Retail Sales Forecasting & Inventory Optimization System

Demo GIF

Architecture

                +-----------------------+
                | data/generate_data.py |
                +-----------+-----------+
                            |
                            v
+-------------+    +--------------------+    +--------------------------+
| data_loader | -> | preprocessor.py    | -> | feature_engineering.py   |
+-------------+    +--------------------+    +-------------+------------+
                                                          |
                                                          v
                                          +-------------------------------+
                                          | forecasting.py (RF + XGBoost) |
                                          +---------------+---------------+
                                                          |
                                   +----------------------+---------------------+
                                   v                                            v
                     outputs/forecasts.csv                       models/best_model.pkl
                                   |
                                   v
                         +--------------------+
                         | inventory.py        |
                         +---------+----------+
                                   |
                                   v
                  outputs/reorder_recommendations.csv
                                   |
                                   v
                        app/dashboard.py (Streamlit)

Installation

python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt

How To Run

python data/generate_data.py
python main.py
streamlit run app/dashboard.py

Screenshots

• docs/overview.png
• docs/forecasting.png
• docs/inventory.png
• docs/store_analytics.png
• docs/eda.png

Tech Stack

Layer	Tools
Data	Pandas, NumPy
ML	Scikit-learn, XGBoost
Stats	SciPy, Statsmodels
Dashboard	Streamlit, Plotly
Serialization	Joblib

Results (Sample)

• MAE: ~4.5 to 6.5
• RMSE: ~6.0 to 9.0
• MASE: < 1.0 for strong SKUs
• Reorder recommendations for all store-SKU combinations