As users take bikes between Capital Bikeshare stations, some stations become unbalanced. Some stations run out of bikes, because they have more riders checking bikes out than riders checking bikes in. Other stations fill up, and when there are no empty docks riders can’t return the bikes. The system thus depends on a crew traveling to the unbalanced stations, transferring bikes from full stations to nearly-empty ones.

I used the open data from the first quarter of 2012 to create an animation, showing how stations gradually become more unbalanced. Beginning from January 1, 2012, stations that have more check-outs than check-ins are shown with red circles, while green circles are shown around stations that have an excess of check-ins. (Overlapping regions blend into yellow.) The size of the circle reflects the lopsidedness of the station’s usage.

When the Bikeshare van transfers bikes, that data is not recorded, so it does not affect the visualizations I created. (If it were included, none of the circles would grow beyond a few pixels.)

The 3-second animation below samples the data once per day, showing a steady growth among the stations most in need of re-balancing. The calculations are cumulative over the 91-day period.

Another version was made from the same data, but sampling every hour. The frame rate was increased from 30 frames per second to 60 FPS. The source video grew from 1.7 MB to 42.4 MB, so I am posting this version via YouTube rather than embedding as QuickTime. In this version, the stations seem to pulse or “twinkle,” since many stations have their own pattern of unbalanced demand over the course of a day (shown in about a third of a second).

[Update: I now have an HD version of this video, at 30fps, and thus twice as slow]

The region includes Washington, DC, and Arlington, Virginia. Customers made 364,757 trips in this quarter. Since their trips aren’t symmetrical, the CaBi staff must re-balance stations by driving their van around and transferring bikes. The system works only if users can be confident that stations will be neither empty nor full.

The most unbalanced station in this period is the one at 16th & Harvard NW (near Columbia Rd, east of Adams Morgan), which had 6,141 bikes checked out, but only 3,283 bikes checked in. This means the CaBi re-balancing crew had to deliver 2,858 bikes to that station (on average, that’s more than 31 per day). Notably, this station is atop Meridian Hill, which might explain the enthusiasm for riding downhill from the station, and avoiding the uphill climb back.

At the other extreme is the much busier station on Massachusetts Ave NW just west of Dupont Circle. It had 13,201 check-ins and 11,248 check-outs, meaning the CaBi staff had to make room by removing 1,953 bikes (an average of over 21 each day).

For similar animations made from CaBi data, see Variations on Displaying Data. Like the other videos, these were made using Processing (see Creating Animation with Processing). My CaBi Dashboard web app shows a live display of how unbalanced stations are (see CaBi Dashboard with Google Chart Tools for more info).

Watching Bikeshare Stations Grow Unbalanced

3 Responses to “Watching Bikeshare Stations Grow Unbalanced”