Store sales forecasting

Project type: Hackathon - individual

Hackathon Link: Job-a-thon September 2021

Overview

The hackathon challenge is to forecast the next 2 months’ sales based on 350+ retail stores of WOMart over the past 18 months. This effective forecasting for store sales gives essential insight into upcoming cash flow, inventory management etc. meaning WOMart can more accurately plan the cashflow at the store level.

I have converted the time-series based forecasting into a regression based on my analysis and employed extreme gradient boost regressor (xgboost regressor) to capture 14^th position among 6828 participants in the hackathon. Here is the link to my solution on GitHub.

Leaderboard position: Leaderboard link

Private leaderboard

Public leaderboard