Hackathon · 2026 · Claude Hackathon · Imperial College London 2026
DarkFleet
A dashboard that tracks fishing vessels across seven public data sources and flags the ones most likely to be fishing illegally.
6th of 64 - Claude Hackathon, Imperial College London
Top-10 finish out of 64 teams at the Anthropic-sponsored hackathon at Imperial College London.
Python · Claude API · PostGIS · Mapbox
The problem
Illegal fishing is a 23 billion dollar business, and the fleets hide by spreading their tracks across separate data sources: switched off transponders, changed flags, opaque owners. No analyst can cross check all of it in real time.
What I built
I pulled seven open data sources into one dashboard that scores every vessel on seven risk signals and ranks the global fleet. An analyst can then ask Claude follow up questions about any ship and get a full report.
What it does
- Seven Risk Signals. Encounters at sea, gaps in the tracking signal, flag changes, opaque owners, loitering, licence status, and sanctions, all scored together.
- Ranked Alerts. Every vessel comes out red, amber, or clear, ranked across the global fleet in real time.
- Ask Claude About Any Ship. An analyst can ask follow-up questions about any vessel and get a structured assessment back.
- One-Click Reports. Everything it knows about a vessel turns into a PDF brief you can share.
Photos
Gallery
The full engineering is on GitHub.