juleskkayima/algerian-forest-fire-fwi-predictor

Predict the Fire Weather Index (FWI) for Algerian forests using meteorological data and a Ridge Regression ML model, deployed with Flask and ready for AWS Elastic Beanstalk. # 🔥 Algerian Forest Fire FWI Prediction (End-to-End ML Deployment) An end-to-end Machine Learning project that predicts the **Fire Weather Index (FWI)** using meteorological data from the Algerian Forest Fires dataset. This project covers: * Data preprocessing * Feature scaling * Model training (Ridge Regression) * Model serialization * Flask web application * Production deployment (AWS Elastic Beanstalk ready) ## 📊 Dataset **Source:** Algerian Forest Fires Dataset (Updated CSV) kaggle.com ### Dataset Description The dataset contains meteorological data from two regions of Algeria: * Bejaia Region * Sidi-Bel Abbes Region ### Features Used * Temperature * RH (Relative Humidity) * WS (Wind Speed) * Rain * FFMC * DMC * ISI * Classes * Region Target Variable: * FWI (Fire Weather Index) ## 🧠 Project Workflow ### 1️⃣ Data Preprocessing * Cleaning missing values * Encoding categorical features * Feature scaling using StandardScaler ### 2️⃣ Model Training * Algorithm used: **Ridge Regression** * Model trained on scaled data * Performance evaluated using regression metrics ### 3️⃣ Model Serialization * Model saved using `pickle` * Scaler saved separately * Loaded inside Flask application ### 4️⃣ Web Application * Built using Flask * Accepts user input via HTML form * Scales input * Returns predicted FWI value ### 5️⃣ Deployment Ready * Configured for AWS Elastic Beanstalk * WSGI path configured * Gunicorn included for production server ## 🚀 Project Structure ``` algerian-forest-fire-fwi-predictor/ │ ├── application.py # Flask app entry point ├── requirements.txt # All Python dependencies ├── models/ # Saved ML models and scaler │ ├── ridge.pkl │ └── scaler.pkl ├── templates/ # HTML templates for Flask │ ├── index.html │ └── home.html ├── notebooks/ # Jupyter notebooks for EDA & training │ └── fwi_model_tr …