A Quick Start Guide to Compute Correlation Matrix in Python using NSEpy & Pandas

Here is a quick tutorial in python to compute Correlation Matrix between multiple stock instruments using python packages like NSEpy & Pandas. Generally Correlation Coefficient is a statistical measure that reflects the correlation between two stocks/financial instruments. Determining the relationship between two securities is useful for analyzing intermarket relationships, sector/stock relationships and sector/market relationships.

Here are some of the essential python libraries required for Correlation Matrix Data Visualization

IPython (Interactive Python)
Pandas (Python Library to handle time-series data )
NSEpy (Fetch Historical data from NSEindia – NSEpy 0.3 ver or higher)
Matplotlib (Python library to handle 2D plotting)

Import the required python modules

i)from nsepy.archives we need to import the get_price_history:-for fetching the stock pricing details
ii) from datetime import we need to import date object :- for giving the date limit for the required stocks
iii) import pandas :- for the creation of dataframe
iv)import matplotlib for plotting the correlation heatmap

Create a list of stocks

https://gist.github.com/bc87b534fb63ec4faae83843fbdc1f39

Fetch Historical Data
Now fetch the price history of each stock based on a specific time limit and appending the last close value into an empty pandas data frame

https://gist.github.com/0256e8d6701a95f9c3d8e21d96ed8b66

Compute and Plot Correlation Matrix
Now compute the percentage change and Pearson correlation using the pandas dataframe functions pct_change() , corr() and plot the correlation matrix using
matplotlib as shown below.

https://gist.github.com/1d2ba65cb5ba1da38cf542947d2885bc

Note : %pylab inline is used as a display variable on ipython notebook.

Sample IPython Notebook to generate Stock Correlation Matrix Map as shown below :