SafeRoute

Inspiration

We noticed how google maps often provides the fastest routes but doesn't say more about it. The goal of this project is to improve on the existing functionality of google maps while providing more information for the user.

What it does

Similar to google maps, the user will enter their starting point and their destination. Afterwards, the website will generate all the routes that fit the criteria (i.e., safety, speed, and lighting).

How we built it

Our team came up with project specifications (i.e., the stack and overall goal) then used AI for the actual coding. The stack we used:

Backend: FastAPI Frontend: Next.js + Javascript Maps +Turf.js APIs: Google geocoding, Google directions

Challenges we ran into

Setting up the api (i.e., figuring out what apis we needed for this project which where geocoding, directions api). We also struggled to get the map to load for some period of time.

Accomplishments that we're proud of

Coming up with the project and designing the overall structure was interesting and enjoyable. For many of us we've never used AI as much as this. We could now focus more on the project specifications and what we wanted/expected. Just let AI handle the implementation and we're good.

What we learned

How to utilize AI to do the actual coding while we worked on project specification, requirements, and debugging. We also improved on our teamworking skills by communicating better with each other, and utilizing the strengths of all out members.

What's next for SafeRoute

Plug in real datasets (lighting, incidents, construction, road quality) and calibrate scores with clear confidence labels. Add personalized safety preferences, turn-by-turn guidance for the chosen route, and production hardening (caching, rate limits, monitoring, safer API key practices). Focus on trust: transparent “what we measure vs infer,” bias review with local context, and pilots (campus/city) before any broad claims.