To build a model to predict when supermarket shoppers will next visit the store and how much will they spend.

Abstract

Supermarkets are one of the most important channels of retail. While they have been popular in Western countries from a long time, they also have shown considerable growth in India in the past decade. However, the problems of inventory management has been of serious concern for Supermarket owners. It is very difficult to predict how many customers are going to visit on a particular day. Also, to know how much and what they are going to buy is difficult. We have tried to predict when consumers are next going to visit the store and how much they will spend. The model tries to predict the purchases for frequently purchased items. OBJECTIVE: 1) Identifying patterns for customer visit and the expenditure from the given data. 2) Identifying the variables that may affect customer’s visit and expenditure. 3) To come up with a model that will predict when a particular customer will next visit the store. Motivation: The motivation behind the project was the importance of inventory management for supermarkets. We will try to predict inventory levels for few important items which will help reducing stockouts and excess inventory costs. Knowing the customer’s expected visit date can help in designing targeted promotions and deals. It may also help in identifying which customers have left visiting the store. Also being able to predict how much a customers is going to spend in his next visit will help the store in designing relevant promotional campaign. Also it may reduce the expenditures on one-to-one marketing like phone calls, sending grocery list etc. The academic motivation behind the project was to get hands on experience of dealing with large datasets. The possibility of dealing with various variables and identifying consumptions pattern would provide immense learning.