FORECASTING DEMAND AT SUPERMARKETS USING QUANTILE REGRESSION

Managing the inventory at supermarket is of utmost importance as it is the majorly responsible for maximizing revenue and maintaining high customer service levels. There are numerous reasons why keeping a stock is fundamental. With the cuttingedge retail industry blasting and ecommerce business infiltrating into the market, it is very important for a supermarket to maintain its product availability. The aim of this project was to forecast daily demand for selected supermarkets in Bengaluru using quantile regression analysis. The basic objective of the study was to develop a quantile regression model to identify the relationship between daily average sales and average number of customers per day. Secondary data was collected for a time period of 3 months from 4 selected supermarkets in Bengaluru. Descriptive Statistics Analysis was performed using Microsoft excel to analyse the collected secondary data. Least Square Analysis, Quantile Regression Analysis for quantiles 0.25, 0.50, 0.80 and 0.90, Residual Analysis, and Demand Forecasting was carried out using E-views. Based on the results of the data analysis, it is seen that the best fit for the quantile regression is at quantile 0.80, making it a tight fit for the regression line.