Albatross

AI-powered navigation for safer, smarter cities using real-time crime data

Comment

Inspiration

There are many ingredients that go into Smart Cities. But one that often gets forgotten amidst the great technological advancements and futuristic dreams is safety. The reality is, without safety for everyone, a city can't truly be considered "smart." Many of us have experienced moments when our navigation apps have led us into areas that felt unsafe or uneasy. That is precisely what we are here to prevent.

What it does

Albatross, like the bird that never fails to find its way home, will use AI to help you reach your destination safely by calculating the safest and fastest routes. Was there a recent robbery? Albatross knows all and will reroute you accordingly! By aggregating criminal history data with real-time traffic and city layout information, Albatross provides quick, safe, and efficient routing.

How we built it

We utilized Databricks for data storage and processing, integrating crime data into a Delta Lakehouse to generate and update crime hot-zones. MLflow powers the clustering and creation of these zones, enabling vector comparisons and embedding models to categorize different types of crimes. Cloudflare Workers handles the routing and navigation due to its speed and lightweight capabilities, while the Here Routing API was leveraged for optimal pathfinding.

Challenges we ran into

Handling large sets of real-time crime data through new platforms like Databricks and Cloudflare Workers was a lot of fun but challenging. Designing a system that could efficiently process a steady stream of crime data and dynamically cluster it into hot-zones was quite complex, and took a lot of iterations. We also faced issues with flexible polyline encoding, which required workarounds due to limited library support.

Accomplishments that we're proud of

Despite the challenges, we’re proud to have completed a project with all the core functionalities we set out to achieve. We embraced the spirit of hacking by diving into new technologies and implementing them successfully within a short timeframe.

What we learned

We gained experience with various frameworks and tools, including Databricks, Cloudflare, and geo-routing APIs. This project also allowed us to revisit previous tech stacks such as Vue.js, Scala, and data warehousing.

What's next for Albatross

Safety is a never-ending battle. We plan to introduce additional features such as real-time notifications, personalized safety recommendations, and advanced AI crime prediction models. Our goal is to continue enhancing urban safety, paving the way for a true Smart City.

Share this project:

Updates