BACK

Best Buy Sales Forecasting

Project description: Time series sales forecasting project incorporating seasonality across multiple levels of hierarchy.

Details: During the Best Buy MSA Project Week, I was part of a team that worked on forecasting slow-selling SKUs using time series analysis and machine learning techniques. Our main goal was to create a final dataset for model training by researching and sourcing external data for reuse and inclusion in the model. We also identified trends and seasonality patterns in the data to improve the model accuracy.

To achieve our objective, we tested multiple models and ultimately selected XGBoost due to its superior performance in both accuracy and computational speed. The project required a combination of technical and analytical skills, including data cleaning, exploratory data analysis, feature engineering, and model selection. By using XGBoost, we were able to generate accurate forecasts that helped Best Buy optimize their inventory and improve their bottom line.

BACK