• 

• 

  Conflicts view

• 

  Hotspots view

planNAV - Air Traffic Intelligence

planNAV is a collision detector and solver for flight schedules.

GH Repo: https://github.com/seofernando25/planNAV

Inspiration

I've always loved information-dense interfaces, but there is a small line 🤏 between "useful" and "overwhelming." With planNAV, I wanted to build something that spelled confidence while avoiding the "90s spreadsheet" look that a lot of those who use industrial software know.

How it works

It's built with a FastAPI backend. The frontend uses Mapbox for the 3D visuals and HTMX to keep interactions snappy.

We ingest a pool of scheduled flights and process them through a custom resolution engine to find "risk of collision" events. Beyond that, we can also visualize them with hotspots and a 3D timeline viewer inspired by Apollo in real time.

I also integrated a live aircraft image fetcher from Planespotters.net I figured if a controller is looking at a screen, they might want to know exactly what the plane looks like if they had to spot it with binoculars.

I used some of my experience from my university rocketry team (uORocketry) to build the collision engine.

To handle Continuous Collision Detection, I used a pruning-and-refining algorithm to process the dataset on the server-side decently fast. For the hotspots, I used Kernel Density Estimation (KDE) to turn the raw coordinates into a heat map of airspace risk.

If I had more time

If I had like… infinite time (or just actually forced myself to keep going 😅), I’d probably throw in some proper fuel burn models, real-time weather, maybe even run a few Monte Carlo sims to deal with all the random noise and uncertainty. But you know… I'm already 2 all-nighters deep… so this is the version you get ❤️

Built With

• fastapi
• htmx
• mapbox
• python