0cagcCACBtiKAdaWtcLessiS97ksMNm95194rO/KQ/b0tBnwngDpHm3q2Alw28YwuUYPk4ZFuU9wdA20xSCIZLQM654J95h+IQc6gV6k4HwGqUn/Rl9YgkLSAtKZWlAoEgEAZEcC6UsZRSd+mJL0eRwyYDJArEEUGGnPq8WbhY783U8eNMJfPmW+SLAZEDHRUIu1GxD3SLqcyGpJ0JeLvOSqRJgZVcT31mX8gBunnQ23e8BGg2oCBjjaNQNIC0nTBeBEIAIHgEMC6EtyQRTY4FL2e10l/s26rZRttn/kHYlC2UStRe3wEqETNRVNKikArFxA65ZG+qlkhcEXO4ASHHAVD18OnSeeePqzTiZerPfI07bU3N1UiPnZW7PF+PRDAulKNcW6lXs8boayn/VkDXeTdnyzl+cy/thCDG2+8MUs/8G4BCBx88MHmuEk/Ng/OONPWPmjQoAJagSqrgEArF5C8iYEL7zhi0MyJVJyvaxXGHH0AAi4EfDYmQK/8CLRSr5e/9+G20Gf+tYUYLF682NBGE5+wENjlsxeCGIQ1ZKVsrc8CIkPIyeyWdFKjM1BSchxKekPZMgce0ddsemuzjW5BHxkCjzNhUtkUT5s+22yzTSOpThRY0mIgyyNf1LUvbTBRFgN5miTjae+yc09D+Rl05s/999nDXrZj/1pK8kOfK+bMNxeOG2FzOeADBKqKgM/GpKp9r1K/fPR6lfpblb74zL+2EAMIUJgi5SNAYfYMrW4nAknzn0/j6cJZVIbMi679iblh2ngblk5Gn+h3wL6WCNCHiMGs2+6z/6coFTJzJX1HGTijIlZoHCQxoPJkpk4iI0w25HvS91QTA1fmz3XrX7OkgZO6UXlV8WFtp2yhrvAQwLoS3phFtThJr1ejl9Xrhc/8AzGo3rjn1iMfAcqtMhRUWQSSFhDtVsMn/Jeee6Z59Y2NnTJQ6jsA/Pd5o082k6+6uXGirzfoMgNnHNBcno6O4ZtZM65e2fYoYiBD51VWGNCx2iOAdaUaIpCk15vppV4LmimjFe/43FGQeRCiDpwWLV1pD62K/vjMPxCDokepxPX7CFCJm4+mlQSBpAVEp5+X2SqJGMiwcfpOgSYGUZkwDz2wT5fTeRc0PkRjwCEHNF7XmTWjXInYKgBiUBKBRDMKRQDrSqHw51Z5kl5vpqIyEgPfO2RxxICwKItF2Gf+gRg0I701ecdHgGoCBbqZAYGkBSTJYiCJQZzF4Bs3/CwyE2aaxSZvi4F0FwIxyCBEeLUyCGBdqcZQJul16iW7icr7VHSHinQhuWbyh1w0o+5eUcQ2csdct/5Vm++A7mPxHSzOO8B3zXbs3s2eyOv8Bq67Zn9zwL7moP1xSC+jAAAgAElEQVT2su6r9HFZbHVEItl2btODS1aYadf91JbDeRD03TiyEt9yzyJzwZhTzUe7bV+YEPjMPxCDwoan/BX7CFD5e4EWFo1A0gKSdMdAEgMZHSjujoG8i8CXgKU/f5LFgO4s3Lt4WSMUKl94TnvHIA0xKMuJUtHygvqrjQDWlWqMb5Je1+6XtPFe8/zLpmPwUTYYBLmK0r0xaQWWLpaEEgVp4IAPpB/5zhc9x2X02mnHRuZkIgbyOdLbfEeNypN3zXSgiStvvsOMOnGgbZP8SL2sLcRxmZPl3Tgqz9fy0Grp8Jl/IAatHoWAy/cRoIC7h6a3CYGkBYSaEReVSGeglCc2U8cNN889/7I9haGPjErEmTCbsRjoKEcUlWjHHt3NsGMPM9KViOpMikoU5UrEmT/pfUQlapMgoppSIIB1pRTDkLkRSXrdV++6LKnPrn2h0/0yWd7/a+9cgO+qqvu/MxTSGIJCQETEZoDJgNDyENT+IQ1J/6JBIP/gGBSGP2GiaGqFIaBNwLTTIoSOJBgeDY1NJSAIqPAPCAidAdIQ6wMJtCCRAQabAkJIABNIw6P5zzp03a67f+exz73n3vP63BlH8rv37Mdn77P3/u691tqyQy8iQ232VXTMmfmJaA7Qm5NVnHzskIlOTEqtIJFFvkaCS7pLxl/M+8LAFxAT9n5vdALhm8fq76qw+RPy/iEM+n49mptASAdqbu2pWVEEsiaQovIpK50idoJwPC6r9ch32ASYV4ZNfDD5ZY3rSYtjKY0skJffem+nYLLxIpsx9sRAhIE1N5Ifi9mQRKhbed/Po2d9YTBr+tSuUwbNSxbscb5mulCXUNTyUbMiLVicELCbWFpuMQ2yPgZJF2UiDExfzOpAg+m2pNovAQbwfgnyvBCo0vtvB3W/dewgn7flQqJWJKXZz7N5y8nvIVA2AeaVslugmPyzxvWkEwNfMKSdGPinxVpyPSHIe2Lgh4iWvO9a83CUbJwZUdamjy0HwiBnv+q1A0k2WTebqgKT3yZ1ojj19/49dnVJNmXyexsy0dqc2caX//7FY091OcTYl0HSueDy67scIq0zjnzvO9SE5mvVttZPTSdyNk/izxnAiyLZ7nSy3v9206H2EGgXAeaVZrR31rju3xava7npU450V97442jdNGb0jl130aT5GMjzssaT58T0pxcfA18YaBn32XN8dGIR5xRsd/nTbqzPEgZZImNYvSLk/au8KVGaMPC9xbPA+sdCaTt1eRbodkGeJgx0t3LOycd2ecIvvemerguc4mKu+4JE6jromLghHSiLOd9DIGsCgRAEINAeAswrzWjrkHE9zndMxcAdqx+KTIPOPf0Et/Len0eXPcpHHI7lowJAoxKpGZFu1GpUItlcnXLEwW7L1q2ZUYl8YSD5ZIUZ9deZmq88a0+Z9e8alcjfqK7KyXDI+1cJYeCrMHUu1DBUW17bOkLN+erLDzP4vt3f07Fhmz1jqlOnFO2MepNqks1XqDDYsnWb+9VT6zse9mnCQPKKW9Dbv4fmizBoxuDahlqETCBt4EAdIQAB50IWJnCqPoGqjOvWwdj3EciimOZMrM8WtdNfBf8CqVPI+1c5YaBxay8+63POD0doj3l89eULAwEgR0M2XJWEtfJNe5JOHUIX6OLUIh/1kE8SBmK6ZL3lbYe1Nnf2iCzNhAlhkPXK831VCJQxgfjH2HlZ9BJRw45PWfnr+PWVz05zC5ff0omikVROW56kCBp568jvIVAGgZCFSRnlIs98BMoY17WE/j0Isvmb14IizoIjbfzt5w6CvNYt+Voi369D3r/KCQPxRE+70Mh2DHvFtC8MJDyVqEerCOOEQdLxTh5hcMwRB3c84W3MdCmrChGNtasxeW1T+pN+iCmR72PgH7Pl6yrxvw7pQEXkQxrNJlDGBJK1MM8iHioMktJJy7+XHag029asuvA9BKpEgHmlSq3Re1nKGNd7Ly1PKoGQ969ywkDi0/70kSc6pkNJYZ98r/SkG0uHIQw0bq0Imv97wjHuqhvv6tjLqTDgxIAXs60EsiYQXfQedsC+bskNd3RC0smJmbzX/g2SetmY8LQ2nvbGS/m7hKATIe5fcOaPKX74uXNOO97NvXSF82/rjGs//5IcvUfB5p/33oNxY8c4MXmUT9KNoPKdxuDWm0Db2r+od70IhCxM6lWjdpY2a1xvJ5Xq1zrk/aucMCjqxEAvuOhVGMTZnvm2bNZpRb+TS5Cef3HTCGEgixzfx+DunzzsJn/4Q27pzXdHvUmOwkLz1d8PshuGdKBB5k/azSCQNYHownza0YdG74ANOGBNAeUd0t9aU0PxJ4q78VJiYMvCOk0YyEVj1rxQ3+m4mNdZwkDeY72Zs9ebksXRToMT+Dd9+o5zVbFZbUYvpRbDIsC8MizSg80na1wfbO6k3iuBkPevcsJAKqvXYPfjYxAqDJJsv3Shr4sOKZf81l5z7Xuz66Jl913HdW4ztYsOa9M27ajD3Kpf/sqde+mKrpClofmWLQw2b97sli1b5s4999xe+yfPNZzAs88+6xYvXuwWr9riHr3lssTa+mZ7VhxvenVz1+2X/m6/+ucsOm+W+9Z1t4+48VLGgTRhIAv4uDDHoaZESSeV+h7rOKSV982I7O+knNaMMCm+t54QcClaw1+ghlZPFiYz93vFzZw50336059uaC3rX61Fixa5M888040bNy62Mr0Kg9CxtQiCWZsnaWWpShShIjjYNGopDGTSs44l82fPcE8/+6L76ukndsWYzYpKFCcM1Jxn7bpnOuFB0zqONU0QsL4df1yYKxtrV57Jusfg5E8e5STqki1TSL5x9xj04oCT1uniOpAIgosvvtjJoLHDDju4rVu3Ft1vSa/mBDqCYPFiN3fu3CBhYHfDfWFgF+5++GId2L92xgx3/hXXR6ZDYrrjL7ht+nbBfdeatV2mi4o+dPLStNT8KC5/a0rknwiGltPG90YY1PwFaXnxVRjcfPPN0fgg/9t7771HUPmHf/iHaGPh8ccfdwceeGD0u89//vMtpze86o8ZM8a9/fbb0ebf+eefP0Ig9CoMhlWDEIffrHG+iZsvtREGvXaUkIZPS7upirBXnv5ztgMdfvjhkSC47LLL3Pbt2yNRsHDhQnf22WcXlR3p1JyALwh0ws+aQPwTA/tvOTFIC0bQy4mBb6pUpRODJAGDMKj5y0HxOwR0XnnppZfcT37yk2jx7wsEEQVLlixxS5cudUcffbR74IEH3Jw5c6L5BnEwnM4k/OfNm+f+67/+y40aNcqdc845XQIha1yXUsrmqW5i6salXYxfs/LezvdaKzH/lM0Uu0Fsfcns3/0LYjWNpJNZ9d2SuwY0aIz6kn1t1nT3t9/5f1ESUl+5R0E+TfPlarww0I436fADo46U59NLZJA86Tfht9KB7rzw/7jvfve77oc//GEkCN54442oauPHj3cysPOBwK9//evIrCxugo8G2SO+mGlKZG3r0xbuaT4G/nPWx0DSjwuBLAtua76jfkDTp3zExV2G47d20ulDrz4GeYRB1jE5PRMCVSTgL0z8DQUxX5kxY0Y0pogo0I+IA/nuV7/6VRWr1cgy7b777m7jxo1R3XbaaacugbDLlPNSx3U/MMM3V9zmTj1uUpRW3NiaFJ1RLT3ErHvW9KldVhhJO/r+pq8NViOntrrYl9DPWhYbKt+uJ5s2zrZCGDTybaxIpaQD/cnYddGuzltvvVWRUlGM2hH48JmZwkAGZ43GY3eB4qKSJe0k+VGJJBDAtKMOjTYN9FZKMQc8/cTJUcQhuedE7h6IS0+cku0NnEmRf9KiEtn8bZvZk85QUyK/PE3cyapdv6bAPRGQecX9clnms7IR5X9kJ5dPuQR23HFHd+KJJ7ofPjM+WBj4d1D5wsA/NfZNRq3J5t9c/X133qzpzt7zFLdhY8PZ+1Es9fchoeKbZk6EMCj3/al97pwY1L4Jh1KBIk4MQnbnh1KZIWRSxGll0yarIWAni4oQ4MSgIg0RUIx+Tgwked2Qkf8W8x0J7e7b9cfd+WKf02LqhpHu7D+3YVNXuGpbnTghYM2a1FzJFwZx81DTxlqEQUDH5yfJBPAxoHfkIdCPj0FVhYG948BnYe1e83CS3/bj39TPs3nLye8hUDQBfAyKJjqY9MTHYP78+ZEDcq8+BlqypJPRMaN3dHL3ix+9LWmH369p0u/S/E9tAAhJz5oSIQzeIVyJcKVFdussL/Mi8wpNK61M/pGZn6bat8nf45wk7e9thxe7PLXpSztyS6sDUYlCW5jfWQJ5oxJBDwIQaA8BohLVo637jUpkF+32lFQX4wu+NDMKMf2xQyZGJwn245sW6Z0us6ZPcVfccGd0T5SYd4b6GNgIkmk+BnHCAB+DAfXXLOfDIrOtojBIq1+aMMgbdckPhdjvzmLakRP3GBTZa5uZVug9Bs2sPbWCAATiCHCPQT36Rb/3GPhh131TIg0xLf5e9qO/S4o+ZM2Mkk5tfXNNNVfSvNSUSP8u+UtUomU/+KeO6JC/WRHRlBvma2VKZBvO2pIlRemQRpMjqLTwUxJuytqjxTk1ioPgTT9eE10yJjHI5QIzsV3Tzin5xDknav6vvva6e+ChddENq0mRkXyxoumJI+SkDx8Y3WOgjpD6gmTd0yAe+jYM2JyZn+jwsPct9KN2QzpQPYY4SlkmgWFuDJRZT/KGAASyCTCvZDOqwy+qPq7n3ViNY940/wKpY8j7VwlTIv+WUD2C8kMG+qEBn3n2RXfOaSd0qTobfkogSGQRsV+Toyo9jpJFuA0nqDcsr39hYxS79qn1z3duW5X0NJyhDZulC3F7M3LSy+w7uCSFTvQ996+/c3XnYjf/hlXJS+uhtzHvusvOIy5U6+flCOlAdRjAKGO5BKo0gSQ5m9lxY5g7Q3GOd0W2lo6tccf1/o6YxBSfsPd7Rxzr2/L4/OwYVWS5Sau5BJhXmtG2VRrXk4j2szHar8VFVVs55P2rhDBIMv+J22lXO3u5sVSFQdrEZU8c/IuTFl/3o0gIyELA2qDZfO9/8NGum1H9BXrShJtUJhEdaRc26XP+gj4pX2s+FCcM+uncIR2oqp2fclWHQJUmkNCxZlj0yhYGeSdOn18Td9SG1fZtzYd5pRktX6VxvRlEh1OLkPevEsJAby/VRbriSRMGsrueFX7K7vzL4t8uov0bVdOEwYIrb+xqMTE7WnTerMhxxvemj2vaEKHhmxL53va+MNB8EQbDeZnIpXcCIROINdebP3uGW7N2XWTr6Qtp/6IaORFUu1H/Zk29F0FKLqZ+++2zV3SCaG+6FJtSa+sq7/Yf7v9Bd8C+e3d2zkMXv0n3K8jY8uvfPBeZHIqJ4rSjDuuY/YmNrJxUihmjmCLaKEi+6eOia293G1/Z4g47YELntFAucZOPb2urY+OfHH5g9P3H/9chsQ5+dsdfx8BDJk6IInVk8WvqzaC993SeDCEQsjAJSYfflEsgZFwvt4TkHkcg5P2rhDAI3cVLctQNDT/lnxjYnfs0YRB3MuGbP6V1QVu/ok4MEAa89HUhkDWB+Lvm8i7+4rGnOmZ99j1VYaCmfHpiZ99t4WJvUrYmhHLTcVx4OmtKZN9RSSskupcu6O3tympmaPPXDY3fvvRK1+LeChcVCb7po5oMSpQxW19r4iimlcJImFkfqzknHztCGPiCxwqDEH5qcpX31KEu/ZZyDoZAyMJkMDmTapEEssb1IvMireIIhLx/lRAG/sJABYCEpjr7b7/j7GQreGR3fenNd3fsYZPCT8lv03wMQoSB9TGQCVkn6/mzT3ILl9+S+8RAyySTv/o2aJ3y+BiECgN8DIp7oUipNwJZE4h/YpgmpJPiVvubA+oXJO+s9U0KEQZSy0uW3+LmzT4pCl4QYkfv39Bs6yS2+/KRRbu/oWD/Lb+x5o0+B/td0s2g/tgoadpND9uC/oLeCoMQfioMQk9Ueus9PNU0AiELk6bVuYn1yRrXB1nnIk0ws8YvP9qjrVcRl1UOklNc2iHvXyWEgRQ+6QhdQ1NJpJ3TT5wcmQHI5Ld125sdswB5Pi78VEhUIjXhSToxkMkvLmxW0sUccQ2RFJVIfitmE08/+2LHyVifz4pKFCcMdOdw7bpn3NULzoyuDO9nNy+kAw27U5Nf/QhkTSD+ojpUGNj3UqhoNC7576RoZiHCQH2OJh1+YGTmIx8/zrbfCv5C3dbBOvX6E5ovDNQ0SNPXOvmmj0k3g6qJo/V9ihMGcSeevimRxgpPElYIg/q9i1UoMfNKFVqh/zJkjev955CcQlHCIMQHM00Y6No1ZPNokDzypB3y/lVGGOSpWFt+289ufxEdNqQDtaUtqGfvBLImkDwnBrJ4lc+s6VOjjQE1u0kyJ1Rhr6eDocJAynTXmoejvE49blIkstM+WScGGu0n68Qg6RJDP/20m0F9IdDLiUGIsEIY9P5OtPlJ5pVmtH7WuO6PO3bMSvOl+sdb3zlh/eeHHu9s9sj4a+9F6Mc3y1pmxJ2apoWBX/IXZ7hrVt7n/DD1/WzADrs3hLx/CIMCWsW/yMMmae8U6CWrXjtcEUdcIR2olzrxTLsIZE0guvtjwwpbHwMbjlfEwJEH7TdCGMiks/Sme6KTMvmELGytX4EfrlTLtM+e40fcMRLXelk+BjYMqD1dsGGT1TlaxY71FfB9k6yPgTVx1BDGanak97jk9TEI4YePQbve46Jqy7xSFMly08ka1+1mhpRU7p2SMd4f57LCyMf5asWNm0lh6a1vliXmX17ml1d9y2y0R9+fS9PrdxN3mC0Z8v4hDIbZIjXLK6QD1axKFLcEAlkTiBTJ7iCd/Mmj3PMvburcQKkRdiRKz5QjDnZbtm6N7PWtOY2Y5ImZoUwOu717XKIwUBNEydPedCn/FtEhn7gQxiHY0qISWWHg73zJJYvTjjp0RFQiu6ngnxhIeeJMHO1iXXa+lNnuu43LHZUozpTI8iMqUUiv4Dc+AeaVZvSJrHHdDwihmw1+tMgsXyoJ/qLBJtSE2p68+uN9Wlh6S943I4obY+X39jRahYEfpj7EJKkqrR7y/iEMqtJaFSxHSAeqYLEpUsUIZE0gfnGTopQNs1pZdqXDLMsg8+r1RFLLlOW4N8iyk3Y9CTCv1LPd/FJnjet28S6+Whrd0fcNk3TTfKnkOd90NC1oQ1pYeluHOCFgN5sktLT4lsUJAz9MPcKghz6d1YF6SJJHhkCAAXwIkFuQRd73v2xhoKcX1vzGnmj4TebfIVCnJvWP0/OUvU6TYZ568dvBEmBeGSzfYaUeMq6rX4GUSYI5yF0tSTvz8pskX6p+TgyS/LbSxi8rPMT06YLLr3fnzZru9MQAYVBALwvpQL1mk3S3Qa/p+c+FpN/P7mLIxJzkPKgv0uqHHo9MK4r+MIAXTbSd6Q3y/W8nUWoNgfoSYF6pb9vZkoeM67qhsvuu4zommnHh6fXelbQLLbN8s5J8DJKEgb/2ss7R1kczzsfAFwb4GPTQp0M6UA/JRo+ELNx7TTs0/X6EQchRfpowkDKGpNELAwbwXqjxjE9gkO8/tCEAgXoRYF6pV3sllTZkXNedd3Ug1rTsCWyaL5W/WBcH5jtWPxTd9J7km+XfGJ8kDPy1kx9ERk2J9O8SBl6jEvnCYFBrsEH0lJD3rzI+BmnhqxZde7vb+MoWd9gBE7puC5UONenDB7otr22N/i4f7Tjy33q3gTqjyN9GjRrVUa5J0NUGzk/fvzth9oypHacY6azawfW4Sf5my+HnF+f8onHMtXPf/+CjbsGVN0aPSkeVKCk+DwnBOIg4uiEdaBAdlzSbRSBkAmlWjakNBCCQRIB5pRl9ownjehHmkEWkMcweEfL+VUIYxIUr/O1Lr3REgA03FRcWUKDqjZ/y32JWI4t7fU7+9qULl0U3KIuNW9onKX296Vi90a3nuz1q0tCBceXwY6Fbxz3/1MF+Z08MbL00vSJCk8YxCelAw+zQ5FVPAkVOIGn2qUqnXx+Ffp+P2wCw4T+LaEXdodLb02UT4tP/+6Nuw6bfdaI55c3HH4Py7oLZ3/c7WYa0c2j9sk5c49Kxu4eyATQIU83Q8jftd8wrzWjRIsf1Mon0E0BhUGuvQfIIef8qIQz8iTgt3FTcRUJyVHTOace7uZeu6Fx45IezuuCKG9xFXzmlp4uK4o6i7CRqhYH8t714Ke6GUW10O5GmmSP5wkBjlGtoQkkv7yQe0vFCOlBIOvym3QTqNoEULQyKbn1rz2rtZP27GPLm649BIf5Pmkecva58l3VjdN4y9vL7XoRBnrr3UqY2P8O80ozWr9u43gzq/dci5P2rhDDwbz5NCzfl+xSoUFBhILHM7UfMbw6ZOKErrnka2qT09eIgNfWRNNR0KE4YxJXDTpJxgsGaU9lIJ74wiBMqCIP+XxhSGAyBkAkk1JRw+pQj3cr7fhGdEFrTPnlfNr+2NYocIR8bMzvpJk09qdR3VXeGQ4SBvr9i8rfitlXuuQ2bIlM/ecd900V7X4KI+aT3POkeBNsqvlOcbEJI+WUsuuALJ7mbfrymc2KQlJ5vSyvs9ETUmkTKiWTobpp/QhA3HinXww7Y1y254Y6uW02lrNZE0raz3FQaF0ZQN0T0plLl7/diKcvOY8a4+x58NGJlTwDi+p08b7nqpXly6izt7NswJ5m6SjppEatsndIuwszbX+x8ERfvPY6/lDXtHo4tW7e51b98PKp/v1G4QhYmgxmJSLVIAiHjepH5kVYxBELev0oIg6wTA7sQTjsx+Jurvx8tDHyTnZCJXpGHnkiknRhoaCu/HHETlobw8r+zTjcIg2JeCFIph0DWBJLHlNC+n0tvvjuqkF52FnfzsUS5EDEv/kZqdiOOcBr+Ls400AoLeyrnL9DFn0nic8tFX7Jbr6eS8jtrumjHH/nOjg/6bsvmhT6fdLumPOsvwO04ZMutNx6L+aStt7CSPJWbjRBiw/Lp2BVqEuSbRV6y/BY3b/ZJzvLTBe60ow/ttNlPH3ki1mTUtrPccqpzgIhB5ffIE884fV78rJLMRaW+dz3wcHQrtj3RTbuB1eajz4TcrGpPu9Pa0d8Ms21i+1mceZdclif9Resb1756oZ4vDOSZOP6WX1x6ehO57VdZZrlJo03IwqSckYpc8xDIGtfzpMVvh0cg5P2rhDDIWhhYYWAnMh3ABKnvY2B9Bfyb8dKaICl931RJJkJdiKT5GNhy+AOpnUx98ZLmY+CfGAzKzi2kAw2vO5NTXQlkTSBZGwPWdM6eENqNgKQFsn/LphXcSQuwPMLA3oAZt8iXBbatnx+OT8uQdlIpO+b68cPipdXb31CJM0FMitetwiDEpMYff5JC99lFs6TvnwzHtbOO6/bmaGGh5Y7j75sv+Ytu7QOyKWPzTBJwfn9IM3UNbUf/XU7ql0m+FnEbWFqXa1be65KEgTWpTTuZt8JF0lMhmWYaGzo+Ma+Ekqr277LGdb/0vhgeRO1C+6d91/s1wYyrR8iGSpXD0FdCGAjYNFMCfyFsjzznz57hnn72RffV00+M2sdGJbJH+3mc/5LSv2vN2k6EIMlXjqVlF0kFioSzkl0pPypR0hG333ns0bI9rtW/a1Qin0dIJ+zlJWQA74Uaz/gEsiaQPKaEuiCaNX2K+5u//37Hbyh0gWwXYPY9lzKrOUceYWDD1llhYMcbOwlJlDHd5bYLfvvuKz9rsmIFhN4gKn9Lqrefjz8RqlmMpisnKnEnBiHCINS/wBeAWQtTGef8oA9aXt8cSv8e5yjsT8BWGFjTUNsH7GbPplc3dwmILFNXjSKX1o5x5Y8re1I4bv/vlm2aMLD90tbDnr5IvwxJjxODdo/1WeN6GcIgtEXyWJGEpmnHpm+uuM2detykVJ/WLN+nQZiHSxlD1nWVEQZ54Tfl90U0fhFpxPEM6UBNaQfqMTgCWRNI1olBnCmhnOCFnhjY53VROGv61K4gAUkL7CxTIt2xtjvYvk9T6ImBXfAntUa/JwaLzpvlvnXd7c4vtwicXoWBLVPa6aV/YuDvvCeZjIrJWNKJgR9PPI6bf2Kg/5YTg7TLj9RkyReKaeVO2vn3y+Uv7Ad1YmBPwOXk3J4YZPHPOoFAGAxuzKxDylnjutTBbviK8FWTNBlXk3xa7N/9zRF5d9WnyJqHvvra6+6Bh9a5ZX/5xcgHTcYFtRTxfWrUNFA2diV93wcsrlxSF9l0FjNUzT8pWpk/RsfVp+ph6FsnDGxH9V++fp2qenmZQ3bk0tJNOrbvpSz+MwiDIiiSRtYEkseUsBcfgxBhYE0D/YVgXAuqEJDvxNxFbLStj0HSiYH1RbChjadP+UjXok0mQA3ZbE8W+vUx8H0rdNJKOjEIOY0MDVOqY++ck4+NnLTt4tj6EUh9bTvLSa2esqjD+dzTjo/8O/y/qx+AbTMpny5I5O8aNc73MZDy6A2s8jsVBlk+BrZ/+eInqR1t3bVORx6034iwqHHpSdmkv6T5GNhw3+pjIwsleSaOf5aPQZxpEsKg3WN7L+O69VWJ86mSDRvrg2VNqq141s2WBV+aGW106AVqeXxqdIy2pkRJvlk6bup4L+OVDaPvjzfqO+r7CKWZiPvplWki3jph0O5XOV/tEQb5ePHreAJZE4i/s5R2c6VdMNqoRGef8im3dt3TsVGJ4oSBOiyr2Yc1DQzxSbI3etodLFksxZ2AWKHQy06ZkvUni7STjpB8ZNdLPrLwm3bUYdGumJpEhkQlCvUv0DYWDuPGjoluLw1tZxEKdqfQmmbavyft4PlRiezzdqPIRgbyJ/ReTV3jzMGEhY2IJb85/cRj3L0/+7dIZFohKL8NaUe7qWXTPvmTR0UXgNod1Dj+afkkRTlCGLR7xM8a19PGQd/UMcl3zI57Mjb5J4S+r1FamPq0QA02il2cb5Z/0uqPDyHjsx+Mxg8qU/aX2q4AACAASURBVKUw9AiDdr/bqbVHGNA5iiCQNYEUkccwHNtsOUOd3Iqom5/GIE8JbV79nmb65R6kXe8gODctzarwZ15pRs/KGtf9Mdn3tUryxdFdexsa154UWkHqj8Nx4XkvOuvUKDpaiDBI8s3SkwkVJknCIO7vdlOhLmHoEQbNeEcHUgsG8IFgbV2iWRNIL0B85820OPC9pO/voPppnH7iZPfSy5tH7GD1mlfe5wblV2TLUXQeVVmY5mXdlN9XhT/zSjN6VNa4nnViEOJTpeZDasqTdGKgf087MbBmedICcaZEftQ4FTd6YpAlDLLMf+oShh5h0Ix3dCC1YAAfCNbWJZo1gbQOCBWGQIsJMK80o/GzxnXfzMf6+sT5WolfjESbu+KGOzuXNPo2+UJOzEB1kf+XX/yMu2blfZ0NmjgfgySfml58DLKEgZTP97myZqR1CUOPMGjGOzqQWjCADwRr6xLNmkBaB4QKQ6DFBJhXmtH4IeO6NaMRP7An1z/fufgwKfpQUth2/5TYRiVKOjFI8mlSPxxpiTxRiUKEgW/qWccw9AiDZryjA6kFA/hAsLYu0ZAJpHVQqDAEWkqAeaUZDV/1cb0s07ksc6LQ1i/alFPzDXn/EAahrdTC34V0oBZioco5CWRNIGUN4GnVSLpxNmfVOz+3YT8l1J2EkBS/iD//3CfdmrXrYqPRhORl2cnx/PV3ro4ue/Qj28Sl5U9g/U5EWRf2hNQn6zd21/DUT01y659/KYpE5Uf8yEqH78sjwLxSHvsic84a14vMq5e0ypxXQsI8Z80/qx96fET44l44+M+EvH8IgyJINzSNkA7U0KpTrQIJZE0gZQ7gBVYzMSl/AW7tTPsVID47m3ZW3ezkJVE/Lll+S+eYP+vZsr63kZKkDDbmeVllIt98BJhX8vGq6q+zxvWqlrvt5Qp5/xAGbe8lKfUP6UDgg0AWgawJRBe3/g2VsgvsL5w1qoN/a23IDk1WPouuvd1tfGWLO+yACW76lCOjGzQ1rrx/46aEzIuzeY2L7W5tTq29qcTdt/UQjhKrW2L8y0dsaDU9a4+r3+kFXXqD51XnfyF6LnSBb8uVxE/57zx2jLvpx2u67h4QJr/+zXPRjaNyN4BEGdGLsPw4/VI2CRmYdB+A34ey6rvX7ru6iRP2cqse/FV08nL1gjOjU4O4+w7UUVHyGDVqlNOyaJ5+O6ZddGnbL+k+Bf27L9psX9YL3eK4SrmS6mFtpv0+kvUeVuV75pWqtER/5cga1/tLnacHRSDk/UMYDIp+A9IN6UANqCZVGDCBrAlEF4vTjj40OjpNuxFXhUHaDZlJ1cnKx9486d+8q6H17GLvmpX3RllJmeX3ITdhyu9tyDr/Jmd7Y62mJ8/oLaF66Zje+Cs31tqoF7qo1Js305rWmg4lnTToAl0Xu/KM1tm/1VdNifSiNHXUs22mtw6L4Em73TmpvnKyYW8k9m9JtVz0ZmC98ffisz7XEVqWi62TChq5WdkXef7lepr3I08807kx2d6c6l+U5wsDMSeL42pvYPbTszcXJ/Eb8Ovcd/LMK30jrEQCWeN6JQpJIUYQCHn/EAZ0nEQCIR0IfBDIIpA1gdj40rLwtZfEbHp1c7To0p17Pw60LIAP3v+D7psrbnOnHjcp1dY8Kx9786Qu4ubPPsktXH7LiPsK/AVk0oVncY5occLAz8cP9WcZ2wVmnDAIMSeyJjljRu+YyC/tkiIrjFSQyInBIRMnjBAr8n1aWnKakPRJWpDvusvOHZHw/j12jU5bPnbIRDdj6kejpFSoSHms0Ejrr2kX1yX5UPh/V6ExfcpHujj4wsD2N19w6smL1kP+X9Kz9ejXDC3rvR3U98wrgyI73HRDxnV/02K4JRyZW9HvTBV9x7IYh7x/CIMsii3+PqQDtRgPVQ8kkHcCCRUGagojC6YQp1vftCMtH51AzjnteDf30hXO30G2pjIWgzUvkb/H3YQZJwzi8rELTmtaIumquUuvwiDUv8CfSLMWsLKg3WfP8c4uepWPbx4kf0+6mC6pvkknBioM1AxL8xRzLX+B7ndba96k31kzLvlbklCL+7u2b5YwsIJXueoNq1bghKQX4mwe+LoO/GfMKwNHPJQM8o7rQynUEDOpqu9YFoKQ9w9hkEWxxd+HdKAW46HqgQRCJhC7E2p39v0TA9/sQ+zp999nL7f7buM6O8VJxfJPDNLyCTkxCHF8LeLEwF9oF3Fi4DtAJ0XA8Hf57b/lxMDf2c46MbCL4aR2istTn8sSBv7NqJJHmnO7v7Af5ImBbyLnn1Dpv+O4Sj3ShAbCIHAw4meFEQgZ1+XEAN+x4fqOZTVwyLoOYZBFscXfh3SgFuOh6oEEQiYQazvtL6DU1l5MR8RG/ciD9uuEcROhcNcDD3ecT9OKpDvDcTdhqjOomiz5tv+Srr1x86KvnOJW3vfzKDv9u9q0+7bpfhjQvD4GIo500ShmP2IyIx8pay8+BnnClOouv+6g+z4GccLgmCMOjtpJT1m0Pf1TEWtLb02JrDDw6xvqYyC/kzKIUEgybYo7CfDra/uT7ZeavtRx/QsbE30MtE+IuZttNw1ZG8c1zcfAmmYUbRYR+Dr3/TPmlb4RViKB0HEd37G73TB9x7I6R8j7hzDIotji70M6UIvxUPVAAiETiN1Z+qOJf9AVOUbNSuTvU4442G3ZurUjDPIsjnTneNzYMVHkH5uPn479t1QzLlqQH83GNyNSPGryJAJCPknCICQfMb059/QT3Mp7f+4uOuvUKD1ZAMsnNCpRqH+BpCll/8db33Gy/ueHHu+YMMkOdZxtvQqFpOhD9u9JZkSWa1x99aRGzYfWrnsmNiqRmBGpaEuzdbYmTvKMfKzgsd08LlqQfJ/0d41iJPU4/cTJTiJIiaATYZDENSm9tChHnBgEDkb8rDACIeO6PQkONRHVjRR8x1ZGY7xumoT4joU0bsi6DmEQQrKlvwnpQC1FQ7VzEMiaQHIkNeKneS7VKuu+hKJuwgzhVNTkYUVNiPlPSNn4zf8QyCNom8aNeaUZLZo1rufx6fI3S8S0Ed8xhEEz3pSG1YIBvGENWlJ1siaQXoqlu8ryrJr/+Dv4Nl3Zsb3gCydFsfjtLkwveffyTMg9C72ka58ZRB5tXsD22x5pz7eZK/PKIHvW8NLOGtfz+HThOza2686eXoNKhLR+yPs3tBODkALzm+oRuH/RKVGhJk+eXL3CUaJaEMiaQGpRCQoJAQgUQiBkYVJIRiQyUAJZ43qWTxe+Y8X6joU2dsj7NxRhoAVetWpVaNn5XcUIIAwq1iA1Kk7WBFKjqlBUCECgTwIhC5M+s+DxIRDIGtf9m+bxHeu+0T7Llyqv71hok4e8fwiDUJot/x3CoOUdoI/qZ00gfSTNoxCAQM0IhCxMalalVhZ3kOM6vmPdXapI37GQ9w9h0MpXOn+lEQb5mfHEOwQGOYGUyTjNmdmGtoyLGKORN6T8Wc69cZeklVnvrLxDy+uHcc1Kl++bQSBkYdKMmja7FoMY1/EdG9lnivYdC3n/hioMmv2aUDsIQCCOwCAmkKqTThMGfvjSrLqELrSz0hnW96HltWFT7T0Gwyon+ZRDIGRhUk7JyDUPgTaO63n4VPW3Ie8fwqCqrUe5INAQAiETiCwSxaZS4rxbW9SkePh+VBcNdzdn5ieiOwckjRW3rXLPbdjk7P0CaenZuPJJMfZtk8TFlZ+14Conz0768IFuy2tbOxGT9Dk/dKl/kdqWrdvc6l8+HpX7U5MOd/Nnn+QWLr8lunchqUw2Dr88o1GaZEf+1795zj3w0LouBlqWtOd2HjPG3ffgo1F7SGx/ieu/4Mobu8og6fu/U/5aXj8SlC8Gijwib8jr0vhqhCxMGg+hARUMGdcbUM3GVSHk/UMYNK7ZqRAEqkUgawLR42O5pVZuDdZF/qzpUzu3186Y+tHoEil7g6Q1wfGFgdxGKxd+vfy7LU4v2dGbkyWfuPRkUS830eotte/b/T2di9TiiFphIPkk3XJrTYn8Y2FfGPzisac65db09ttnL6eXeu37gT27iqJC5+KzPjei3JaXb85kQwnqRWFaX3nOL4feXurffKy3TitbuQnYllfT1rb1T0uKPiavVs+nNHEEQhYmkKs+gaxxvfo1aGcJQ94/hEE7+wa1hsDQCGRNIEm2+v7f7WJ206ubu2zzfWHwsUMmRot/+agj2yETJzh7A66f3uLrftS5cdleuJMEypbvqfXPd5UnKU69vzD2hYHkJbf1WrGUJgzibmzWelyz8p0bi/XGZVsP39QpqRy+WZDlYkWCpG0FnRUy9lTA9yvAnGhor2FlMgpZmFSmsBQkkUDWuO4/KGOMHWMHiXaYl1lmOUpnnYqmmV4O4nLMkPcPYTDI3knaEIBApvNx0oTh/90OoFnCQHeorTDYZ8/xXRNTSHpxi2ptUjv53P/go+6njzzRMeNJEga+4PAX5GKyI4ImVBj4C3xbJhEGmp7fDeU5MQ2yHzXhss9lCQObfpIw0FOBz5/0p+5b1/3IzZt9klOfAoRB+waIkIVJ+6jUr8YIg3faLE0YhJyIZvlkhaSRp/eEvH8IgzxE+S0EIJCbQNYEUsSJge5eq427nhjoAlv+nXViEGeaFCoMijoxyCsMsk4M0oTBM8++GHuaYCe6LGFgTyS0DcQEzJ4Y6K7Xvnu/17246dWuPBEGuV+n2j8QsjCpfSVbUIGscV0QWJ8u8VVSE0XZGIjzcZJn0nzE7DPyWzH9FPNT+fuia293G1/Z4g7a7wPu+ZdecY8++e+Rr9nXZk13y37wT+6wA/Z1S264I/KTEtNLOb1QHyod5/P6oD3yxDOdDRbry6bN75+Qyr+X3/rOSa7w8H2ylvzFGe6alfe5V197PfIN0/oVGcEt5P1DGLTgBaaKECiTQNYEoo7HYp+uPgay+37Oace7uZeucEk+AXE3Z+pAK/UVJ1y5Wj7Ux6AfYSD5ifO01EF9FLQMeXwM8gqDLB+DJGFgzajEb8H6Iyy9+e7OSUOWMNCJ3tY/zvRJTyh0otP+WPRuWJn9nLzDCIQsTMJS4ldlEggd1+34bX2XdFy2Pk46fsf5iEld9RkZs+xpqYgAnQ/kuzj/rzknH9vxLVPfKJum9ZOSecj3aUvyQUs6MfA3PewmjuT7zRW3uVOPm+QkX91I8f29tH3zRrJL6xch7x/CoMw3i7wh0AICWROIv7MUEpVIntHdF/n9lCMOdlu2bu3swIgjre7M2MVo3ihHoScG/g7Y/Nkz3NPPvui+evqJzgqDrKhEccJAhcbadc+4qxec6XwH5LToQknCQPjZ5yzzPCcGNoqS7pjpKY0tb9KpUJb9bQtej9ZVMWRh0jooNaxw1rge5yOmPl5JppcahS3OR0x9xuxiWTdzRBhY/wVfGPiCQk9LfXPSpDTkRDjJBy1JGKQFmrBzgi2DCgNbf50fr79z9Yj5pJduE/L+IQx6IcszEIBAMIGsCSQ4oYAf+hGOAh4Z+k+K3P0ZeuFNhllOd7Zscc7cmBGV2Xrl5R2yMCmvdOQcSiBrXPd9xHyfrDgfp0XnzXLfuu726JRYdu11A0g3OKwpjnyn4ZlFGNgTX18Y2KATdizyhYGcCtiPhohO8mmTjaOkcTDOz8z6dulGSpwwsPVHGIT2SH4HAQjUhkDWBFJkRYoUBpqWxOT3PyH3HKTVq0ib0SL55UkrVBjY8Kf2IrMmMMjDi9++QwBh0IyekDWuZ50YxPk4WZ8wG4RBdtD94BF24V2UMEi6hT7p3pw0YZBmJpkUXMIP76w9pUiTy5D3jxODZryj1AIClSWQNYFUtuAUDAIQKJxAyMKk8ExJsHACWeO6v8j370ex5j1qz6+mRFJY30dMdu3VnGfM6B0jJ2X9XRHCwPcxkN19EQpyH44fXMIP2xxnshl3maOKIWtSGudj4J8YFHnKHPL+IQwKf11IEAIQsASyJhBoQQAC7SEQsjBpD4361jRkXLc+XWef8in35PrnO+GK43ycdMEf5yNmT3DlxPbc009wK+/9ubvorFNHLNw1oIXQ1ahE8js5sUwyJVKnZblYUm6et6fCaScGah6UFZXIP4GO88nSqES+MCjydDXk/UMY1Pe9pOQQqAWBkAmkFhWhkBCAQN8EQhYmfWdCAgMnMIhxvUhT0IEDCMigCBOgItKwRQ15/xAGAY3LTyAAgd4JDGIC6b00PAkBCJRJIGRhUmb5yDuMwCDG9aYJAyHZT+Q1bj4O64v8CgIQqBmBQUwgNUNAcSEAgf8mgDBoRldgXK9nO4a8f5wY1LNtKTUEakOACaQ2TUVBITBwAiELk4EXggz6JsC43jfCUhIIef8QBqU0DZlCoD0EmEDa09bUFAJZBEIWJllp8H35BBjXy2+DXkoQ8v4hDHohyzMQgEAwASaQYFT8EAKNJxCyMGk8hAZUkHG9no0Y8v4hDOrZtpQaArUhwARSm6aioBAYOIGQhcnAC0EGfRNgXO8bYSkJhLx/CINSmoZMIdAeAkwg7WlragqBLAIhC5OsNPi+fAKM6+W3QS8lCHn/EAa9kOUZCEAgmAATSDAqfgiBxhMIWZg0HkIDKsi4Xs9GDHn/EAb1bFtKDYHaEJAJhA8EIAABJXD/olOi/5w8eTJQakoAYVDPhkMY1LPdKDUEGklg1apVjawXlYIABHojgDDojVsVnkIYVKEV8pcBYZCfGU9AAAIDIoAwGBBYkoVATQkgDGracM45hEE92w5hUM92o9QQaCQBhEEjm5VKQaBnAgiDntGV/iDCoPQm6KkACIOesPEQBCAAAQhAAAIQgEASAYRBPfsGwqCe7UapIQABCEAAAhCAQGUJIAwq2zSpBUMY1LPdKDUEIAABCEAAAhCoLAGEQWWbBmFQz6ah1BCAAAQgAAEIQKCeBBAG9Ww3Tgzq2W6UGgIQgAAEIAABCFSWAMKgsk3DiUE9m4ZSQwACEIAABCAAgXoSQBjUs904Mahnu1FqCEAAAhCAAAQgUFkC3Ghf2abJLFjWzeOjtm/fvj0zFX4AAQhAAAIQgAAEIAABQ6Bp99O8/vrr7pJLLnHz5s1z73rXuxrd1kn3iCAMGt3sVA4CEIAABCAAAQgMhkDThMGyZcvczTff7GbOnOnOPPPMwUCrSKoIg4o0BMWAAAQgAAEIQAACTSDQJGEgpwXTp093b775pttxxx3dypUrG31qgDBowhtIHSAAAQhAAAIQgAAECicwf/58t3jxYvfGG2+4nXbayc2dO9ctXLiw8HyqniCmRFVvIcoHAQhAAAIQgAAEIDAwAps3b3Z77LGH27ZtWyeP0aNHuw0bNrhx48YNLN8qJowwqGKrUCYIQAACEIAABCAAgaEQsKcFmmFbTw0QBkPpcmQCAQhAAAIQgAAEIFA1AnJaMH78eLfDDju4sWPHuo0bN0b/Fp+Dt956K/p3m04NEAZV66GUBwIQgAAEIAABCEBgKAQWLVrkvv71r0dhSs8++2w3atQoJ5H8lyxZEoUt/cY3vuHOPffcoZSlCpkgDKrQCpQBAhCAAAQgAAEIQKB0AioMSi9ISQVAGJQEnmwhAAEIQAACEIAABKpFAGHAzcfV6pGUBgIQgAAEIAABCECgFAIIA4RBKR2PTCEAAQhAAAIQgAAEqkUAYYAwqFaPpDQQgAAEIAABCEAAAqUQQBggDErpeGQKAQhAAAIQgAAEIFAtAggDhEG1eiSlgQAEIAABCEAAAhAohQDCAGFQSscjUwhAAAIQgAAEIACBahFAGCAMqtUjKQ0EIAABCEAAAhCAQCkEEAYIg1I6HplCAAIQgAAEIAABCFSLAMIAYVCtHklpIAABCEAAAhCAAARKIYAwQBiU0vHIFAIQgAAEIAABCECgWgQQBgiDavVISgMBCEAAAhCAAAQgUAoBhAHCoJSOR6YQgAAEIAABCEAAAtUigDBAGFSrR1IaCEAAAhCAAAQgAIFSCCAMEAaldDwyhQAEIAABCEAAAhCoFgGEAcKgWj2S0kAAAhCAAAQgAAEIlEIAYYAwKKXjkSkEIAABCEAAAhCAQLUIIAwQBtXqkZQGAhCAAAQgAAEIQKAUAggDhEEpHY9MIQABCEAAAhCAAASqRQBhgDCoVo+kNBCAAAQgAAEIQAACpRBAGCAMSul4ZAoBCEAAAhCAAAQgUC0CCAOEQbV6JKWBAAQgAAEIQAACECiFAMIAYVBKxyNTCEAAAhCAAAQgAIFqEUAYIAyq1SMpDQQgAAEIQAACEIBAKQQQBgiDUjoemUIAAhCAAAQgAAEIVIsAwgBhUK0eSWkgAAEIQAACEIBASwlMmTLF3X///aXV/phjjnH33XdfafmXnfGo7QiDstuA/CEAAQhAAAIQgAAEnHNt37EvuxMgDMpuAfKHAAQgAAEIQAACEIgIIAzK7QgIg3L5kzsEIAABCEAAAhCAwH8TQBiU2xUQBuXyJ3cIQAACEIAABCAAAYRBJfoAwqASzUAhIAABCEAAAhCAAAQ4MSi3DyAMyuVP7hCAAAQgAAEIQAACnBhUog8gDCrRDBQCAhCAAAQgAAEIQIATg3L7AMKgXP7kDgEIQAACEIAABCDAiUEl+gDCoBLNQCEgAAEIQAACEIAABDgxKLcPIAzK5U/uEIAABCAAAQhAAAKcGFSiDyAMKtEMFAICEIAABCAAAQhAgBODcvsAwqBc/uQOAQhAAAIQgAAEIMCJQSX6AMKgEs1AISAAAQhAAAIQgAAEODEotw8gDMrlT+4QgAAEIAABCEAAApwYVKIPIAwq0QwUAgIQgAAEIAABCECAE4Ny+wDCoFz+5A4BCEAAAhCAAAQgwIlBJfoAwqASzUAhIAABCEAAAhCAAAQ4MSi3DyAMyuVP7hCAAAQgAAEIQAACnBhUog8gDCrRDBQCAhCAAAQgAAEIQIATg3L7AMKgXP7kDgEIQAACEIAABCDAiUEl+gDCoBLNQCEgAAEIQAACEIAABDgxKLcPIAzK5U/uEIAABCAAAQhAAAKcGFSiDyAMKtEMFAICEIAABCAAAQhAgBODcvsAwqBc/uQOAQhAAAIQgAAEIMCJQSX6AMKgEs1AISAAAQhAAAIQgAAEODEotw8gDMrlT+4QgAAEIAABCEAAApwYVKIPIAwq0QwUAgIQgAAEIAABCLSPwKJFi9zXv/51d8kll7izzz7b6YnBkiVL3Lx589w3vvENd+6557YPTEk1RhiUBJ5sIQABCEAAAhCAQNsJbN682Y0fP9793u/9nnvXu97lNm7cGP379ddfd2+99Vb073HjxrUd09DqjzAYGmoyggAEIAABCEAAAhDwCcyfP98tXrzYvfHGG52vdtppJzd37ly3cOFCgA2RAMJgiLDJCgIQgAAEIAABCECgm4CcGuyxxx5u27ZtnS9Gjx7tNmzYwGnBkDsLwmDIwMkOAhCAAAQgAAEIQKCbgD014LSgvN6BMCiPPTlDAAIQgAAEIAABCDjn7KkBpwXldQmEQXnsyRkCEIAABCAAAQhA4L8JyKmBRCmSKET4FpTTLRAG5XAnVwhAAAIQgAAEIAABQ0BODc444wz3ne98B9+CknoGwqAk8GQLAQhAAAIQgEDzCIw64ovNqxQ1qhWB7Q/+fc/lRRj0jI4HIQABCEAAAhCAQDcBEQaP3nIZWCBQCoGDTzrHIQxKQU+mEIAABCAAAQhAAGFAH6gOAYRBddqCkkAAAhCAAAQg0HICnBi0vAOUXH0RBvcvOiUqxeTJk3OXBlOi3Mh4AAIQgAAEIAABCMQTQBjQM8okgDAokz55QwACEIAABCAAAUMAYUB3KJMAwqBM+uQNAQhAAAIQgAAEGi4Mnv6PF9yXLlzm5px8rJsx9aNDb+//3PaGW3rz3W7W9Klu113GDj3/OmWIMKhTa1FWCEAAAhCAAAQaTYATg+Kb97Lrbne/eOwpd9X5X0AYZOBFGBTf/0gRAhCAAAQgAAEI9ESgicLAnhhMO+ow91d/d1PEZuexY9xNP17j/mjiH0SL9jGjd+x899d/drL7/dE7Obuol2e+fPG33b8+8Zvo+Wsu/LI74qD9nab/mY//sbvvwUej72fPmOrOOe0Ed+u9P3MLrrwx+v3799jNXb3gTLfvB/aM0l1+673R3z816XCn+T342JNu1oKroufl+0mHH+CeWv9i57myTz966lQ5HkIY5IDFTyEAAQhAAAIQgMAgCbRFGKxd90y02JaPNTOShfz37/mXSCioEDjyoP3cnJmf6BINd61Z65bedE9XGocdMCFa4NvvVATYE4M4sSF5iJCwwkD+7QsBKZ/mK2k37YMwaFqLUh8IQAACEIAABGpLoC3CYP0LG0cs/u1C/OKzPhe1oezey8nAbu8e1yUgdMEuv9Pvph19aNfiXk8U0oSA5GEX+5te3dzJU04jxD9BTzjmzT7JXbL8lqhcesJQ246WUHCEQdNalPpAAAIQgAAEIFBbAm0RBrq43rrtzcg8SHfsdSH+vt3fE7Wh7vQ/tf75aMHufy7888+6QyZO6BINuusfJwxe/t2WEY7QacJAhYOcYsw97Xh3/uXfK82JehidGmEwDMrkAQEIQAACEIAABAIItF0Y2B388e/Z2X3k4P2jU4A0237/uzRhIOlbIWLzE9Mm/8RAvtf05UTirgce7vgbBDRn7X6CMKhdk1FgCEAAAhCAAASaSgBh8D8L8ec2bOo4GFuTHt+PQPqC9VNIEwYSrjTEx0BPGyRtzfuO1Q91OSo3sQ8iDJrYqtQJAhCAAAQgAIFaEkAY/M9CXP0Q9O6Bl3/3WmpUIr0nwRcG+m/pENa8SKMSaQQj+d5/VjuRRjcS06Uy7mIYVmdGGAyLNPlAAAIQgAAEIACBDAJNFAZNaPSmRyPSNkIYNKG3UgcIQAACEIAABBpBoInCQBabdf3cdvm8yExJzJr0tEDr8+gtl9W1WonlRhg0rkmpEAQgAAEI/WTjDwAADTdJREFUQAACdSXQRGFQ17ZoY7kRBm1sdeoMAQhAAAIQgEAlCSAMKtksrSkUwqA1TU1FIQABCEAAAhCoOgGEQdVbqNnlQxg0u32pHQQgAAEIQAACNSKAMKhRYzWwqAiDBjYqVYIABCAAAQhAoJ4EEAb1bLemlBph0JSWpB4QgAAEIAABCNSeAMKg9k1Y6wogDGrdfBQeAhCAAAQgAIEmEUAYNKk161cXhEH92owSQwACEIAABCDQUALf/va3G1ozqlV1Ah/60Ifc0Wdf6+5fdEpU1MmTJ+cu8qjt27dvz/0UD0AAAhCAAAQgAAEIjCCwZs0aJws0PhAog8BuH5+HMCgDPHlCAAIQgAAEIAABnwCmRPSJMglgSlQmffKGAAQgAAEIQAAChgDCgO5QJgGEQZn0yRsCEIAABCAAAQggDOgDFSGAMKhIQ1AMCEAAAhCAAAQgwIkBfaBMAgiDMumTNwQgAAEIQAACEODEgD5QEQIIg4o0BMWAAAQgAAEIQAACnBjQB8okgDAokz55QwACEIAABCAAAU4M6AMVIYAwqEhDUAwIQAACEIAABCDAiQF9oEwCCIMy6ZM3BCAAAQhAAAIQKPnE4OXfvea+fPG33dzTjndHHLR/7vZ4+j9ecJdes9JddNapbtddxiY+/+BjT7rv3/Mv7q//7GT3+6N36vwuK39J//o7V7uvfHaaW7j8FveZY/84tZy2PC//bkv07FdPP7ErT838sutuj/7znNNO6JTHPi9/FDZHHrRf129uvfdn7plnX4z+Fld/qeusBVd10pw9Y2rX86H5/usTv+ni+f49dnNXLzjT7fuBPXO3U8gDCIMQSvwGAhCAAAQgAAEIDIFAGScGWQvzrGqHCoOkdNLy/89tb7hvrrjNnXrcpODFsF8eWcTLZ8bUj44oQugC/aWXN3ctyNOEgXy39KZ7Or+XOvzV390U5a2iKDTfXsVaVpslfY8w6JUcz0EAAhCAAAQgAIGCCWQJA130HnbAvm7JDXc4u4Msu9SLrr3dbXxlizvsgAnRIvTRJ/+9s3P9qUmHdxamuli9Y/VDTv6+/oWN0YnBbu8e17X77+/yS/5funCZe27Dpui5c0473s29dIWTne0/mvgH7qrzv5B4amDTEmyyWPbz908s5JnVDz0e7bZrmeXEQMs5buyYKA35XHPhl91+++wV7fDb8sh3lyy/xc2bfdKIsoUs0C+4/Hq313t3c1te29rhlyQMJK+40xdf/ITk288pTq/dEmHQKzmegwAEIAABCEAAAgUTCBEGsjCfdvSh0WJZFqg/feSJjgg4//LvdXaqdRF/8Vmfcwfv/8FoIf6+3d8TPScL09++9EqXeJCFdZow2LrtTSeL5PNmTY927yWNCXu/1x0ycUJuU6KlN98dm78vDCSPSYcfGJkO+cJAOMw5+djoJMDW57kNL48oj03HNlnIAl3q/OXPTnPX3n5/x4wpSRiI6VKSWZXNKyRfhEHBLxfJQQACEIAABCAAgToRCBEGF1xxg7voK6dEi3PZidbF+qZXN7vF1/2os2vv7/bLv+X7RefNct+67vbOIjduJ179BWwacvoQ5yMQakqkac2ffVKXr4DN3woD34zIL6flYMsZJwySzIlCFuhxfO9/8NFYH4On1j/f1Qa271kxEZKvnnzYNOypzyD6NScGg6BKmhCAAAQgAAEIQKAHAiHCwO5I+8LALtztaYI4++oC/mtnzHDnX3F9x9k4VBjctWZt53TCOg/nFQZqfqT280nCwNZNRFBoOQclDOwpiTRtnPMxJwbbt2/vod/zCAQgAAEIQAACEICARyBEGNidclmU67/lxMAKgyJODHxTpSqdGFiB1OuJgd3F16bQkxXxl5CPNZ9S3h/abx+385jRI6ISye+tCZCIm3978t/dH+7/wa6/h+SLKRHDAwQgAAEIQAACEGgxgRBhYG3r0xbuaT4G/nMSWlN9DCR965cgzSGOzLITb0WJmsNMn/KR3D4G9vRBHaQl/zw+BnmEQZKPgYgA65ehpxLqi+GfWggLYbfgyhudhiCNi4KkUYl23WVnd+WNd7mbfrym83tJIyRfhEGLBwKqDgEIQAACEIAABEKEgSyINRqPjQQUd0+AjaefFpVo57Fj3LSjDo0W5rrwlYhHp584OYrwo2E249ITp2RZxMqn16hENn/bC7KiEsX5QvjlkfSSohLpIj3pzoE4YeCLh5B7DObPnuFW3LaqEy1KTLHS7jrQKEb+PQZS3jgBVdSbg49BUSRJBwIQgAAEIAABCPRJIFQYZF0m1mcxKvN4L/cY+IVPu8egMhWtSEEQBhVpCIoBAQhAAAIQgAAE6iwM7B0Hfkv2E01Hbz5Our04rdf082wbeyPCoI2tTp0hAAEIQAACEKgkgSxhUMlCU6jGEEAYNKYpqQgEIAABCEAAAnUngDCoewvWu/wIg3q3H6WHAAQgAAEIQKBBBBAGDWrMGlYFYVDDRqPIEIAABCAAAQg0k0CVhIGNtmMv7kq7xCutVdTe/yufndZ183HSM37+19+52vXiZ9DMnjKYWiEMBsOVVCEAAQhAAAIQgEBuAlUVBrvuMrZTl9Cbjm3le4kuFHc/gKQ5Y+pHc3PlgTACCIMwTvwKAhCAAAQgAAEIDJxAiDCw8e8lPv6ateuchC99av3zXTcf29t1/bj4/uVcei+CVFDi5O+3z17R3QQSR1/uSvjarOlu2Q/+yX3tjBnu/Cuu7/xdbvQ9YN+9O4v1pNCgWfcRZOWvtxCn3Ucw8MZpQQYIgxY0MlWEAAQgAAEIQKAeBLKEgS7w5552fHQZmdzo+4vHnoouFksSBnNmfsL91d/d5D52yMRoAS878XqDsVCxNylLer996ZXOTcd6u3CSKZHNU9L65orb3KnHTXL7fmDPLuD25mG9IOwzx/6x2+3d44Ly1xOLpBuM69G61S8lwqD6bUQJIQABCEAAAhBoCYEsYSA774uv+1HnhmFrbpN2YmDx2dt85e8qEmQxb29Pfm7Dyy5LGMjzuosv4iHOD8A3I/KFQUj+Kgy4rGywLwLCYLB8SR0CEIAABCAAAQgEEwgRBt+/51+iHf3fH71TtPuvi/c0YWDNj6Qw799jN3f1gjOjcunzsvjOKwzkGd3FX//Cxig93wfAChERH74wCMkfYRDchfr6IcKgL3w8DAEIQAACEIAABIojECIMQk8MZMEun1nTp0b+Amp+5J8YhCzM06ISiZi4a83DUV5xZkRZJwYh+SMMiutjaSkhDIbDmVwgAAEIQAACEIBAJoEsYaA+BmKfLzvzvo/B+Zd/LzoJ2HWXnSMxcORB+40QBmKOs/Smewo7MdAy7bPn+M5Jhl/RNB+DPMIAH4PMLtTXDxAGfeHjYQhAAAIQgAAEIFAcgSxhIDmJ+ZA4DD+3YZM7+ZNHuedf3BRFJVKznuW33htFEppyxMFuy9at7pzTTnAiBhZceWNUUIlkJNGG1Pk3aWG+ddubkbiQj0Ylknzko38Xp2fNd8Le700MJZoVlUjLb02ZbP5EJSquj3FiMByW5AIBCEAAAhCAAAT6IhAiDGwGvdwp0FcBYx72fQji0u/lHgM/HRyPi265kelxYjB4xuQAAQhAAAIQgAAEggjUTRjo6cWck4/NvHhMbz7u5fbifp4NAs+PIgIIAzoCBCAAAQhAAAIQqAiBvMKgIsWmGA0hgDBoSENSDQhAAAIQgAAE6k8AYVD/NqxzDRAGdW49yg4BCEAAAhCAQKMIIAwa1Zy1qwzCoHZNRoEhAAEIQAACEGgqAYRBU1u2HvVCGNSjnSglBCAAAQhAAAItIIAwaEEjV7iKCIMKNw5FgwAEIAABCECgXQQQBu1q76rVFmFQtRahPBCAAAQgAAEItJYAwqC1TV+JiiMMKtEMFAICEIAABCAAAQg4hzCgF5RJAGFQJn3yhgAEIAABCEAAAoYAwoDuUCYBhEGZ9MkbAhCAAAQgAAEIIAzoAxUhgDCoSENQDAhAAAIQgAAEIMCJAX2gTAIIgzLpkzcEIAABCEAAAhDgxIA+UBECCIOKNATFgAAEIAABCEAAApwY0AfKJIAwKJM+eUMAAhCAAAQgAAFODOgDFSGAMKhIQ1AMCEAAAhCAAAQgwIkBfaBMAgiDMumTNwQgAAEIQAACEODEgD5QEQIIg4o0BMWAAAQgAAEIQAACnBjQB8okgDAokz55QwACEIAABCAAAU4M6AMVIYAwqEhDUAwIQAACEIAABCDAiQF9oEwCCIMy6ZM3BCAAAQhAAAIQ8E4Mpk85wr311ttwgcDQCdyxeq27f9EpUb6TJ0/Onf+o7du3b8/9FA9AAAIQgAAEIAABCIwg8N07fxb9bd26ddCBQCkEPv7hCQiDUsiTKQQgAAEIQAACEIghsGrVKrhAoFQCnBiUip/MIQABCEAAAhCAwDsERBjssMMO4IBAKQTefvttTIlKIU+mEIAABCAAAQhAwCPwwgsvwAQCpRLYc889c+ePj0FuZDwAAQhAAAIQgAAEIACB5hH4/x6RRak79zfCAAAAAElFTkSuQmCC\n",
+      "text/plain": [
+       "<IPython.core.display.Image object>"
+      ]
+     },
+     "execution_count": 28,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "Image('data_diagram.png')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "generic-myanmar",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dir = 'data_0407_0'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "biblical-honolulu",
+   "metadata": {},
+   "source": [
+    "#### Step 1. Based on the `coupons` table, create a `coupon_dates` dataframe which maps a date to all coupons available on that date."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "aquatic-processing",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 1688 entries, 0 to 1687\n",
+      "Data columns (total 7 columns):\n",
+      " #   Column      Non-Null Count  Dtype \n",
+      "---  ------      --------------  ----- \n",
+      " 0   id          1688 non-null   int64 \n",
+      " 1   type        1688 non-null   object\n",
+      " 2   department  77 non-null     object\n",
+      " 3   discount    1688 non-null   int64 \n",
+      " 4   how_many    1688 non-null   int64 \n",
+      " 5   start_date  1688 non-null   object\n",
+      " 6   end_date    1688 non-null   object\n",
+      "dtypes: int64(3), object(4)\n",
+      "memory usage: 92.4+ KB\n"
+     ]
+    }
+   ],
+   "source": [
+    "coupons = pd.read_csv(os.path.join(data_dir, 'coupons.csv'))\n",
+    "coupons.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "bound-mercury",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>type</th>\n",
+       "      <th>department</th>\n",
+       "      <th>discount</th>\n",
+       "      <th>how_many</th>\n",
+       "      <th>start_date</th>\n",
+       "      <th>end_date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>90</th>\n",
+       "      <td>91</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>18</td>\n",
+       "      <td>3</td>\n",
+       "      <td>2010-02-16</td>\n",
+       "      <td>2010-03-15</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1072</th>\n",
+       "      <td>1073</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>50</td>\n",
+       "      <td>3</td>\n",
+       "      <td>2011-12-01</td>\n",
+       "      <td>2011-12-25</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>646</th>\n",
+       "      <td>647</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>38</td>\n",
+       "      <td>5</td>\n",
+       "      <td>2011-02-27</td>\n",
+       "      <td>2011-03-08</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1034</th>\n",
+       "      <td>1035</td>\n",
+       "      <td>buy_more</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>16</td>\n",
+       "      <td>2</td>\n",
+       "      <td>2011-11-04</td>\n",
+       "      <td>2011-11-14</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1291</th>\n",
+       "      <td>1292</td>\n",
+       "      <td>just_discount</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>27</td>\n",
+       "      <td>1</td>\n",
+       "      <td>2012-04-18</td>\n",
+       "      <td>2012-05-17</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "        id           type department  discount  how_many  start_date  \\\n",
+       "90      91       buy_more        NaN        18         3  2010-02-16   \n",
+       "1072  1073       buy_more        NaN        50         3  2011-12-01   \n",
+       "646    647       buy_more        NaN        38         5  2011-02-27   \n",
+       "1034  1035       buy_more        NaN        16         2  2011-11-04   \n",
+       "1291  1292  just_discount        NaN        27         1  2012-04-18   \n",
+       "\n",
+       "        end_date  \n",
+       "90    2010-03-15  \n",
+       "1072  2011-12-25  \n",
+       "646   2011-03-08  \n",
+       "1034  2011-11-14  \n",
+       "1291  2012-05-17  "
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "coupons.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "abstract-absorption",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "coupons.start_date = pd.to_datetime(coupons.start_date, format='%Y-%m-%d')\n",
+    "coupons.end_date = pd.to_datetime(coupons.end_date, format='%Y-%m-%d')\n",
+    "coupons.rename(columns={'id': 'coupon_id'}, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "excellent-webcam",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "coupon_dates = coupons.drop(['type', 'department', 'discount', 'how_many'], axis=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "therapeutic-booking",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Start: 2010-01-01, end: 2013-01-24. 1120 days\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Get the earliest and latest date in the dataset\n",
+    "start = coupon_dates.start_date.min()\n",
+    "end = coupon_dates.end_date.max()\n",
+    "days = (end-start).days + 1\n",
+    "print(f'Start: {start.date()}, end: {end.date()}. {days} days')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "graduate-merchant",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Create a dataframe with row for each day from the earliest to the latest date in the set\n",
+    "all_dates = pd.DataFrame(pd.date_range(start=start, end=end, freq='D'), columns=['date'])\n",
+    "assert days == len(all_dates)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "proved-scope",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-02</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-03</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-04</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2010-01-05</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1890540</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>2013-01-05</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1890541</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>2013-01-06</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1890542</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>2013-01-07</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1890543</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>2013-01-08</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1890544</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>2013-01-09</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>25956 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         coupon_id       date\n",
+       "0                1 2010-01-01\n",
+       "1                1 2010-01-02\n",
+       "2                1 2010-01-03\n",
+       "3                1 2010-01-04\n",
+       "4                1 2010-01-05\n",
+       "...            ...        ...\n",
+       "1890540       1688 2013-01-05\n",
+       "1890541       1688 2013-01-06\n",
+       "1890542       1688 2013-01-07\n",
+       "1890543       1688 2013-01-08\n",
+       "1890544       1688 2013-01-09\n",
+       "\n",
+       "[25956 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Step 1.1: Perform a cross join of `all_dates` and `coupon_dates` - which contains info on validity periods\n",
+    "coupon_dates['key'] = 1\n",
+    "all_dates['key'] = 1\n",
+    "coupon_dates = pd.merge(coupon_dates, all_dates, on='key').drop('key', axis=1)\n",
+    "\n",
+    "# Step 1.2 Drop rows where a date does not fall within the validity period of a coupon\n",
+    "coupon_dates = coupon_dates[(coupon_dates['date'] >= coupon_dates['start_date']) & \\\n",
+    "                            (coupon_dates['date'] <= coupon_dates['end_date'])]\n",
+    "coupon_dates.drop(['start_date', 'end_date'], axis=1, inplace=True)\n",
+    "coupon_dates"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "combined-apache",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "coupon_dates.date = coupon_dates.date.dt.date"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "blind-membrane",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Validate date_coupon is consistent with the original data in terms of coupon validity dates\n",
+    "coupons['days_valid'] = (coupons.end_date - coupons.start_date).dt.days + 1\n",
+    "df = pd.merge(coupons, coupon_dates.groupby(by='coupon_id').count().rename(columns={'date': 'days_valid'}), on='coupon_id')\n",
+    "assert 0 == len(df.loc[df.days_valid_x != df.days_valid_y])\n",
+    "coupons.drop(['days_valid'], axis=1, inplace=True)\n",
+    "\n",
+    "# Validate no coupon has been lost\n",
+    "assert len(coupons) == len(coupon_dates.coupon_id.unique())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "directed-municipality",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "coupon_id    24.0\n",
+       "dtype: float64"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "coupon_dates.groupby('date').count().median()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "aggressive-grave",
+   "metadata": {},
+   "source": [
+    "#### Step 2. Merge `coupon_dates` with `orders` in order to create a dataframe mapping `order_id` to all coupons available on the date that order was made -> `order_coupons_available`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "recovered-decrease",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 324683 entries, 0 to 324682\n",
+      "Data columns (total 3 columns):\n",
+      " #   Column       Non-Null Count   Dtype \n",
+      "---  ------       --------------   ----- \n",
+      " 0   id           324683 non-null  int64 \n",
+      " 1   customer_id  324683 non-null  int64 \n",
+      " 2   order_date   324683 non-null  object\n",
+      "dtypes: int64(2), object(1)\n",
+      "memory usage: 7.4+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "orders = pd.read_csv(os.path.join(data_dir, 'orders.csv'))\n",
+    "orders.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "attempted-palestine",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>customer_id</th>\n",
+       "      <th>order_date</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>321942</th>\n",
+       "      <td>321943</td>\n",
+       "      <td>111</td>\n",
+       "      <td>2012-12-21 16:46:04</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>259661</th>\n",
+       "      <td>259662</td>\n",
+       "      <td>544</td>\n",
+       "      <td>2012-05-27 12:58:35</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>263595</th>\n",
+       "      <td>263596</td>\n",
+       "      <td>637</td>\n",
+       "      <td>2012-06-09 13:50:04</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>76819</th>\n",
+       "      <td>76820</td>\n",
+       "      <td>856</td>\n",
+       "      <td>2010-09-16 10:39:13</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>439</th>\n",
+       "      <td>440</td>\n",
+       "      <td>776</td>\n",
+       "      <td>2010-01-02 13:10:04</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            id  customer_id           order_date\n",
+       "321942  321943          111  2012-12-21 16:46:04\n",
+       "259661  259662          544  2012-05-27 12:58:35\n",
+       "263595  263596          637  2012-06-09 13:50:04\n",
+       "76819    76820          856  2010-09-16 10:39:13\n",
+       "439        440          776  2010-01-02 13:10:04"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "orders.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "informed-lindsay",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "orders.rename(columns={'id': 'order_id', 'order_date': 'date'}, inplace=True)\n",
+    "orders.date = pd.to_datetime(orders.date, format='%Y-%m-%d').dt.date"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "static-skill",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupons_available = pd.merge(coupon_dates, orders[['order_id', 'date']], on='date', how='left')\\\n",
+    "                            .dropna().drop(['date'], axis=1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "suburban-venture",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>coupon_id</th>\n",
+       "      <th>order_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>1</td>\n",
+       "      <td>2.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1</td>\n",
+       "      <td>3.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1</td>\n",
+       "      <td>4.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>1</td>\n",
+       "      <td>5.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7624415</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>324679.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7624416</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>324680.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7624417</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>324681.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7624418</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>324682.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7624419</th>\n",
+       "      <td>1688</td>\n",
+       "      <td>324683.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>7624186 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         coupon_id  order_id\n",
+       "0                1       1.0\n",
+       "1                1       2.0\n",
+       "2                1       3.0\n",
+       "3                1       4.0\n",
+       "4                1       5.0\n",
+       "...            ...       ...\n",
+       "7624415       1688  324679.0\n",
+       "7624416       1688  324680.0\n",
+       "7624417       1688  324681.0\n",
+       "7624418       1688  324682.0\n",
+       "7624419       1688  324683.0\n",
+       "\n",
+       "[7624186 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupons_available"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "announced-mandate",
+   "metadata": {},
+   "source": [
+    "#### Step 3. From the `order_details` table, select only `order_id` - `coupon_id` pairs, resulting in `order_coupons_used` - a dataframe mapping an order to the coupons used in that order, if any."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "naughty-denver",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 2395354 entries, 0 to 2395353\n",
+      "Data columns (total 7 columns):\n",
+      " #   Column            Dtype  \n",
+      "---  ------            -----  \n",
+      " 0   id                int64  \n",
+      " 1   order_id          int64  \n",
+      " 2   product_id        int64  \n",
+      " 3   quantity_ordered  int64  \n",
+      " 4   original_price    float64\n",
+      " 5   buy_price         float64\n",
+      " 6   coupon_id         float64\n",
+      "dtypes: float64(3), int64(4)\n",
+      "memory usage: 127.9 MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "order_details = pd.read_csv(os.path.join(data_dir, 'order_details.csv'))\n",
+    "order_details.info()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "human-candy",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>product_id</th>\n",
+       "      <th>quantity_ordered</th>\n",
+       "      <th>original_price</th>\n",
+       "      <th>buy_price</th>\n",
+       "      <th>coupon_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1078232</th>\n",
+       "      <td>1078233</td>\n",
+       "      <td>145447</td>\n",
+       "      <td>1962</td>\n",
+       "      <td>5</td>\n",
+       "      <td>4.01</td>\n",
+       "      <td>1.7243</td>\n",
+       "      <td>734.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1463383</th>\n",
+       "      <td>1463384</td>\n",
+       "      <td>196697</td>\n",
+       "      <td>2168</td>\n",
+       "      <td>18</td>\n",
+       "      <td>2.03</td>\n",
+       "      <td>2.0300</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1597884</th>\n",
+       "      <td>1597885</td>\n",
+       "      <td>215224</td>\n",
+       "      <td>584</td>\n",
+       "      <td>2</td>\n",
+       "      <td>4.07</td>\n",
+       "      <td>1.3431</td>\n",
+       "      <td>1101.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>73244</th>\n",
+       "      <td>73245</td>\n",
+       "      <td>10317</td>\n",
+       "      <td>1374</td>\n",
+       "      <td>5</td>\n",
+       "      <td>12.18</td>\n",
+       "      <td>4.6284</td>\n",
+       "      <td>46.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>680560</th>\n",
+       "      <td>680561</td>\n",
+       "      <td>92155</td>\n",
+       "      <td>1864</td>\n",
+       "      <td>2</td>\n",
+       "      <td>2.65</td>\n",
+       "      <td>2.6500</td>\n",
+       "      <td>NaN</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "              id  order_id  product_id  quantity_ordered  original_price  \\\n",
+       "1078232  1078233    145447        1962                 5            4.01   \n",
+       "1463383  1463384    196697        2168                18            2.03   \n",
+       "1597884  1597885    215224         584                 2            4.07   \n",
+       "73244      73245     10317        1374                 5           12.18   \n",
+       "680560    680561     92155        1864                 2            2.65   \n",
+       "\n",
+       "         buy_price  coupon_id  \n",
+       "1078232     1.7243      734.0  \n",
+       "1463383     2.0300        NaN  \n",
+       "1597884     1.3431     1101.0  \n",
+       "73244       4.6284       46.0  \n",
+       "680560      2.6500        NaN  "
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_details.sample(5)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "sophisticated-million",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_details.drop(['id'], axis=1, inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "protective-hopkins",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupon_used = order_details[['order_id', 'coupon_id']].dropna().drop_duplicates()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "classified-garbage",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>order_id</th>\n",
+       "      <th>coupon_id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>7.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>3</td>\n",
+       "      <td>22.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>6</td>\n",
+       "      <td>3.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>14</th>\n",
+       "      <td>7</td>\n",
+       "      <td>8.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>17</th>\n",
+       "      <td>7</td>\n",
+       "      <td>7.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2395344</th>\n",
+       "      <td>324681</td>\n",
+       "      <td>1682.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2395345</th>\n",
+       "      <td>324681</td>\n",
+       "      <td>1672.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2395351</th>\n",
+       "      <td>324682</td>\n",
+       "      <td>1677.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2395352</th>\n",
+       "      <td>324682</td>\n",
+       "      <td>1682.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2395353</th>\n",
+       "      <td>324683</td>\n",
+       "      <td>1682.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>400236 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "         order_id  coupon_id\n",
+       "2               3        7.0\n",
+       "4               3       22.0\n",
+       "10              6        3.0\n",
+       "14              7        8.0\n",
+       "17              7        7.0\n",
+       "...           ...        ...\n",
+       "2395344    324681     1682.0\n",
+       "2395345    324681     1672.0\n",
+       "2395351    324682     1677.0\n",
+       "2395352    324682     1682.0\n",
+       "2395353    324683     1682.0\n",
+       "\n",
+       "[400236 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupon_used"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "liked-republican",
+   "metadata": {},
+   "source": [
+    "#### Step 4. Combine `order_coupons_available` and `order_coupons_used`, add column `coupon_used` which stores `True` for coupons used in an order, and `False` for coupons not used in an order."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "beneficial-garbage",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupon_used.set_index(['order_id', 'coupon_id'], inplace=True)\n",
+    "order_coupons_available.set_index(['order_id', 'coupon_id'], inplace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "organized-sewing",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupons_available['coupon_used'] = order_coupons_available.index.isin(order_coupon_used.index)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "ranking-hypothesis",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "order_coupons = order_coupons_available.reset_index()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "reasonable-banking",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "False    0.947504\n",
+       "True     0.052496\n",
+       "Name: coupon_used, dtype: float64"
+      ]
+     },
+     "execution_count": 26,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "order_coupons.coupon_used.value_counts(normalize=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "worth-quarterly",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/training-with-artificial-data/Dockerfile b/training-with-artificial-data/Dockerfile
new file mode 100644
index 0000000..7532a70
--- /dev/null
+++ b/training-with-artificial-data/Dockerfile
@@ -0,0 +1,18 @@
+FROM ubuntu:18.04
+
+RUN apt update && \
+    apt install -y software-properties-common && \
+    apt install -y default-jre && \
+    apt install -y python3-pip
+
+RUN mkdir data-mining
+
+COPY data_0408_0 data-mining/data_0408_0
+COPY *.ipynb data-mining/
+COPY requirements.txt data-mining/
+
+RUN pip3 install -r /data-mining/requirements.txt
+
+WORKDIR /data-mining
+
+CMD jupyter notebook --ip 0.0.0.0 --port 8000 --no-browser --allow-root
diff --git a/training-with-artificial-data/README.md b/training-with-artificial-data/README.md
new file mode 100644
index 0000000..222f62b
--- /dev/null
+++ b/training-with-artificial-data/README.md
@@ -0,0 +1,36 @@
+# Data preparation and training - artificial data set
+
+
+This directory contains jupyter notebooks for preparing data and training a model for the coupon recommendation service.
+
+`01_data_prep.ipynb` - data preparation. This notebook contains data cleaning, feature engineering, as well as merging datasets. It requires the original dataset to be present in a directory - oath to the directory is specified at the top of the script.
+`02_training_automl.ipynb` - Training using H2O AutoML.
+`03_training.ipynb` - Training model using scikit-learn. Algorithm (GBM) and parameters are selected based on AutoML result. The notebook compares training on unbalanced and balanced dataset.
+
+In order to run the jupyter notebooks, use (specify `ip` and `port` according to your needs):
+```
+python3 -m venv .venv
+source .venv/bin/activate
+pip install -r requirements.txt
+
+jupyter notebook --ip 0.0.0.0 --port 8000 --no-browser
+```
+
+The notebooks should be run in the order they are numbered.
+
+
+## Docker images
+
+This repository contains a Dockerfile for building a docker image. To build it, use:
+
+```
+docker build -t coupon-rec:0.0.1 .
+```
+
+To run it, use:
+
+```
+docker run -it -p 0.0.0.0:8002:8000 coupon-rec:0.0.1
+```
+
+
diff --git a/training-with-artificial-data/automl_output_about_gbm.txt b/training-with-artificial-data/automl_output_about_gbm.txt
new file mode 100644
index 0000000..a3b150b
--- /dev/null
+++ b/training-with-artificial-data/automl_output_about_gbm.txt
@@ -0,0 +1,191 @@
+Model Details
+=============
+H2OGradientBoostingEstimator :  Gradient Boosting Machine
+Model Key:  GBM_4_AutoML_20210408_164243
+
+
+Model Summary: 
+number_of_trees	number_of_internal_trees	model_size_in_bytes	min_depth	max_depth	mean_depth	min_leaves	max_leaves	mean_leaves
+0		70.0	70.0	578602.0	0.0	10.0	7.571429	1.0	994.0	654.6
+
+
+ModelMetricsBinomial: gbm
+** Reported on train data. **
+
+MSE: 0.2711710163001577
+RMSE: 0.5207408341009544
+LogLoss: 0.8622691624238035
+Mean Per-Class Error: 0.2690251558453727
+AUC: 0.8156996013643698
+AUCPR: 0.8394483144715442
+Gini: 0.6313992027287396
+
+Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.05733060682487723: 
+0	1	Error	Rate
+0	0	1729723.0	1297628.0	0.4286	(1297628.0/3027351.0)
+1	1	513782.0	2513447.0	0.1697	(513782.0/3027229.0)
+2	Total	2243505.0	3811075.0	0.2992	(1811410.0/6054580.0)
+
+Maximum Metrics: Maximum metrics at their respective thresholds
+metric	threshold	value	idx
+0	max f1	0.057331	7.351083e-01	336.0
+1	max f2	0.032548	8.452362e-01	374.0
+2	max f0point5	0.133512	7.645287e-01	255.0
+3	max accuracy	0.095349	7.309765e-01	292.0
+4	max precision	0.880308	1.000000e+00	0.0
+5	max recall	0.007737	1.000000e+00	398.0
+6	max specificity	0.880308	1.000000e+00	0.0
+7	max absolute_mcc	0.139263	4.796984e-01	250.0
+8	max min_per_class_accuracy	0.078557	7.252985e-01	310.0
+9	max mean_per_class_accuracy	0.095349	7.309748e-01	292.0
+10	max tns	0.880308	3.027351e+06	0.0
+11	max fns	0.880308	3.027017e+06	0.0
+12	max fps	0.006237	3.027351e+06	399.0
+13	max tps	0.007737	3.027229e+06	398.0
+14	max tnr	0.880308	1.000000e+00	0.0
+15	max fnr	0.880308	9.999300e-01	0.0
+16	max fpr	0.006237	1.000000e+00	399.0
+17	max tpr	0.007737	1.000000e+00	398.0
+
+Gains/Lift Table: Avg response rate: 50.00 %, avg score: 21.13 %
+group	cumulative_data_fraction	lower_threshold	lift	cumulative_lift	response_rate	score	cumulative_response_rate	cumulative_score	capture_rate	cumulative_capture_rate	gain	cumulative_gain	kolmogorov_smirnov
+0	1	0.010002	0.837540	1.981545	1.981545	0.990753	0.846832	0.990753	0.846832	0.019819	0.019819	98.154496	98.154496	0.019634
+1	2	0.020009	0.828769	1.975119	1.978331	0.987540	0.832717	0.989146	0.839772	0.019766	0.039585	97.511902	97.833106	0.039151
+2	3	0.030006	0.822702	1.969374	1.975347	0.984667	0.825749	0.987654	0.835101	0.019686	0.059272	96.937366	97.534695	0.058531
+3	4	0.040002	0.817128	1.967028	1.973268	0.983494	0.819833	0.986614	0.831285	0.019663	0.078935	96.702794	97.326806	0.077864
+4	5	0.050000	0.812258	1.963398	1.971295	0.981679	0.814630	0.985627	0.827955	0.019630	0.098565	96.339839	97.129454	0.097127
+5	6	0.100008	0.785588	1.956363	1.963828	0.978162	0.799591	0.981894	0.813772	0.097833	0.196398	95.636342	96.382841	0.192776
+6	7	0.150011	0.751169	1.942474	1.956710	0.971217	0.769543	0.978335	0.799029	0.097131	0.293529	94.247368	95.671019	0.287029
+7	8	0.200001	0.346516	1.862231	1.933095	0.931097	0.641898	0.966528	0.759754	0.093093	0.386622	86.223067	93.309509	0.373233
+8	9	0.300003	0.126728	1.329560	1.731916	0.664767	0.164262	0.865941	0.561256	0.132958	0.519579	32.956029	73.191627	0.439145
+9	10	0.400000	0.098188	1.110646	1.576603	0.555312	0.111048	0.788285	0.448707	0.111062	0.630641	11.064569	57.660273	0.461273
+10	11	0.500001	0.078522	0.959023	1.453085	0.479502	0.088037	0.726528	0.376572	0.095904	0.726545	-4.097720	45.308544	0.453078
+11	12	0.600000	0.061271	0.817325	1.347127	0.408654	0.069658	0.673550	0.325421	0.081732	0.808276	-18.267544	34.712655	0.416544
+12	13	0.700001	0.049013	0.705426	1.255455	0.352706	0.054639	0.627715	0.286737	0.070543	0.878819	-29.457405	25.545453	0.357630
+13	14	0.800000	0.040772	0.577073	1.170657	0.288531	0.044764	0.585317	0.256491	0.057707	0.936526	-42.292719	17.065740	0.273046
+14	15	0.900000	0.032451	0.439379	1.089404	0.219685	0.036779	0.544691	0.232078	0.043938	0.980464	-56.062056	8.940442	0.160925
+15	16	1.000000	0.002556	0.195360	1.000000	0.097678	0.024191	0.499990	0.211290	0.019536	1.000000	-80.463982	0.000000	0.000000
+
+
+ModelMetricsBinomial: gbm
+** Reported on cross-validation data. **
+
+MSE: 0.06498795354987083
+RMSE: 0.2549273495525163
+LogLoss: 0.24260618450943383
+Mean Per-Class Error: 0.2770904651860334
+AUC: 0.8021582436506262
+AUCPR: 0.4995332242622148
+Gini: 0.6043164873012523
+
+Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.22919100477943258: 
+0	1	Error	Rate
+0	0	2977495.0	49856.0	0.0165	(49856.0/3027351.0)
+1	1	206636.0	134827.0	0.6051	(206636.0/341463.0)
+2	Total	3184131.0	184683.0	0.0761	(256492.0/3368814.0)
+
+Maximum Metrics: Maximum metrics at their respective thresholds
+metric	threshold	value	idx
+0	max f1	0.229191	5.125079e-01	167.0
+1	max f2	0.098798	5.028615e-01	259.0
+2	max f0point5	0.541164	6.446633e-01	104.0
+3	max accuracy	0.533739	9.262245e-01	105.0
+4	max precision	0.883700	9.166667e-01	0.0
+5	max recall	0.007146	1.000000e+00	398.0
+6	max specificity	0.883700	9.999993e-01	0.0
+7	max absolute_mcc	0.477452	5.089233e-01	115.0
+8	max min_per_class_accuracy	0.077382	7.161830e-01	287.0
+9	max mean_per_class_accuracy	0.098798	7.229095e-01	259.0
+10	max tns	0.883700	3.027349e+06	0.0
+11	max fns	0.883700	3.414410e+05	0.0
+12	max fps	0.005775	3.027351e+06	399.0
+13	max tps	0.007146	3.414630e+05	398.0
+14	max tnr	0.883700	9.999993e-01	0.0
+15	max fnr	0.883700	9.999356e-01	0.0
+16	max fpr	0.005775	1.000000e+00	399.0
+17	max tpr	0.007146	1.000000e+00	398.0
+
+Gains/Lift Table: Avg response rate: 10.14 %, avg score:  9.87 %
+group	cumulative_data_fraction	lower_threshold	lift	cumulative_lift	response_rate	score	cumulative_response_rate	cumulative_score	capture_rate	cumulative_capture_rate	gain	cumulative_gain	kolmogorov_smirnov
+0	1	0.010001	0.816049	8.421531	8.421531	0.853606	0.830672	0.853606	0.830672	0.084220	0.084220	742.153097	742.153097	0.082591
+1	2	0.020000	0.794009	8.171586	8.296564	0.828272	0.805167	0.840940	0.817920	0.081713	0.165933	717.158645	729.656427	0.162393
+2	3	0.030000	0.768283	7.894065	8.162397	0.800142	0.781534	0.827341	0.805791	0.078943	0.244876	689.406521	716.239659	0.239112
+3	4	0.040000	0.728944	7.604888	8.023025	0.770831	0.751144	0.813214	0.792130	0.076046	0.320922	660.488801	702.302462	0.312608
+4	5	0.050000	0.376858	5.845286	7.587469	0.592478	0.588207	0.769066	0.751344	0.058454	0.379376	484.528645	658.746923	0.366527
+5	6	0.100000	0.140183	1.934508	4.761005	0.196082	0.177726	0.482576	0.464537	0.096725	0.476101	93.450753	376.100516	0.418523
+6	7	0.150000	0.116331	1.335611	3.619212	0.135378	0.127054	0.366843	0.352043	0.066780	0.542882	33.561146	261.921178	0.437196
+7	8	0.200001	0.101635	1.145526	3.000784	0.116110	0.108554	0.304159	0.291170	0.057277	0.600159	14.552575	200.078385	0.445293
+8	9	0.300000	0.081609	0.936360	2.312647	0.094909	0.091048	0.234410	0.224463	0.093635	0.693794	-6.364005	131.264731	0.438211
+9	10	0.400000	0.065673	0.765235	1.925794	0.077564	0.073385	0.195198	0.186694	0.076524	0.770318	-23.476472	92.579373	0.412087
+10	11	0.500000	0.053659	0.611310	1.662897	0.061962	0.059244	0.168551	0.161204	0.061131	0.831449	-38.869035	66.289660	0.368833
+11	12	0.600001	0.045913	0.502951	1.469572	0.050979	0.049540	0.148956	0.142593	0.050295	0.881744	-49.704892	46.957168	0.313522
+12	13	0.700000	0.040044	0.417208	1.319235	0.042288	0.042899	0.133718	0.128351	0.041720	0.923465	-58.279224	31.923506	0.248670
+13	14	0.800000	0.034543	0.354182	1.198604	0.035900	0.037320	0.121490	0.116972	0.035418	0.958883	-64.581772	19.860360	0.176804
+14	15	0.900000	0.027748	0.284277	1.097012	0.028814	0.031404	0.111193	0.107465	0.028428	0.987310	-71.572289	9.701187	0.097159
+15	16	1.000000	0.001796	0.126895	1.000000	0.012862	0.019823	0.101360	0.098700	0.012690	1.000000	-87.310507	0.000000	0.000000
+
+
+Cross-Validation Metrics Summary: 
+mean	sd	cv_1_valid	cv_2_valid	cv_3_valid	cv_4_valid	cv_5_valid
+0	accuracy	0.9239439	6.1908754E-4	0.92440516	0.92351615	0.9231228	0.9240608	0.9246143
+1	auc	0.8021638	0.0012602289	0.8009509	0.80137473	0.80144393	0.8033893	0.80366004
+2	aucpr	0.4995746	0.0022548349	0.49709672	0.49797425	0.49965167	0.5002288	0.5029216
+3	err	0.076056145	6.1908754E-4	0.07559483	0.07648387	0.076877184	0.07593916	0.07538567
+4	err_count	51243.8	417.1387	50933.0	51532.0	51797.0	51165.0	50792.0
+5	f0point5	0.6247343	0.0031962688	0.624532	0.6235206	0.62106586	0.62473017	0.6298228
+6	f1	0.51262814	0.0020927442	0.51327825	0.50949	0.5123657	0.5127004	0.51530653
+7	f2	0.43464056	0.0022552824	0.4356685	0.43071926	0.43604782	0.43474048	0.43602678
+8	lift_top_group	8.421473	0.06115962	8.466085	8.333362	8.395673	8.4234	8.488844
+9	logloss	0.24260618	0.0010718547	0.24179175	0.243898	0.24364957	0.24194153	0.24175005
+10	max_per_class_error	0.6053806	0.0024194021	0.6042265	0.6095273	0.603341	0.6052736	0.6045347
+11	mcc	0.5020918	0.0025971201	0.5025298	0.4998116	0.4998053	0.5021768	0.50613576
+12	mean_per_class_accuracy	0.6891337	0.001123119	0.68969077	0.68717736	0.6897318	0.68919575	0.68987286
+13	mean_per_class_error	0.3108663	0.001123119	0.31030926	0.31282264	0.3102682	0.31080425	0.3101271
+14	mse	0.06498795	3.5334285E-4	0.064677946	0.06543439	0.06530259	0.064802855	0.06472199
+15	pr_auc	0.4995746	0.0022548349	0.49709672	0.49797425	0.49965167	0.5002288	0.5029216
+16	precision	0.73137695	0.005741021	0.73002064	0.7328715	0.7233771	0.7312541	0.7393614
+17	r2	0.28652275	0.0020135231	0.2858809	0.28392112	0.28594545	0.28759506	0.28927112
+18	recall	0.3946194	0.0024194021	0.39577347	0.3904727	0.39665905	0.39472643	0.39546534
+19	rmse	0.2549266	6.9268135E-4	0.2543186	0.25580147	0.2555437	0.25456405	0.25440517
+
+See the whole table with table.as_data_frame()
+
+Scoring History: 
+timestamp	duration	number_of_trees	training_rmse	training_logloss	training_auc	training_pr_auc	training_lift	training_classification_error
+0		2021-04-08 17:33:34	4 min 10.858 sec	0.0	0.639457	1.197953	0.500000	0.499990	1.000000	0.500010
+1		2021-04-08 17:33:39	4 min 15.279 sec	5.0	0.616837	1.091785	0.792265	0.824312	1.976849	0.322680
+2		2021-04-08 17:33:43	4 min 19.044 sec	10.0	0.593316	1.024019	0.799346	0.827751	1.969823	0.313274
+3		2021-04-08 17:33:46	4 min 22.648 sec	15.0	0.569878	0.971272	0.802795	0.830288	1.972070	0.311855
+4		2021-04-08 17:33:50	4 min 26.494 sec	20.0	0.551765	0.934923	0.804866	0.831898	1.973781	0.310356
+5		2021-04-08 17:33:54	4 min 30.256 sec	25.0	0.539484	0.910068	0.806598	0.833337	1.976033	0.304712
+6		2021-04-08 17:33:58	4 min 34.008 sec	30.0	0.532110	0.893191	0.808993	0.834841	1.976818	0.303763
+7		2021-04-08 17:34:01	4 min 37.697 sec	35.0	0.527083	0.881022	0.810823	0.836076	1.979631	0.304047
+8		2021-04-08 17:34:05	4 min 41.397 sec	40.0	0.524119	0.873296	0.812040	0.836947	1.980430	0.299840
+9		2021-04-08 17:34:09	4 min 45.047 sec	45.0	0.522347	0.868030	0.813448	0.837918	1.980489	0.298310
+10		2021-04-08 17:34:12	4 min 48.648 sec	50.0	0.521098	0.863513	0.815260	0.839107	1.981774	0.301199
+11		2021-04-08 17:34:14	4 min 50.656 sec	55.0	0.520741	0.862269	0.815700	0.839448	1.981545	0.299180
+12		2021-04-08 17:34:15	4 min 51.297 sec	60.0	0.520741	0.862269	0.815700	0.839448	1.981545	0.299180
+13		2021-04-08 17:34:16	4 min 51.956 sec	65.0	0.520741	0.862269	0.815700	0.839448	1.981545	0.299180
+14		2021-04-08 17:34:16	4 min 52.613 sec	70.0	0.520741	0.862269	0.815700	0.839448	1.981545	0.299180
+
+Variable Importances: 
+variable	relative_importance	scaled_importance	percentage
+0	coupon_type_department	1.227769e+06	1.000000	0.534741
+1	coupon_mean_prod_price	2.428440e+05	0.197793	0.105768
+2	coupon_how_many	1.721953e+05	0.140251	0.074998
+3	coupon_type_just_discount	1.354846e+05	0.110350	0.059009
+4	coupon_type_buy_all	1.111026e+05	0.090492	0.048390
+5	coupon_discount	1.098288e+05	0.089454	0.047835
+6	coupon_type_buy_more	8.621582e+04	0.070222	0.037550
+7	coupon_days_valid	7.937698e+04	0.064651	0.034572
+8	cust_unique_coupons_used	2.874513e+04	0.023412	0.012520
+9	cust_mean_discount	2.412111e+04	0.019646	0.010506
+10	cust_total_products_bougth	2.372456e+04	0.019323	0.010333
+11	cust_mean_product_price	1.860420e+04	0.015153	0.008103
+12	cust_unique_products_bought	1.549414e+04	0.012620	0.006748
+13	cust_credit	4.810139e+03	0.003918	0.002095
+14	cust_age_mid	4.308483e+03	0.003509	0.001877
+15	cust_age_old	3.579277e+03	0.002915	0.001559
+16	cust_gender_F	3.050989e+03	0.002485	0.001329
+17	cust_age_young	2.436828e+03	0.001985	0.001061
+18	cust_gender_M	2.314412e+03	0.001885	0.001008
\ No newline at end of file
diff --git a/training-with-artificial-data/data_diagram.png b/training-with-artificial-data/data_diagram.png
new file mode 100644
index 0000000..994a23d
Binary files /dev/null and b/training-with-artificial-data/data_diagram.png differ
diff --git a/training-with-artificial-data/requirements.txt b/training-with-artificial-data/requirements.txt
new file mode 100644
index 0000000..c67a43a
--- /dev/null
+++ b/training-with-artificial-data/requirements.txt
@@ -0,0 +1,6 @@
+h2o
+imblearn
+notebook
+matplotlib
+pandas
+scikit-learn