Wildfire Risk Assessment Project

Project Overview

This project leverages Canada's open-source forest wildfire data with two primary objectives:

Objectives

1. Predict the Cause of Forest Fires:

   • Many forest fires have unknown causes. We've developed a machine learning model that predicts the cause of a fire (Human or Lightning) based on historical fire data and temperature data.

2. Predict the Probability of Fire Occurrence:

   • Our model can also predict the likelihood of a fire occurring at a specific latitude, longitude, and day of the year. This prediction is based on a voxel approach.

Code Structure

Data Files

• forest_fire.txt: Contains raw data about forest fire incidents.

Jupyter Notebooks

• data_cleaning.ipynb: Notebook detailing data cleaning and pre-processing steps.

Directories and Models

• Prong1_Predicting_Unknown_Fires:

  • Contains models and scripts for predicting the causes of fires with unknown origins.

• Prong2_Predicting_Fire_Probability:

  • Houses models and scripts for calculating the probability of fire occurrences at specific locations and times.