Demand Forecasting using ML
Objective: To optimize workforce allocation for 1.1 million associates to tackle high attrition rates.
Project: Developed a predictive model across 100+ stores that achieved a 6.3% MAPE and a near-zero positive bias, providing stable 5-week-ahead inbound case forecasts to drive more efficient staffing allocation reducing overtimes.
Result: This data-driven approach directly addresses employee attrition, where even a 1% improvement in retention translates into double-digit million-dollar savings in annual operational costs.
- XGBoost
- CatBoost
- LSTM
- Prophet
- Python