Machine learning algorithms for analyzing debit card purchases

Abstract

This paper exposes the presence of certain important trends in consumer spending which the society has been largely unaware of. Firstly the paper talks about the significance of spending volatility of the Indian consumers/households and the importance of risk management by each of these households in order to manage volatilities in spending & consumption. Secondly, the paper tries to cluster the Indian urban consumers in different clusters based on their purchase patterns. These clustering results help in segmenting the consumer base available into broad categories, after which we were able to infer general insights from each cluster as to their unique characteristics such as consumption, internet banking usage, demographics etc. Thirdly, we move on to assess the impact of macroeconomic shocks on the purchasing patterns of the Indian households. This would help us understand the impact of shocks such as RBI rate cuts, crude oil prices etc. on consumer aggregate consumption & the source of majority of that consumption/type of goods spent on. Fourth, we try to assess the internet banking transaction data & extract details regarding the types of transactions done online and the demographics of customers who frequently transact online. These insights can be used by companies for selective & targeted marketing of their online products to customers & also by banks to encourage the use of online banking in lieu of a move towards a cashless Indian society.