The New York Citi Bike scheme is the largest collected at SchemeStats.bike, with over 6,000 stations and more than 23,000 vehicles. The scheme is very well subscribed, seeing over one thousand bikes in circulation simultaneously during the daily commute, and a strong usage pattern at weekends too. I look forward to an in-depth analysis of the scheme soon, but today I’d like to focus on how some poor weather recently affected the scheme.

The scheme is in a city very familiar with comparatively large snowfall: the first week in January alone saw a blanketing of snow. Around this time, real-time data from the scheme started to indicate several stations sporadically going offline. These stations started showing up as closed on the dedicated app, and users were unable to release nor receive bikes at those locations. The Citi Bike team released an update on Twitter on Jan 9^th:

A quick glance at the stations reveal that indeed they are accompanied by elevated solar panels, which no doubt suit this kind of low-power application:

Unfortunately, the poor weather meant that for many stations there simply wasn’t enough current flowing from the panels to keep them online, and they started showing up on our maps as closed. Here’s a snapshot from Jan 9, with closed stations in black:

So, how many stations were affected by this, and how long did it go on for?

SchemeStats.bike collects station status information every minute, thanks to the data made available publicly by Citi Bike. Looking at the data, we can see that the average daily quantity of closed stations saw a significant increase over the course of January, after increasing slightly in December:

The peak value of approximately 280 stations represents roughly 5% of all stations (see note 1), so this isn’t exactly taking the scheme offline by any means. By January 18^th, numbers of closed stations had returned to normal.

One point to note is that the effects of this were sporadic – it wasn’t the same set of stations that were offline day to day. To visualise this, here’s a timelapse December and January on the scheme, with closed stations in black, free spaces in white, and available bikes in blue. Each frame captures 12 hours elapsed time:

It should be said that the snow which visited the city later in January didn’t result in a similar spike in offline stations. This probably indicates that there is a combination of things which came together to take the stations offline: a long period of dark weather, short days, and snowfall on the panels.

Perhaps the main take away from this is that in weather conditions that would have ground many cities to a halt, NYC managed to keep over 95% of its bike share stations open throughout.

Looking forward to a more in depth analysis of this scheme soon!

Notes:

1. The baseline number of closed stations is approximate, as it is not possible for the data-collection bot to differentiate between a temporarily offline station and one that has been permanently removed. Nevertheless, the figures quoted are accurate to within approximately 20 stations.