Rossmann Retail Sales Prediction

Machine Learning project for forecasting future sales using Rossmann Store Sales data, feature engineering, Random Forest, XGBoost, and business insights.

End-to-end sales analytics on 844K+ Rossmann retail records across 1,115 German stores, using PostgreSQL, Python, and Power BI.

Rossmann product price extractor

End-to-end ML pipeline forecasting daily retail sales on the Rossmann Store Sales dataset (1M records, 1,115 stores). Compares Linear Regression, Random Forest, and XGBoost with time-series cross-validation. Tuned XGBoost achieves R² = 0.89 on a 6-week-ahead test window.

Time series sales forecasting — Prophet vs LightGBM on Rossmann dataset. MAPE 9.47%. Deployed on Streamlit Cloud.

Time series sales forecast for Rossmann Pharmaceuticals

🏪 The project's idea is to use a machine learning model to predict the sales quantity that each store will have in the next six weeks, assisting managers in their future decision-making.

DEPI team project predicting Rossmann store sales — EDA, feature engineering, and an ML pipeline tuned with GridSearchCV (final XGBoost model: 97.9% R2). Deployed via Streamlit, FastAPI, and Power BI.

Rossmann store sales forecasting: Prophet vs SARIMA vs LightGBM. MAPE 8.20% (LightGBM winner, 2x better than classical models). Walk-forward CV + anomaly detection on 1M rows.