The idea
Philita was a mobility startup I co-founded around a simple observation: the fastest route for one passenger is not always the best outcome for the whole transport network. We explored congestion-responsive journey planning across London’s roads, buses, trams, and trains, directing passengers toward routes that balance travel time and crowding. For commuters, the proposition was a more comfortable and resilient journey; for transport operators, it was a way to distribute demand more intelligently across existing infrastructure.
My contribution
I worked on the venture’s technical and commercial development, using Python, Pandas, CARTOframes, and TfL journey data to support transport modelling and simulation, alongside strategic discussions with TfL. Our team’s proof of concept modelled London’s roads and public transport provision and compared 100,000 simulated journeys with a parallel scenario using intelligent routing. The research framework used A* search with a Tung–Chew heuristic and a generalised journey cost, making the trade-off between speed and congestion explicit and testable.
The output
Initial prototype analysis indicated a 4% average reduction in busyness across London’s 100 busiest routes and up to a 15% reduction in peak-time congestion at London Bridge station within the model. These were simulation findings, not measured reductions from a live deployment. They gave us a quantitative basis for the startup’s value proposition and a research roadmap covering stronger commuter-behaviour models, disruption scenarios, and a proposed expansion to six million journeys. The lesson was how a scientifically testable routing decision could become a product opportunity for passengers and operators alike.