Nehmyabiruk/ethiopia-commodity-price-prediction

# Ethiopia Commodity Price Prediction An ML-powered web application for predicting commodity prices in Ethiopian markets using historical market data, time-series features, and XGBoost. ## 🚀 Live Demo **Open Ethiopia Commodity Price Predictor →** ## Features * Ethiopian commodity price prediction * Market and product selection * Time-series feature engineering * Automatic lag feature generation * Rolling price statistics * Price-change features * XGBoost machine learning model * FastAPI prediction API * Single-page frontend interface * Live deployment on Render ## Overview This project predicts commodity prices in Ethiopian markets using historical market data and machine learning. The system uses historical prices to generate time-series features such as: * Lag 1, 3, 6, and 12 months * Rolling 3, 6, and 12 month averages * Price changes over different time periods * Calendar features such as year, month, and quarter * Market and geographic information The trained XGBoost model is exposed through a FastAPI backend and connected to a simple web-based frontend. ## Project Architecture ```text User │ ▼ Frontend │ │ Prediction Request ▼ FastAPI API │ ├── Feature Engineering ├── Historical Market Data └── XGBoost Model │ ▼ Predicted Commodity Price │ ▼ Frontend ``` The application is deployed as a Python web service on Render. ## Repository Structure ```text commodity-predict/ ├── data/ │ ├── raw/ │ └── processed/ ├── models/ ├── notebooks/ ├── src/ │ ├── app.py │ ├── static/ │ │ └── index.html │ └── data/ │ ├── __init__.py │ ├── eda.py │ ├── evaluate.py │ ├── feature_engineering.py │ ├── predict.py │ ├── train.py │ └── utils.py ├── requirements.txt └── README.md ``` ## Setup Clone the repository: ```powershell git clone github.com cd ethiopia-commodity-price-prediction ``` Create a virtual environment: ```powershell python -m venv .venv …