Sales and demand forecasting is one the most critical tasks of enterprises. It lays the foundation for many other essential business assumptions, such as cash flows, profit margins, turnover, capacity planning, and capital expenditure. This report presents a solution for the case study of forecasting monthly sales of one of the largest retail stores in Europe, Rossman chain stores. There are three steps in which this problem will be handled. First, a complete and comprehensive exploratory data analysis will be done to understand the data and perform feature engineering. Secondly, a time series will be modeled by autoregressive models using machine learning and neural networks. Thirdly, these models will be evaluated with standard time series evaluation metrics. Some of the commonly used approaches to achieving the prediction value or models include the ARIMA and classification-based modeling techniques for forecasting. The literature review indicates that these aspects are hard to choose due to the need for matching the supply and demand of consumers being critical as more consumers prefer more reliable companies and companies that can consistently deliver on what they need and when they need. People have underappreciated machine learning’s ability to make data-driven predictions. Still, given the analysis results, companies are now starting to realize its potential and are investing more in this technology.
Professional Studies (MS)
Department, Program, or Center
Graduate Programs & Research (Dubai)
Al Ali, Mohamed, "Retail Demand Forecasting" (2021). Thesis. Rochester Institute of Technology. Accessed from