SafeRoute

Inspiration

Walking alone at night can often feel unsafe, especially for women and other vulnerable groups. Many navigation apps optimize for speed, not safety - ignoring how factors like lighting, open spaces, or nearby crowds impact how secure someone feels. We wanted to change that. SafeRoute is a web app that was inspired by the idea of empowering people to make informed choices about their routes, not just based on distance, but on real indicators of safety.

What it does

SafeRoute analyzes multiple environmental and social factors to recommend safer walking routes. It calculates safety scores based on three main data sources:

Lighting data from the Overpass API (to detect well-lit streets, especially relevant at night).
Crowd and business activity from Foursquare, representing nearby open establishments and potential foot traffic.
Crime data from public databases to assess the relative safety of each area.

The app suggests two routes and rates each with a clear safety level (High, Medium, or Low). It then uses the Gemini API to generate a user-friendly summary that explains the route’s safety rating and offers practical walking advice depending on conditions such as time of day.

How we built it

We used Mapbox for routing and visualization, Overpass API for lighting data, Foursquare Places API for nearby open venues, and a mock crime dataset for proof of concept. The backend computes safety scores for each route segment and adjusts weights dynamically depending on whether it’s day or night - lighting, for example, is only considered at night. The Gemini API processes the final route data to generate natural-language summaries that make the information accessible and easy to understand.

Challenges we ran into

Our biggest challenges were integrating multiple APIs smoothly and handling incomplete or uneven datasets. Lighting data, for example, can vary drastically between cities, and crime data is often region-specific. Balancing performance with real-time API calls also required optimization. Additionally, tuning the safety-scoring algorithm to feel realistic and intuitive took a lot of iteration.

Accomplishments that we're proud of

We’re proud of creating a prototype that goes beyond basic navigation - one that prioritizes safety and empathy. Integrating multiple data sources and using Gemini for natural summaries helped make complex analysis understandable for users. We’re also proud that SafeRoute centers the experiences of women walking alone at night, turning data into actionable awareness.

What we learned

We learned how to combine geospatial data, machine intelligence, and UX design to create something meaningful. Working with APIs like Mapbox, Overpass, and Foursquare taught us how different data ecosystems can come together to solve real human problems. We also learned that safety is not just about metrics - it’s about context, communication, and trust.

What's next for SafeRoute

Next, we plan to integrate live data sources, such as city safety feeds, user-reported incidents, and real-time crowd density from social platforms. We also hope to launch a mobile version with a community reporting feature that allows users to flag unsafe spots or share updates about lighting and activity. Ultimately, we envision SafeRoute as a global safety companion - helping everyone walk with confidence, no matter the time or place.