ML-powered demand forecasting + Streamlit dashboard + GenAI chatbot
Reduce food waste by 30-40% | 95%+ stock availability | €26K-104K annual ROI
End-to-end solution for 989 grocery SKUs:
-
XGBoost: 97% accurate demand prediction (MAE: 2.6)
-
Optimization: EOQ, safety stock → 213 URGENT | 776 REORDER
-
Dashboard: Live KPIs, search, CSV export
-
GenAI: GPT-4o chatbot ("What expires today?")
├── data_exploration.ipynb # EDA & cleaning ├── feature_engineering.ipynb # 30+ ML features ├── model_training.ipynb # XGBoost models ├── invnetory_optimization.ipynb # EOQ, reorder points ├── genai_inventory.py # Main Streamlit app ├── genai_chatbot.py # GPT-4o assistant └── requirements.txt
| Metric | Impact |
|---|---|
| Waste Reduction | 30-40% ↓ |
| Stock Availability | 95%+ ↑ |
| Weekly Savings | €500-2000 |
| Time Saved/Manager | 30-40 min/day |
pandas numpy scikit-learn xgboost streamlit openai matplotlib seaborn plotly python-dotenv joblib
Mitesh Parab
miteshparab89@gmail.com
LinkedIn