Inspiration
- With increasing wildfires, a proposal was created to have commercial aircraft be equipped with imaging equipment and to fly over wildfires to provide enhanced imagery for a better emergency response. Our project integrates the concepts of this proposal with pre-existing wildfire detection tools to more efficiently address and monitor ongoing wildfires and find those in remote areas whilst they are still of a smaller, more manageable size.
What it does
- Our project takes in data from 2 sources: NASA’s FIRMS data for possible wildfires, and FlightAware’s AeroAPI to locate commercial aircraft near points of interest. The wildfire data is first sorted by priority(by descending confidence) and data entries with a lower confidence (and thus lower likelihood of being wildfires) are removed. After finding points of interest, it analyzes flight data to find the flight that will pass closest to the fire within a reasonable amount of time. This data could be used to ping the flight and request that it fly over the fire to provide enhanced imagery of the fire to better plan an emergency response- in exchange for the delay/detour, airlines could be provided tax credits.
How we built it
- We researched the APIs provided by each vendor and how to parse them, and proceeded to build code in python to parse and analyze the data. We used React, Leaflet, and Bulma to provide a real time map of detected fires and commercial planes in the air, updated each minute, along with forms to allow users to adjust the desired confidence interval for detected fires.
Challenges we ran into
- We are either first time hackers or have not attended a hackathon in a long time. We were learning and relearning technologies as we built this project, which of course slowed us down. It was a great opportunity to brush up on some pretty essential concepts in SWE, from frontend UI to simple data analysis and sorting.
Accomplishments that we're proud of
- We are proud to have demonstrated a proof of concept with additional features with our limited time, even with a later start time than initially planned.
- Have implemented and used a variety of different technologies we were unfamiliar with.
- Constructing multiple possible solutions and choosing/optimizing for better time complexity and memory efficiency.
What we learned
- Front-end concepts and syntax.
- Back-end integration and API querying.
What's next for FireFlys
- In the future, possibly implement a grid or geoindexing in the map to narrow down the amount of flights being checked per fire, and implement multi-threading to solve multiple fires in parallel, along with better optimizing the code to be more efficient, possibly change flight data APIs so we can query more often and produce more “real time” results.
- Alternatively build a more plane-centric model where each plane goes to a wildfire if a waypoint is within a certain distance of said waypoint.
Log in or sign up for Devpost to join the conversation.