Flask ML Deployment - Algerian Forest Fires Prediction

A Flask web application to deploy a machine learning model for predicting Algerian forest fire weather indices.

Project Structure

Flask_App/
├── app.py                  # Main Flask application
├── config.py              # Configuration settings
├── export_model.py        # Script to export trained models
├── requirements.txt       # Python dependencies
├── Procfile              # Heroku deployment configuration
├── .gitignore            # Git ignore file
├── templates/
│   └── index.html        # Main web interface
├── static/
│   └── css/
│       └── style.css     # Styling
└── models/               # Directory for trained models (*.pkl files)
    ├── scaler.pkl        # Fitted StandardScaler
    └── lasso_cv_model.pkl # Trained LassoCV model

Features

• Web Interface: User-friendly form to input forest fire features
• Real-time Predictions: Get FWI (Fire Weather Index) predictions instantly
• API Endpoints: RESTful API for programmatic predictions
• Model Deployment: Ready for Heroku, Docker, or other cloud platforms

Installation

1. Create Virtual Environment

# Windows
python -m venv venv
venv\Scripts\activate

# macOS/Linux
python3 -m venv venv
source venv/bin/activate

2. Install Dependencies

pip install -r requirements.txt

3. Prepare Models

First, train your model in the Jupyter notebook and export it:

# In your notebook after training
from export_model import export_models
export_models(lasso_cv, scaler)

This will create:

• models/scaler.pkl - Fitted StandardScaler
• models/lasso_cv_model.pkl - Trained model

Running the Application

Local Development

python app.py

Visit: http://localhost:5000

Production (Heroku)

# Install Heroku CLI
# Login to Heroku
heroku login

# Create app
heroku create your-app-name

# Deploy
git push heroku main

# View logs
heroku logs --tail

Docker Deployment

Create a Dockerfile in the Flask_App directory:

FROM python:3.9-slim

WORKDIR /app

COPY requirements.txt .
RUN pip install -r requirements.txt

COPY . .

CMD ["gunicorn", "--bind", "0.0.0.0:5000", "app:app"]

Build and run:

docker build -t forest-fire-predictor .
docker run -p 5000:5000 forest-fire-predictor

API Usage

Prediction Endpoint

POST /predict

Request body (JSON):

{
  "Region": 0,
  "FFMC": 50,
  "DMC": 100,
  "DC": 200,
  "ISI": 10,
  "BUI": 150,
  "FWI": 50,
  "Rain": 0,
  "Temperature": 25,
  "RH": 60,
  "Ws": 15
}

Response:

{
  "success": true,
  "prediction": 45.32,
  "message": "Predicted Fire Weather Index (FWI): 45.32"
}

Features Endpoint

GET /api/features

Returns information about all input features and their valid ranges.

Feature Descriptions

• Region: Region code (0 or 1)
• FFMC: Fine Fuel Moisture Code (0-100)
• DMC: Duff Moisture Code (0-300)
• DC: Drought Code (0-1000)
• ISI: Initial Spread Index (0-60)
• BUI: Buildup Index (0-300)
• FWI: Fire Weather Index (0-100)
• Rain: Rainfall in mm (0-100)
• Temperature: Temperature in °C (2-40)
• RH: Relative Humidity in % (15-100)
• Ws: Wind Speed in km/h (0-40)

Model Information

• Algorithm: LassoCV (Cross-validated Lasso Regression)
• Dataset: Algerian Forest Fires
• Target Variable: Fire Weather Index (FWI)
• Features Standardized: Yes (StandardScaler)

Environment Variables

Create a .env file:

FLASK_APP=app.py
FLASK_ENV=development
SECRET_KEY=your-secret-key

Troubleshooting

Models not loading

• Ensure model files exist in the models/ directory
• Check file permissions
• Verify pickle files are valid

Port already in use

# Change port in app.py
app.run(port=5001)

# Or kill process on port 5000
lsof -ti:5000 | xargs kill -9  # macOS/Linux
netstat -ano | findstr :5000   # Windows

Import errors

pip install --upgrade -r requirements.txt

Contributing

1. Create a feature branch
2. Make your changes
3. Test locally
4. Submit a pull request

License

This project is for educational purposes.

Support

For issues and questions, please open an issue on the repository.