FireWatch AI utilizes an advanced machine learning model to predict future wildfires based on key features identified by the same AI.
Fire Prediction: Predict the occurrence of wildfires based on historical data and weather conditions.
Real-Time Heat Maps: Visualize predicted fire locations using interactive heat maps.
To run this project, you need to install the following libraries:
- pandas
- numpy
- joblib
- folium
pip install xgboost pandas scikit-learn folium geopandas joblib- Load the Dataset: Update the file_path variable with the path to your dataset.
- Train the Model: The script will preprocess the data, train the XGBoost classifier, and evaluate its performance.
- Generate Predictions: The script will generate predictions for the entire dataset and create a heat map of predicted fire locations.
- Visualize: The heat map will be saved as an HTML file (fire_prediction_heatmap.html).
- Data Preprocessing: Handles missing values and converts feature columns to integers.
- Model Training: Splits the data into training, validation, and test sets. Trains the XGBoost model and evaluates its performance.
- Model Evaluation: Outputs accuracy, classification report, and confusion matrix.
- Model Saving: Saves the trained model using joblib.
- Visualization: Creates a heat map of predicted fire locations using Folium.