Book: Demand Prediction in Retail - A Practical Guide to Leverage Data and Predictive Analytics

Dataset

The concepts and demand prediction methods covered in the book "Demand Prediction in Retail - A Practical Guide to Leverage Data and Predictive Analytics" are illustrated using a retail dataset. More specifically, we provide two versions of the datatset:

This dataset reports the weekly sales of 44 stock-keeping units (SKUs) from a tech-gadget e-commerce retailer over a period of 100 weeks, from October 2016 to September 2018. It thus has a total of 4,400 rows.

The raw dataset includes the eight following columns (from left to right): week, SKU, weekly sales, featured on the main page (i.e., whether the SKU was featured on the main homepage), color, price, vendor, and functionality.

The processed dataset includes several pre-processing steps such as dealing with missing data, adding lag-prices, and transforming some of the features to dummy variables.

More details about this dataset and the pre-processing steps can be found in the book.

This dataset can be used to estimate and test demand prediction algorithms. If you use this dataset, please reference the book as follows:

Cohen, M. C., Gras, P.E., Pentecoste, A., & Zhang, R. (2022). Demand Prediction in Retail - A Practical Guide to Leverage Data and Predictive Analytics. Springer Series in Supply Chain Management 14, 1-155.