Betwise Smart Market Performance Analysis

Betwise_smart_markets_PandL.py

import os
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
from datetime import datetime as dtdt
pd.set_option('display.width', 160)
pd.set_option('display.max_rows', 50)
pd.set_option('display.max_info_rows', 20)
pd.set_option('display.max_info_columns', 20)


csv_file = 'smart_markets.csv'
# loading from csv is much faster than html, so we use that if we can
if os.path.exists(csv_file):
    print 'loading from csv...'
    df = pd.read_csv(csv_file)
else:
    # smart_markets_download.html is just http://www.betwise.co.uk/smart_markets_pandl saved
    print 'loading from html...'
    df = pd.read_html('smart_markets_download.html', header=0, infer_types=False)[0]
    df.to_csv(csv_file)
df = df.reindex( index=df.index[ ::-1 ] )
df['Date'] = df['Date'].apply(lambda x: dtdt.strptime(x, '%d-%m-%Y'))
df = df.set_index('Date')
df['BSP Win'] = df['BSP Win'].astype(float)
df['loss'] = df.apply(lambda row: (0,1)[row['Result'] != 'NR'], axis=1)
df['loss2'] = df.apply(lambda row: (0,1)[row['Result'] != 'NR' and pd.notnull(row['BSP Win'])], axis=1)
df['profit'] = df.apply(lambda row: (0, row['BSP Win'])[row['Result'] == '1' and pd.notnull(row['BSP Win'])], axis=1)
df['PL'] = df.apply(lambda row: row['profit']-row['loss'], axis=1)
df['PL2'] = df.apply(lambda row: row['profit']-row['loss2'], axis=1)
cpl1 = 'cum P&L'
df[cpl1] = df.PL.cumsum()
cpl2 = 'cum P&L NULL removed'
df[cpl2] = df.PL2.cumsum()
print df
print df.dtypes
df[[cpl1, cpl2]].plot()
plt.savefig('smart_markets_PandL.jpg')
plt.show()