Fireside

• 

  Landing page

• 

  Organization view

• 

  User view

Inspiration

Wildfires are becoming increasingly frequent and severe, putting lives at risk. While existing systems track fires, there’s a lack of community-driven evacuation support. Our goal was to create a platform that helps users find the safest escape routes to nearby safe zones (“Firesides”) during an emergency.

What it does

Fireside is a web application that allows users to:

✅ Select a nearby Fireside (for normal users) or create a Fireside (for organizations) - a designated safe zone created by organizations such as food banks and hospitals

✅ View available resources at a Fireside

✅ View a route to each Fireside that avoids the ongoing fire

How we built it

Frontend:

• T3 web stack to set up server interactions with the frontend

• Leaflet.js for interactive map rendering and routing between points

Backend:

• Supabase for storing data

• Prisma object relational manager to query data

Challenges we ran into

• Obtaining real, current fire data - our initial approach was using an ML model to generate a fire prediction heatmap, but the scale of doing so required time and compute power beyond the scope of a hackathon (i.e., large environmental dataset that gets fed into a deep learning model). Then, we tried using the Fire Weather Index (FWI) as a layer in our Leaflet map, through OpenWeatherAPI, but this requires a paid subscription (or it just wasn't working for some reason).

Accomplishments that we're proud of

• Clean UI

• Database set up with authentication

• Distinguishing between roles of user and organization

• Implementing a routing system that avoids (or at least tries to avoid) fires

What we learned

• How to use the T3 web stack and set up OAuth with various providers

• How to use Leaflet, display custom icons on the map, and route between points

What's next for Fireside

• Training a deep learning model, such as a ConvLSTM, to predict fire probability across an area of land and deploying it via a Flask server that our website can interact with - display heat map over land and determine which areas are unsafe to navigate through

• Implement real-time updates to Firesides on the map and Alerts, through Supabase realtime functionality

• Implement Alert functionality on the backend

Built With

• next.js
• prisma
• react-leaflet
• supabase
• tailwindcss
• trpc
• typescript