FireWatch

Fire stations often use old technology. With Firewatch, they would be instantly notified of new fires, shown where they should allocate their resources, and predict where the next fires might occur.

Comment

Inspiration

We were inspired by the ongoing LA fires and the fires that happen so close to home here in Quebec every summer. We wanted to find a solution that would be extremely efficient in saving lives, the environment, and resources.

What it does

Our all-inclusive web-app monitors forest fires on a dashboard, cost-efficiently sends out units to combat them, and predicts where there could be future forest fires. On top of all of this, we have many interesting analytics and metrics to show.

How we built it

We have a Python backend that uses Redis for queue management and communicates to our React.js frontend with webhooks on FastAPI. We used Random Forest Classifier for predicting future fires, Streamlit for our embedded analytics, and MongoDB for some data storage.

Challenges we ran into

The main challenges were designing an efficient optimization algorithm and keeping course on the right direction, as we had so many ideas to implement, that the scope of our project admittedly went a bit overboard.

Accomplishments that we're proud of

We are proud of the resource allocation algorithm that uses " ". We are also proud of our FastAPI and Webhooks and our nice use of React for a Real-time UI. For once we kind of kept a good code structure at a hackathon.

What we learned

We learned about webhooks and FastAPI. We learned about Redis as well and Streamlit.

What's next for FireWatch

Polishing our product and then launching drones to survey the forests with computer vision and instantly report the occurrence of new fires to our servers to have an even quicker response.

Built With

Share this project:

Updates