Predicts weekly Walmart store sales using historical, economic, and weather data. Incorporates feature engineering, modeling, and scenario simulations to uncover trends, optimize promotions, and improve staffing for higher profitability.
Key features engineered for the model include:
- Holiday_Flag: Marks weeks with major promotions or holidays.
- Temperature: Estimates weather impact on store traffic.
- Fuel_Price, CPI, Unemployment: Represent economic conditions.
- Lag Features (1–5 weeks): Capture recent sales momentum.
- Store_Cluster: Groups stores with similar sales behavior.
- Week_Number & Month: Identify seasonal patterns.
- Data Cleaning & Exploration – Analyzed sales patterns and correlations.
- Feature Engineering – Created lag variables, store clusters, and seasonality indicators.
- Model Selection – Tested multiple algorithms for accuracy and interpretability.
- Scenario Simulations – Assessed the impact of shifting promotions or adjusting staffing.
- Store clusters revealed distinct demand profiles (urban vs. rural, high vs. low traffic).
- Lag features improved short-term forecast accuracy.
- Seasonal and holiday effects were strong drivers in certain clusters.
- "What-if" scenarios quantified the revenue impact of business decisions.
├── data/ # Raw and processed datasets ├── notebooks/ # Jupyter notebooks for EDA, modeling, and visualization ├── scripts/ # Python scripts for preprocessing and model training ├── results/ # Model outputs and visualizations └── README.md # Project documentation
bash Copy Edit
- Clone the repository:
git clone https://github.com/bs6396/walmart-sales-forecasting.git cd walmart-sales-forecasting
Install dependencies:
bash Copy Edit pip install -r requirements.txt Run the Jupyter notebooks in /notebooks.
📜 License This project is licensed under the MIT License.
🤝 Contributions Pull requests and suggestions are welcome. Please open an issue for major changes.
yaml Copy Edit
If you want, I can also make the requirements.txt so that anyone cloning your repo can run it without dependency issues.