Update README with project overview

README.md

@@ -1,16 +1,120 @@
-# Inventory Analysis
+# Inventory Demand Forecasting
 
-End-to-end demand forecasting pipeline (dataset: retail_store_inventory.csv).
+End-to-end inventory analytics project that predicts daily units sold and provides explainability, evaluation reports, and production-ready serving via API and Streamlit.
 
-## Docker
+## Why This Project
+- Forecasts demand for store-product pairs to reduce stockouts and overstock
+- Tracks experiments and saves the best model for deployment
+- Provides SHAP explanations for transparent predictions
+- Includes automated tests, CI workflow, and Docker support
+
+## Highlights
+- Pipeline: ingestion → transformation → training → evaluation
+- Evaluation reports by store, product, category, and region
+- FastAPI prediction and explanation endpoints
+- Streamlit UI for interactive predictions
+- MLflow tracking and artifacts management
+
+## Tech Stack
+- Python, pandas, scikit-learn
+- FastAPI, Streamlit
+- MLflow
+- Pytest, GitHub Actions
+- Docker, Docker Compose
+
+## Project Structure
+```
+.
+├── artifacts/
+├── data/
+├── src/
+│   ├── app.py
+│   ├── data_ingestion.py
+│   ├── data_transformation.py
+│   ├── model_evaluation.py
+│   ├── model_pipeline.py
+│   ├── model_trainer.py
+│   ├── predict_pipeline.py
+│   └── streamlit_app.py
+├── templates/
+├── tests/
+├── docker-compose.yml
+├── Dockerfile
+└── requirements.txt
+```
+
+## Quickstart (Local)
+1) Create and activate a virtual environment
+2) Install dependencies
+3) Train and evaluate
+
+```cmd
+python -m venv .venv
+.\.venv\Scripts\activate
+python -m pip install -r requirements.txt
+python src\model_pipeline.py
+```
+
+## Run FastAPI
+```cmd
+python -m uvicorn app:app --reload --app-dir src
+```
+
+Endpoints:
+- `GET /health`
+- `POST /predict`
+- `POST /explain`
 
-Build and run API + Streamlit:
+Docs: `http://127.0.0.1:8000/docs`
 
-```bash
+## Run Streamlit
+```cmd
+python -m streamlit run src/streamlit_app.py
+```
+
+App: `http://localhost:8501/`
+
+## Docker
+```cmd
 docker compose up --build
 ```
 
 Ensure `artifacts/models/best_model.pkl` exists on the host before running.
 
-API: http://127.0.0.1:8000  
-Streamlit: http://127.0.0.1:8501
+API: `http://127.0.0.1:8000`  
+Streamlit: `http://localhost:8501`
+
+## Tests
+```cmd
+python -m pytest
+```
+
+## MLflow
+Runs are stored in `mlruns/`.  
+Launch UI:
+```cmd
+python -m mlflow ui
+```
+Open: `http://127.0.0.1:5000`
+
+## Evaluation Reports
+Generated in `artifacts/reports/`:
+- `evaluation_summary.json`
+- `metrics_by_store.csv`
+- `metrics_by_product.csv`
+- `metrics_by_store_product.csv`
+- `metrics_by_category.csv`
+- `metrics_by_region.csv`
+
+## Screenshots
+Add screenshots here after running the app:
+- `docs/screenshots/streamlit_home.png`
+- `docs/screenshots/prediction_result.png`
+- `docs/screenshots/mlflow_runs.png`
+
+## Notes
+- `Demand Forecast` is included as a feature and is highly predictive in this dataset.
+- If you want a more realistic model, remove it from features and retrain.
+
+## Author
+BabaKaElijah