Many tube stations are too busy to use much of the time, partly because there is uneven demand across tube stations. This is because there is no way for the user to plan their journey ahead of time and see which tube stations are busier (existing journey planners help you to plan your route, but don't tell you which stations will be busy). Data that shows the number of people at different tube stations would fix this problem, although the data would need to real time to make it useful to the user. Transport for London currently collects data from every ticket and Oyster card barrier at tube stations - we know this because they release an annual overview of the traffic at different tube stations. We've created a real time tube map with circles over each station growing larger as more people visit the station (because TfL won't release the data they collect, we've had to make up our owndata). We've written the web app in PHP, and it connects to a MySQL database, gets the required data and plots SVG circles over a tube map using it. We've used JS to make the map draggable, and the Google Charts API to integrate a chart showing usage.

Harry Rickards, Lawrence Job and Damon Hayhurst

Ideas for taking this project forward

If National Rail were to start collecting data from ticket barriers from around the country (we do not know if they already do), National Rail could be easily implemented within TubeSmart. TubeSmart could also be coupled with schemes such as Barclays Cycle Hire (@barclayscycle).

Estimated costs for taking this project forward

Very low. TfL already has the data available, they just need to use it, and if 2 developers and a designer could write TubeSmart pretty messily in a week then surely TfL could do similar (but with better code!).

About the data used for this project

Although we know that TfL collects data from the ticket and Oyster Card barriers at tube stations from the annual reports they release, TfL does not release this data. So instead we simply generated random data. If we had had more time we would have written an algorithm that would have made the random data more like real data, but as it was we just generated a random number between 200 & 1000 (for the graph we used a moving average of this number plus the previous two).

Project URL:

Twitter: @yrsbrighton

Created at

Young Rewired State 2010

Flickr Photos

Add photos by tagging them with rewiredstate:project="tubesmart"