Golden Eye

Feeling safe in a city means having early awareness of hazards. It's about turning data into action—monitoring streets, holding officials accountable, and ensuring timely repairs.

Comment

💡 Inspiration

Late-night walks in the city can feel unpredictable. You might ask: “Is this street safe?” We wanted to create a way to visualize real-time urban data — from police presence to trash hotspots — to make navigating cities safer, smarter, and more transparent. With access to public cameras from OrchestraOS, we realized we could turn this passive feed into an interactive map for public awareness, city operations, and even analytics.

🗺️ What it does

Golden Eye is a mobile-friendly web app that shows a real-time heatmap of key urban events:

  • Police activity
  • Homeless encampments
  • Trash accumulation
  • Parking availability
  • Public camera feeds

Users can:

  • View a live map on their phone
  • Filter by category (police, trash, etc.)
  • Tap into live street cameras
  • Submit and see community reports
  • Use AI-powered prompts to track custom objects or activity

🛠️ How we built it

  • Frontend: Angular + Mapbox for mobile-friendly mapping and UI
  • Backend: Node.js + Express API to serve geo-tagged data
  • AI/ML: Integrated LLM prompts + object detection to dynamically track user-defined categories
  • Cameras: Streamed real-time public feeds from OrchestraOS
  • Hosting: Netlify (frontend), Render/Azure (backend)

🧱 Challenges we ran into

  • Handling multiple live video streams efficiently in a browser
  • Building a generalized system to support arbitrary category prompts
  • Ensuring mobile responsiveness while overlaying dynamic data
  • Mapping object detection output to accurate heatmap coordinates
  • CORS and security policies for embedding camera streams

🏆 Accomplishments we're proud of

  • Built a fully functional mobile heatmap web app in under 48 hours
  • Integrated real-time public camera streams
  • Created a flexible LLM prompt system for tracking any user-defined entity
  • Delivered a clean, intuitive UX with real impact potential
  • Used open public infrastructure in a meaningful, civic-minded way

📚 What we learned

  • How to stream RTSP camera feeds into web apps securely and performantly
  • Techniques for spatially clustering real-time data (Supercluster, grid-based binning)
  • How to merge AI, geospatial data, and user interface design in a useful way
  • Designing for multiple users: public, responders, and city planners — all with different needs

🚀 What’s next for Golden Eye

  • Add report verification and upvoting to improve data reliability
  • Train custom vision models for local city issues (e.g. illegal dumping, blocked sidewalks)
  • Enable anonymous reporting from citizens directly from mobile
  • Partner with local governments for pilot programs
  • Build a progressive web app (PWA) for offline usage and home screen access
  • Explore privacy-safe tracking to detect movement patterns for safety analytics

Built With

Share this project:

Updates