sohanlv/retail-demand-forecasting

Retail Demand Forecasting

Next-day sales prediction per store–item using historical sales, promotions, and transactions. Model: XGBoost (trained on Favorita dataset sample).

How to use

1. 1.Select a store and item from the sidebar.
2. 2.Adjust the date range.
3. 3.See KPIs (RMSE, SMAPE, wMAPE) and Actual vs Predicted chart.
4. 4.Residuals and feature importance help explain the model.

Why it matters: better inventory planning → fewer stockouts & less overstock.

pinned: False