safely

Landing Page
Pathfinding
3D Visualization

Why safely❓

In a city grappling with high crime rates, safely empowers residents by transforming how they understand their surroundings. We're providing peace of mind, enabling users to generate safer paths and navigate Atlanta confidently.

How it works 💻

safely automatically clusters historical crime data in Atlanta with data provided by the Atlanta police department. It runs on the MapBox API (used by BMW, Rivian, Toyota) which provides seamless and responsive frontend interaction. Users are able to filter based on crimes, and also generate directions that avoid high crime areas.

The front end is built with React.js and styled using Chakra UI components. A custom pathfinding algorithm integrates with the Mapbox Directions API to guide users along safer routes, dynamically bypassing high-crime clusters. The geojson data is parsed and loaded as a layer on top of the MapBox visualization.

Challenges 👎

One of the most significant challenges we faced was the Mapbox Directions API's limitation, which allows avoidance of only up to 50 points per request. Given Atlanta's high crime density, this posed a serious obstacle to creating genuinely safe paths for our users. To overcome this, we needed to develop a sophisticated approach that adapts to different modes of transportation.

For driving, we adopted a broader clustering method, allowing for more general avoidance of high-crime areas since vehicles can cover greater distances quickly. For walking and biking, the challenge was even greater due to the need for more precise route adjustments—walking and biking users are more vulnerable and can’t bypass large areas as quickly. This required us to implement a dynamic filtering system that intelligently prioritizes the most dangerous clusters, ensuring that even within the API’s limitations, users can navigate safely.

What's next for safely? 🤔

Deploying safely as a mobile app would unlock real-time user location tracking, enabling seamless, on-the-go navigation and safer route adjustments based on dynamic user movement.

Additionally, while the current Atlanta PD website lacks direct access to live GeoJSON data, implementing a web scraping algorithm would allow Safely to stay updated with the latest crime reports. By continuously extracting and integrating this data, we could ensure that our crime clusters and safety recommendations remain accurate and up-to-date, providing users with the most reliable information possible.