Winner of the 2020 Ontario-Wide Software Engineering Capstone Competition 🏆

Forestcasting is a forest fire management tool that provides vital information to aid fire fighting resource allocation. The Forestcasing tool consists of a machine learning model that calculates the probability a forest fire could occur in a given area, and a mathematical damage algorithm that estimates the severity of damages caused by said fire. Forestcasting’s results give forest fire managers the tools and data required to take appropriate precautions.

All design and implementation details including information on data gathering, data pipelines and processing, machine learning tuning and metrics can be found inside the Final Report.

Specifically:

• Page 24-38 Software Design Specification
• Page 41-48 Implementation

• The fire risk model and damage algorithm were built and tuned on Google Colab, the notebooks have been exported as Python scripts into this repository (damage-alg-notebooks and fire-prediction-notebooks).
• The model was deployed as a separate API (ML) while the damage algorithm was directly integrated into the NodeJS backend.
• Lastly, the user interface is coded in ReactJS and can be found in the frontend folder.

See diagram below for architecture and technology stack information:

• Samantha Campbell - Data Science and Backend
• Shima Kananitodashki - Data Science and Frontend
• Tharmiga Loganathan - Data Science and Backend
• Ivan Zvonkov - Data Science and Machine Learning