Nyeanchi/cameroon-food-forecasting

This project demonstrates the practical application of machine learning to address real-world food security challenges. By providing reliable price forecasts, we can support better decision-making across the agricultural value chain in Cameroon. # 🌾 Cameroon Food Price Forecasting ## 📌 Project Overview A comprehensive machine learning system for forecasting food prices across different regions in Cameroon using time series analysis and multiple ML models. ### 🎯 Key Features - **Data Processing**: Clean and standardize WFP food price data - **Model Training**: XGBoost, Random Forest, CatBoost, and LSTM models - **Forecasting**: 6-month price predictions with confidence intervals - **Web Interface**: Interactive Streamlit dashboard - **Risk Assessment**: Automated risk level classification ### 📊 Models Used 1. **XGBoost** - Gradient boosting with regularization 2. **Random Forest** - Ensemble decision trees 3. **CatBoost** - Categorical feature handling 4. **LSTM** - Deep learning for sequence prediction ### 🚀 Quick Start #### 1. Clone Repository ```bash git clone github.com cd cameroon-food-forecasting ``` #### 2. Install Dependencies pip install -r scripts/requirements.txt #### 3. Run Analysis Pipeline ### Open Jupyter Lab bash jupyter lab Run notebooks: 1. FootPredictionModel.ipynb 2. Run the codes inorer: 1-4 ### 4. Launch Web App bash streamlit run scripts/PredictionApp.py ### 📁 Project Structure cameroon-food-forecasting/ ├── notebooks/ # Jupyter notebooks for analysis ├── scripts/ # Python scripts and app ├── data/ # Raw and processed data ├── models/ # Trained ML models └── visualizations/ # Generated plots and charts ### 📈 Results Average Forecast Error: 12.5% SMAPE Best Performing Model: XGBoost (for most commodities) Most Predictable: Beans in Extrême-Nord region Most Volatile: Fresh fish in coastal regions ### 🛠️ Technologies Used Python: pandas, numpy, matplotlib, seaborn ML Libraries: scikit-learn, xgboost, catboost, tensorflow Visualization: Plotly, Streamlit Deployment: GitHub, Streamlit Cloud ### 👥 Contributors Nsaichie Berinyuy Nyeanchi - Data Scientist/Analyst ### 📄 License MIT L …