This repository consists of all files related to internship at HDFC Bank. The bank statistics data were taken from official Reserve Bank of India website (https://rbi.org.in/Scripts/Statistics.aspx). The data consists of the Point of Sales Machine transaction and ATM transactions for the years 2018 - 2019 (month of May). The objective of this project work was to analyze the performance of the Indian banking system.

Data was segregated into government and private bank datasets Feature Engineering was implemented and certain statistics were extracted from the dataset: like per transaction value of credit cards for a particular bank

Version: Jupyter Notebook Packages: pandas, numpy, sklearn, matplotlib, seaborn, json

After reading the data, I needed to clean it up so that it was usable for further analysis. I made the following changes and created the following variables:

• Segregated the bank dataset into Private and Government Banks
• Parsed numeric data out of ATM and POS transactions
• Made columns for transaction value for CC and DC transactions
• Removed rows without any transaction values
• Parsed rating out of company text
• Used outstanding CC and DC to aggregate transaction values
• Made columns if transaction levels were higher than the average:
  • Credit Cards: Private and Government
  • Debit Cards: Private and Government

Distributions of the data and the value counts for the various categorical variables.Also, data was sorted according to various peaks. Binning Data: Data was binned according to various categories and then the top performing banks were visualized.

In this step, I created a different jupyter notebook file wherein different visuals of the data are taken into cognizance. It gives an in-depth research of other factors as well which are not present in the primary file such as violin plots, join plots of various categorical features.