Calculated most-potential fire hazard data. Pictured: coordinates and fire hazard levels
With the recent rise of wildfires disrupting wildlife, evacuating thousands of families and impacting many more, our team was inspired to create the next wildfire forecasting tool, FireWatch. Using metrics and predictive analysis, FireWatch displays possible future California wildfires onto a Google Maps API using Machine Learning with our mascot Disel the Dog!
Our process (with 4 team members and 36 hours):
- [Data] Extract relevant data (wildfire history, altitude, vegetation, and precipitation patterns) from the CalFire Database and use Python scripts to parse data for model training.
- [Backend] Host a Flask server with Python, which takes care of Machine Learning calculations and handling SendGrid email requests.
We had some troubles merging data, integrating the .geojson file onto our Google Maps API, and deciding on algorithms.
What's next for FireWatch?
• Market our product for those in California wildfire-prone areas • Finalizing and selling our product to CalFire • Securing a .tech domain