I’ve been working with the TfL data over the past few weeks to better understand the geographical distribution of journeys and how this affects the load on individual segments across the transport network. In other words, we all know that some parts of some trips are particularly unbearable, and that disruptions in those areas (e.g. cancellations, delays) can seem to bring the entire system to a halt in a matter of 15 minutes.
The visualisation below shows how some links in the network are particularly heavily-loaded at various times of the day because of the way that people are travelling around the network based on the probable shortest path that they would take between points A and B. In other words, if someone boards the Tube at Bethnal Green at 9am and exits at Shepherd’s Bush then, in all likelihood, they have travelled along the Central Line and the map below would increment the counter for each segment in the 9am time interval between stations by one. Summing up over all travel for working days produces this time-based visualisation.
Personal experience suggests that the prominence of the leg between Highbury and Kings Cross is right on the money — the Victoria line at this point becomes nearly unbearable, but the train empties out rapidly by the time you approach Oxford Circus. The busiest segment of all looks to be between Kings Cross and Euston, but it’s followed rather closely by Kings Cross to/from Farringdon. I want to emphasise, however, that I’m not saying that most people are starting their journey at Kings Cross and finishing at Farringdon; rather, they are taking a trip that is likely to pass through those stations, and that’s why those lines become so heavily-travelled.
Because of the way that London’s transit system works, for now only the rail and Tube portions of our journeys are mappable. But in the longer term I hope to come up with ways to infer data about bus journeys based on the ways that people surge on to alternative routes or modes when delays pass some threshold of bearability.