Charts

[Henrietta1989]

Hi

Your python-poloniex is so simple to use that newbie like me can use it!!!

Is it possible to create similar charts that Poloniex website has? If not, do you know any newbie friendly way to create candlestick charts like that?

Thanks

[s4w3d0ff]

The easiest way to plot candlestick charts is to probably use matplotlib and mpl_finance https://github.com/matplotlib/mpl_finance

pandas also works well with this kind of data and has built-in functions that makes plots using matplotlib easier.

Some links to check out:
https://pythonprogramming.net/candlestick-ohlc-graph-matplotlib-tutorial/
https://pythonprogramming.net/data-analysis-python-pandas-tutorial-introduction/
http://pandas.pydata.org/pandas-docs/stable/visualization.html
https://stackoverflow.com/questions/35677173/charting-with-candlestick-ohlc

To get the raw chart data into a pandas dataframe you just need to:

from poloniex import Poloniex
import pandas as pd
from time import time

api = Poloniex(timeout=120, jsonNums=float)

raw = api.returnChartData('BTC_LTC', period=api.MINUTE * 5, start=time() - api.DAY)
df = pd.DataFrame(raw)

# adjust dates format and set dates as index
df['date'] = [pd.to_datetime(c['date'], unit='s') for c in raw]
df.set_index('date', inplace=True)

# show the end of dataframe
print(df.tail())

                        close      high       low      open  quoteVolume  \
date                                                                       
2017-06-01 18:10:00  0.011370  0.011400  0.011340  0.011400  2121.235002   
2017-06-01 18:15:00  0.011330  0.011370  0.011330  0.011350  1305.376729   
2017-06-01 18:20:00  0.011375  0.011376  0.011331  0.011340   522.556688   
2017-06-01 18:25:00  0.011400  0.011431  0.011344  0.011375  2424.692791   
2017-06-01 18:30:00  0.011414  0.011414  0.011385  0.011385    24.396853   

                        volume  weightedAverage  
date                                             
2017-06-01 18:10:00  24.129772         0.011375  
2017-06-01 18:15:00  14.803764         0.011341  
2017-06-01 18:20:00   5.931372         0.011351  
2017-06-01 18:25:00  27.650402         0.011404  
2017-06-01 18:30:00   0.278263         0.011406  

I'm working on a module that saves chart data in mongodb and returns data in a dataframe with indicators. But I haven't worked in the plotting bit yet.

[s4w3d0ff]

This seems like a good start: pandas-dev/pandas#783 (comment)

[s4w3d0ff]

Not pretty but it it works:

from poloniex import Poloniex
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from mpl_finance import _candlestick
from time import time

api = Poloniex(timeout=120, jsonNums=float)

raw = api.returnChartData('BTC_LTC', period=api.MINUTE * 5, start=time() - api.DAY)
df = pd.DataFrame(raw)

# adjust dates format and set dates as index
df['date'] = [pd.to_datetime(c['date'], unit='s') for c in raw]
df.set_index('date', inplace=True)

# show the end of dataframe
print(df.tail())
df.dropna(inplace=True)
fig, ax = plt.subplots()
idx_name = df.index.name
dat = df.reset_index()[[idx_name, "open", "high",
                        "low", "close", "volume"]]
print(dat)
dat[df.index.name] = dat[df.index.name].map(mdates.date2num)
ax.xaxis_date()
ax.xaxis.set_major_formatter(mdates.DateFormatter("%Y-%m-%d %H:%M:%S"))
plt.xticks(rotation=45)
_candlestick(ax, dat.values,
             width=0.002, colorup='g', colordown='r',
             alpha=1.0, ochl=False)

ax.grid('on')
plt.subplots_adjust(left=.09, bottom=.14, right=.94,
                    top=.95, wspace=.20, hspace=0)
plt.xlabel('Date')
plt.ylabel('Price')

# plt.savefig("candle.png")
plt.show()

[s4w3d0ff]

Here is a nicer looking chart using bokeh

from math import pi
from time import time
from poloniex import Poloniex
import pandas as pd
from bokeh.plotting import figure, output_file, show

api = Poloniex(timeout=None, jsonNums=float)

period = api.HOUR * 4
raw = api.returnChartData('BTC_LTC', period=period, start=time() - api.DAY*7)

df = pd.DataFrame(raw)

df['date'] = pd.to_datetime(df["date"], unit='s')
df.dropna(inplace=True)
print(df.tail())

w = (period * 1000) - 5000
tools = "pan,wheel_zoom,box_zoom,reset,save"

output_file("BTC_LTC.html", title="BTC_LTC-Poloniex")

p = figure(x_axis_type="datetime", tools=tools, plot_width=1500, title="BTC_LTC")
p.xaxis.major_label_orientation = pi / 4
p.grid.grid_line_alpha = 0.7

inc = df.close > df.open
dec = df.open > df.close

p.segment(df.date, df.high, df.date, df.low, color="black")

p.vbar(df.date[inc], w, df.open[inc], df.close[inc], fill_color="green", line_color="black")
p.vbar(df.date[dec], w, df.open[dec], df.close[dec], fill_color="red", line_color="black")

show(p)

Dataframe:

       close                date      high       low      open    quoteVolume  \
37  0.011591 2017-06-15 04:00:00  0.011929  0.011400  0.011520  144043.619743   
38  0.011591 2017-06-15 08:00:00  0.011806  0.011510  0.011597   93924.507037   
39  0.011878 2017-06-15 12:00:00  0.011981  0.011558  0.011591  208059.835072   
40  0.012256 2017-06-15 16:00:00  0.012533  0.011740  0.011894  333514.493589   
41  0.012166 2017-06-15 20:00:00  0.012329  0.012000  0.012256   73753.454514   

         volume  weightedAverage  
37  1676.411564         0.011638  
38  1093.897811         0.011647  
39  2453.011715         0.011790  
40  4071.072424         0.012207  
41   896.245362         0.012152

Chart:

[techietrader]

raw = api.returnChartData('BTC_LTC', period=api.MINUTE * 5, start=time() - api.DAY)
df = pd.DataFrame(raw)

In reply to this message of yours, i have a small query.
You have saved data of a single currency-pair into the variable raw i.e. 'BTC_LTC'
Is there a way you can suggest that I can iterate through all the currency pair in one shot and save data into its own name variable.
For eg
BTC_LTC data goes to var LTC
BTC_ETH goes to var ETH
etc
Something like-

 coinlist=['BTC_ETH','BTC_ZEC','BTC_XMR','BTC_LTC','BTC_ETC','BTC_BTS','BTC_GNT','BTC_XRP','BTC_FCT','BTC_SC','BTC_DCR','BTC_DASH']
 for i in coinlist:
      altcoins=public.returnChartData(i,period=14400)

But that var altcoins only provides me data of BTC_DASH which is last in the list.

Can you think of any solution where data of a particular currencypair goes into its own variable which is created in advance.
Also than we would be required a loop to create 80-100 dataframes

Do you have easy way out ?

[s4w3d0ff]

Maybe store them in a dict? You were just overwriting the the altcoins var each iteration:

coinlist = ['BTC_ETH',
            'BTC_ZEC',
            'BTC_XMR',
            'BTC_LTC',
            'BTC_ETC',
            'BTC_BTS',
            'BTC_GNT',
            'BTC_XRP',
            'BTC_FCT',
            'BTC_SC',
            'BTC_DCR',
            'BTC_DASH']

altcoins = {}

for coin in coinlist:
    altcoins[coin] = public.returnChartData(coin, period=14400)

for coin in altcoins:
    print(altcoins[coin])

# or rather
print(altcoins)

[techietrader]

Thanks a ton.
Well I took it a little forward by making a new Dict of Dataframe of every element in altcoins dict
Here is the code

data={}
for i in altcoins:
    data[i]=pd.DataFrame(altcoins[i])

for i in data:
    data[i]=data[i].set_index("date", inplace=True)
    data[i]["date"] = pd.to_datetime(data[i]["date"],unit='s')
    data[i]=data[i][['open','high','low','close']]    # These are simply to remove unwanted column

While running this code I keep getting some 'date' error every now and than and I dont know why.
Can you see some mistake in the code?

[s4w3d0ff]

It's because you set date as the index, remove that line or reference the index values.

[s4w3d0ff]

or rather:

for i in data:
    data[i]["date"] = pd.to_datetime(data[i]["date"],unit='s')
    data[i]=data[i].set_index("date", inplace=True)
    data[i]=data[i][['open','high','low','close']]    # These are simply to remove unwanted column

[techietrader]

Right!.
Thanks
This worked well

for i in data:
    data[i]["date"] = pd.to_datetime(data[i]["date"],unit='s')
    data[i]=data[i].set_index("date")
    data[i]=data[i][['open','high','low','close']]     

[s4w3d0ff]

Added a bokeh plotting example with basic indicators, it also can merge candles into whatever candle period you want using the method discussed at #129.

7a6ff02

[slidenerd]

Doesnt look live, how do you build candles from websocket

[Villiking]

good project