Disaster Response Pipeline Project

Project motivation

This project analyzes disaster data to build a model for an API that classifies disaster messages.

The project analyzes messages for disaster response. It is intended to know which type of responses organizations should take in case of a disaster. These responses are based on an individual's message in a direct request for help or on social media.

Moreover, slow responses are a matter of concern in a disaster context. Therefore, this type of project is very important because it can help organizations to prioritize resources accurately and quickly.

Files

• data
  • disaster_categories.csv # the raw data containing the labels for each message
  • disaster_messages.csv # the raw data with the messages and its genres
  • process_data.py # the script that performs the ETL operation
• graph
  • utils.py # creates cached objects and builds serialized representations of the visuals
• models
  • train_classifier.py # the script that builds and trains the model
• templates
  • master.html # main page of web app
  • go.html # classification result page of web app
• run.py # Flask file that runs the app
• README.md
• requirements.txt # project dependencies

Instructions to reproduce it locally:

1. First of all, you must assure to install the dependencies:

    pip install -r requirements.txt

2. Run the following commands in the project's root directory to set up the database and model.

   • To run ETL pipeline that cleans data and stores in database:

     python data/process_data.py data/disaster_messages.csv data/disaster_categories.csv data/DisasterResponse.db
     

   • To run ML pipeline that trains classifier and saves it:

     python models/train_classifier.py data/DisasterResponse.db models/classifier.pkl
     

3. Run the web app: python run.py

Other info

The messages data is skewed, hence the accuracy of the predictions will be affected. You can check the bar plot to make sure of it.

Project preview

Check the live demo