The objective of this project is to forecast weekly retail store sales based on historical data using XGBoost Sagemaker.
This dataset used in this project contains weekly sales from 99 departments belonging to 45 different stores. The aim is to forecast weekly sales from a particular department. The data contains holidays and promotional markdowns offered by various stores and several departments throughout the year. Markdowns are crucial to promote sales especially before key events such as Super Bowl, Christmas and Thanksgiving. Developing accurate model will enable make informed decisions and make recommendations to improve business processes in the future.
Xgboost algorithm is used for model building, training and deploying on AWS Sagemaker for the predictive task.
Dataset source: https://www.kaggle.com/manjeetsingh/retaildataset