April 9, 2020 Python Q&A

Topics covered in this session include:

  • Continued work on Risk on / Risk off tool
  • Preparing and merging two dataframes
  • Rolling Correlation between 2 assets
  • Histogram of rolling correlation

RiskOnOff.py

# -*- coding: utf-8 -*-
"""
Created on Sun Jan  5 11:02:36 2020

@author: Bruce1
"""

# =======================================================================
# Import Libraries
# =======================================================================
import pandas as pd 
import pandas_datareader.data as web
from   pathlib import Path
import matplotlib.pyplot as plt
#import numpy as np

# =======================================================================
# Setup Porgram Variables
# =======================================================================
symList = ['SPY', 'TLT']

startDate = '01/01/2000'

refreshData = False

lookback = 21

# =======================================================================
# Gather data function
# =======================================================================
def gatherData(sym, startDate):
    #import pdb; pdb.set_trace()

    savedFile = Path('./{}.xlsx'.format(sym))
    
    if savedFile.exists() == False or refreshData == True:
        print("")
        print("-> Fetching data from the web")
        df = web.DataReader(sym, data_source='yahoo', start=startDate)
    
        print("")
        print("-> Save data to file")  
        df.to_excel("{}.xlsx".format(sym))

    else:
        print("")
        print("-> Fetching data from file") 
        df = pd.read_excel(savedFile, index_col='Date', parse_dates=True)

    # =======================================================================
    # Inspect/Report on data
    # =======================================================================
    firstIndex = df.index.min()
    lastIndex  = df.index.max()
    records = len(df)
    print("")
    print("-> Importing ", sym)
    print("First Date = ", firstIndex)
    print("Last Date  = ", lastIndex)
    print("Total Days = ", records)

    if df.isnull().values.any() == True:
        print("WARNING: there are {} NaN in the data".format(df.isnull().values.sum()))
        print(df.isnull().values)
        
    return df

dfDict = {}

for sym in symList:
    
    dfDict[sym] = gatherData(sym, startDate)

import pdb; pdb.set_trace()

# =======================================================================
# Format/Normalize Data
# =======================================================================
cdf = pd.DataFrame() 
first = True

for sym in symList:
    
    tdf = dfDict[sym].copy()
    tdf.rename(columns={'Close':sym}, inplace=True)
    
    if first:
        cdf = tdf.loc[:,sym]
        first = False
        
    else:
        cdf = pd.merge(cdf, tdf.loc[:,sym], left_index=True, \
                       right_index=True,  how='inner')
    
import pdb; pdb.set_trace()

# =======================================================================
# Calculate Correlation
# =======================================================================
xdf = cdf.loc[:,symList[0]].rolling(lookback).corr(cdf.loc[:,symList[1]])

# =======================================================================
# Visualize Data
# =======================================================================
plt.figure()
t = 'Correlation of {}'.format(symList)
xdf.plot(kind='hist', bins=20, title = t)
plt.show()


cdf.to_excel('temp.xlsx')

Scroll to Top