Inspiration
Navigating public transportation, especially in cities like NYC, can be daunting and sometimes unsafe, especially for women, children, and tourists who may be unfamiliar with the area or concerned about safety. The lack of accessible, real-time safety insights for transit routes and the helplessness felt by us as travelers inspired us to create saferMaps - a tool that prioritizes safe navigation so users can take their safety into their own hands.
What it does
SaferMaps helps users navigate public transportation safely, by optimizing routes based on open-source crime data, MTA population density data, AI-driven analysis, and data collected from real-time CCTV camera feeds.
How we built it
We developed SaferMaps as a web application using:
Frontend: React for a dynamic and responsive user interface Backend: Flask to handle API requests, and integrate various data sources Database: MongoDB for scalability Data Sources: Sourced crime data from NYC Crime OpenData, MTA population density data from MTA Subway Hourly Ridership, and cameras feed from NYC DOT Real Time Traffic Information APIs: several Google APIs such as Routes and Maps for route calculation and visualization, and GroqAPI (llama 3.37 and llama 3.2-11b-vision ) to generate a route summary, and summarize images from cameras AI & Optimization: Used the A* algorithm to graph and find optimal routes
Challenges we ran into
Modeling graphs for routes proved to be difficult - we initially ran into the problem of getting significant overlap in graphs, so we switched to separate graphs and ran the A* algorithm on each one of them. This significantly increased our run time, which is something we hope to optimize more in the future. It was also difficult to handle the multiple APIs from Google to create a map and use polylines to display multiple routes on it, but eventually we figured out how to structure the components to make it work.
What we learned
The challenges of working with multiple APIs and databases efficiently, modeling graphs and using A* in a practical application, and making the UI more responsive while keeping the UX in mind.
What's next for saferMaps
For this Hackathon, we opted to build our solution as a web app, to quickly deploy the core functionality, an ideal starting point for our MVP. But we know the value of mobile-first experiences, especially for future real-time features we plan to implement like navigation, emergency alerts, and crowd-sourced data. We plan to provide native-like experiences, including offline use, push notifications, and home-screen access, which would be ideal for a transportation app.
We also plan to make it more personalized, allowing users to prioritize features such as crime density, population density, accessibility, etc., and give them more tailored recommendations based on user prefrences.
Log in or sign up for Devpost to join the conversation.