Inspiration
With the news on the wildfires in Australia, we wanted to see if we could predict the probability and prevalence of wildfires, and other factors such as spread speed and spread area. Then using these target values, we would be able to simulate future forest fires to create awareness of global warming. We would also be able to spot which fires are more prone to spread at a faster rate.
What it does
Displays a map.
How I built it
.
Challenges I ran into
Unfortunately, data on wildfires is surprisingly difficult to find. The data only had cases with forest fires and contains no data on temperature before nor after. Thus, it we did not have the sufficient time to preprocess it.
Accomplishments that I'm proud of
- Actually having something show
What I learned
- Better familiarity with maps and leaflet
- Learning about cool python libs
- Learning about Plotly
What's next for ML Fire
- Successfully engineering useful features
- Segmentation of forest fires to capture spread
- Modify sizes of dots on the map for different brightness values
- Predicting spread speed for an initial brightness data point
Log in or sign up for Devpost to join the conversation.