(edit: I’ve updated the video to include the river and the larger parks – thanks again to Ollie O’Brien for this)
I’ve been working with Ollie O’Brien here in CASA on this animation showing hire bike journeys on October 4th 2010, the day of a major tube strike, and hence a very popular day for the London Bicycle Hire Scheme.
We gleaned this data from the public available first one million journeys (actually nearly 1.4 million journeys) – detailing the start and end stands, start time and duration for each individual trip (this data runs from late July to November, but we’ve just looked at October 4th so far). Ollie has the positions of each stand, and was able to calculate the shortest reasonable route for each combination – as there are upwards of 350 stands, that’s over 100,000 potential routes! By dividing the route length by the duration of the journey, we were able to calculate an average speed for each trip.
In the visualization, the fixed circles represent stands – when a stand flashes red, it means that one or more bikes have left it – and a yellow flash means a bike has arrived. The bikes themselves are represented by the Boris Barclays Blue Tadpoles whizzing around – leaving at the right time, travelling at their correct average speed, and taking a (generally) realistic route.
Bikes which start and end and the same place are represented by the tadpole circling around the stand a few times and then disappearing – faster circling means the bike was not used for very long; a slower orbit means the bike was out for longer.
There are a few provisos – the route finding is not always 100% accurate (though it’s very good 99% of the time, but the Embankment proves confusing). I’ve yet to test the “number of bikes active” statistic properly – but I’ve reason to think it should be accurate most of the time. There is the occasional “blip” – bikes taking ten minutes to cross London – which could be due to bad data, or possibly the funny vans that redistribute them. We will no doubt tweak these on subsequent versions, and for the technically minded we can talk about the approach we used in more detail in future posts.
I think on that particular day I walked for 40 minutes before I could find a spare bike, and somewhere around Wenlock Rock I found a single usable bicycle in a rack of six dysfunctional ones and took furiously to the road. Hindsight is always 20/20.
Great work Martin! Is it possible to get a version with some background mapping? Difficult with the movie resolution I know, but would aid interpretation.
Also, for the future, interesting to think about relationships between bike infrastructure (lanes, cyclesuperhighways… ) and where cyclists really go/want to go.
Hi Duncan
Yep, this animation contains very little analysis – in order to get anything sensible for future planning, we need to do some serious aggregation of journey preference by time of day, total, etc. Some of that’s started already, and I’ll add it to the blog as we get more results.
I’m working with Ollie O’Brien to add a map backing – there are a couple of technical hurdles we haven’t yet overcome* but it’s on the to-do list.
*If anyone who is experienced in setting alpha values for images in OPENGL Processing sketches wants to point me in the right direction, it would be much appreciated.
Martin, really enjoy this. Can you give a brief rundown of the software you used to put this together? Very well executed!
-Anthony
Pingback: Bike Share Route Fluxes | Suprageography