Realtime Barcode Based Sales Tracking and Forecasting

Abstract

Supply chain management in any domain has been the key to the growth of businesses. Efficient supply chains are known to benefit each level of members of the chain. Yet, there are many untapped inefficiencies that if resolved can tremendously improve the chain performance. Problems like rapid demand surge as in recent times, have led to early stock outs while for some items the unusual decrease in demand has led to increased inventory holding costs and dead stock. At the very core of these problems is the Bullwhip Effect. Our solution and study is divided in two parts. The first one focuses on solving the issues created by the Bullwhip Effect using an efficient manner of sales tracking using an android application, which keeps track of the chain at each level. The second part focuses on enhancing the above solution with provision of intuitive sales forecast generated from machine learning models and user friendly visualizations. This system can be used at every level of supply chain in order to keep track of sales and maintain sufficient inventory based on the demand forecast. The main focus of the study is to discuss in detail the demand forecasting using machine learning and deep learning approaches. We aim to identify a strong baseline model for sales forecasting and also provide intuitive inferences as gathered from exploratory analysis.