An interactive supply chain forecasting tool designed for FMCG/Retail clients. This project bridges the gap between machine learning metrics and business value by allowing category managers to simulate promotional "What-If" scenarios and visualize their impact on inventory risk and predicted revenue.
- Inventory Optimization: Flags high-risk stock-out periods based on non-linear holiday and promotional trends.
- Revenue Simulation: Translates XGBoost demand predictions into projected financial lift based on discount elasticity.
- Data Ingestion: Automated fetching of historical retail data.
- Feature Engineering: Extracts seasonality, promotional depth, and day-of-week matrices.
- Inference Engine: XGBoost regressor optimized for time-series retail spikes.
- Client Interface: Streamlit dashboard for non-technical stakeholder interaction.
git clone [https://github.com/Atri2-code/retail-demand-optimizer.git](https://github.com/Atri2-code/retail-demand-optimizer.git)
cd retail-demand-optimizer
pip install -r requirements.txt
python model_trainer.py
streamlit run app.py