Citi Bike data errors

[Zach Rausnitz]

If any of you are working with the trip history data that Citi Bike released recently -- I've been finding a lot of flaws in the data. More details here: http://www.rausnitz.com/blog/2014/04/bad-data/

I'd be interested to know if any of you have found the same issues.

[Noel Hidalgo | BetaNYC]

It appears that this is part of rebalancing. You should come to our CitiBike Hacknight where we're going to discuss this data.

http://www.meetup.com/betanyc/events/174010072/

- Noel

[Dani Simons]

The mostly likely source for these errors is that when we move a docking point from one station to another, these re-located docking points sometimes "think" they are still part of their original station. So, for example, a trip might look improbably short between say Williamsburg and Hells Kitchen, but it might just be because we moved a docking point from Williamsburg to midtown and the docking point still thinks it's in Williamsburg. 

This issue will be resolved in an upcoming software update. 

Best, 

Dani

On Tue, Apr 8, 2014 at 9:40 AM, Noel Hidalgo | BetaNYC <no...@betanyc.us> wrote:

FYI. 

--

Sent from my TI–85

Begin forwarded message:

--
You received this message because you are subscribed to the Google Groups "NYC Bike Data Hackers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to citibike-hacke...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/citibike-hackers/192ef373-e9ea-4ca5-b8ed-2930a56064ff%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Dani Simons

Director of Marketing and External Affairs

NYC Bicycle Share, LLC

  Operator of Citi Bike

5202 3rd Ave. Brooklyn, NY 11220

[Zach Rausnitz]

Makes sense. Thanks Dani -- I really appreciate the info. Do you know if Citi Bike will be able to release updated trip data for July 2013 thru Feb 2014, or will the software fix only affect data on trips that occur in the future?

Zach

[Abbott Katz]

There also appear to be some date of birth errors; see my blog post at

http://spreadsheetjournalism.com/2014/05/15/whats-up-dock-new-yorks-citi-bike-data-part-1/

On Tuesday, April 8, 2014 1:56:12 PM UTC+1, Zach Rausnitz wrote:

[R Antonio]

Is there any clue as to whether this will be corrected by Citi? or Bikeshare?

[Guilherme Oliveira]

Hi all.

I've been collecting the stations' footprint json at intervals of 30 seconds and found some inconsistencies when comparing it with the trip history dataset.

When aggregating both data in intervals of 10 minutes, there is usually more registered trips than activity (according to the json footprint), even though both barcharts show the same curve (correlated).

Since rebalance and 'false' trips are removed from the trip csv files provide, the number of trip should be usually less than the amount activity implicit in the json footprint, right?

I thought it might be  due to my sampling rate not being high enough (the 30 segs interval between jsons), but it happens even at times of low activity, and at stations that are not that popular..

Also, i've been using the json feed from the http://api.citybik.es/citi-bike-nyc.json    not the original one (http://www.citibikenyc.com/stations/json).

I've compared recent instances of both and the data match (number of available bikes and slots), but I found that some of the stations had info, that should be static, modified over the months (like name and geo-coords).

I dont have older citibike original json instances (http://www.citibikenyc.com/stations/json) to verify this, but I'm wondering if the stations have really been moved around or if it is some mistake made in the http://api.citybik.es/ API

Anyone?