Once Again, DNAinfo Lowballs Ridership for a New South Side Divvy Station

IMG_4125
The station at 76th and Jeffery in South Shore. Photo: John Greenfield

I appreciate DNAinfo’s efforts in crunching Divvy ridership data to produce a series on usage patterns in neighborhoods like Uptown, Logan Square, and Lakeview. However, their method has resulted in a couple of articles that dramatically underreported the number of rides taken to and from new South Side stations.

On Friday morning, DNA ran a piece by Andrea V. Watson that initially claimed Englewood’s five bike-share stations have only been used a total of 267 times since they opened this spring. The station at 56th Street and Halsted has seen a mere 22 trips, Watson reported.

After I double checked the numbers with Divvy management and the Chicago Department of Transportation, it became clear that the DNA numbers only accounted for rides taken by June 30. That’s the latest date for which ridership data is available for download on the bike-share system’s website.

However, more recent data was readily available from CDOT and Divvy, if DNA had asked  for it. It turns out that the five stations were actually used a total of 798 times by September 30, about three times what DNA reported. The station at 56th and Halsted has actually seen 107 trips this year.

Old Divvy
The station at 76th and Jeffery is one of the southernmost in the system. The fact that it’s on the outskirts of the coverage area helps explain its low ridership numbers.

While it’s true the Englewood ridership numbers are low, they’re not nearly as dismal as what Watkins originally stated. DNA eventually edited the piece to note that their initial numbers only covered the roughly 2.5-month period between station installment and June 30, and they updated the article with the more recent numbers.

This morning, DNA made the same mistake again. An article by Sam Cholke about Divvy use in South Shore included the interesting observation that most local residents are using the bikes within the neighborhood, or to get to the lakefront or Hyde Park.

However, the piece stated that a station at Jeffery Boulevard and 76th Street, one of the southernmost in the system, has gotten little use, “with 70 rides so far this year.” “Perhaps downtown is just too much of a slog from Jeffery Boulevard and 76th Street,” the author theorized, noting that the station is mostly used for trips to or from seven stations in or near Jackson Park.

That figure also seemed suspiciously low, so I checked it with Divvy general manager Elliot Greenberger. Again, it turns out that DNA only used numbers for trips taken by June 30. Since the 76th and Jeffery station debuted on April 23, what Cholke called “so far this year” was really only about nine weeks. Here are the more recent figures for the station:

Screen Shot 2015-10-19 at 5.03.31 PM

With 195 trips taken to or from the station this year, ridership is – once again – nearly three times what DNA reported. It’s low, but it’s not that low.

As with the Englewood article, analyzing ridership data for a roughly two-month period after a station debuts isn’t very useful. And when an article suggests state that those numbers represent the entire year, that’s misinformation.

That said, CDOT and Divvy could help prevent this problem in the future by releasing ridership numbers more frequently. Currently, Chicago bike-share data is only released to the public twice a year, so only the first two quarters of 2015 are currently available on the Divvy site.

Instead, they should release the data each quarter, at a minimum. Washington D.C.’s Capital Bikeshare and New York’s Citi Bike are both operated by Motivate, the same parent company as Divvy, and they release data more frequently than Divvy. In D.C., it’s released every quarter, and in NYC it’s released each month.

  • cjlane

    “With 195 trips taken to or from the station this year, ridership is – once again – nearly three times what DNA reported. It’s low, but it’s not that low.”

    Well, to have some context on whether it is low, *that* low, or just kinda normal, shouldn’t you compare it to other “edge” stations? I’d be interested in a comparison to the south Halsted stations, the most westerly Lawndale and Humboldt Park stations, and the more remote Albany Park stations–especially any that also opened this year.

    If some/many of those have well under 500 trips, 76/Jeff is “low”, but perhaps only barely so. If the lowest other ‘edge’ location is north of 1,000 , then it’s certainly low and may be *that* low. If the lowest non-south-of-Hyde-Park edge station had 2,000 trips, then it becomes a very fair question.

    Adding capacity were there is not so much nominal use is ok, but adding it where it is something like 10% of use elsewhere (in semi-comparable locations–ie, other edge spots) can be a suspect allocation of resources.

  • I get the feeling that a science of bike sharing is still a ways off. It seems that there is still a lot of experimentation needed.

    The other significant factor is station density. I believe that the new edge stations up north suffer a lack of density as well.

  • cjlane

    Yeah, of course. And I’m curious if the edge stations, relatively (and actually) far from the lakefront, all suffer similarly.

    I would guess that they do, and some of them are probably pretty close to as ‘slow’ as the Jeffrey stations, and that then invalidates singling out the Jeffrey stations–it becomes just an issue of the inherent nature of the Divvy frontier. But if my hunch is wrong, and the Jeffrey stations have 90% less traffic than the other edge stations, then it is a genuine question of the wisdom of resource allocation.

  • skyrefuge

    DNAinfo provided a map that nicely answers your questions here: http://www.dnainfo.com/chicago/20151012/downtown/how-chicagoans-use-divvy-see-where-bikes-from-your-station-went-map

    In general, it looks like the new south side edge stations got about 20-30% of the traffic of the new north side (Rogers Park, Albany Park) edge stations. Which in turn get about 10% of the traffic of a busy core station.

  • cjlane

    cool. thanks for the link.

    Now, would be interested in seeing them adjusted for population and destination (other stations, shopping/recreation nodes, etc) density.

  • Keep in mind that the DNA map only uses data from July 1, 2014 to June 30, 2015. The new stations were installed in spring 2015, so it’s pretty useless to compare their ridership data for new stations, which only covers two or three months, with their data for older stations, which covers 12 months.

  • Anne A

    From my experience in using stations at the edge of the system, it seems pretty normal for usage to be lower there.

  • Except that (a) the data only goes to June 2015, and (b) a lot of the northside edge stations were there for several months longer than the southside ones.

    They don’t note when it was installed. IMO every station that was there for a subset of their dataset (July 2014-June 2015) should be marked somehow, even if we can’t necessarily see month of installation.

    If I’m recalling correctly, the Keystone/Montrose station went in in late April/early May, but some of the others went in in fall 2014, which will pump your numbers appreciably.

    Most of the southside stations were only there a month or two in the data DNA provided, which is misleading when you compare to 6mo+ stations.

  • cjlane

    Looks like most of the low usage northside stations show zero rides in 2014 on that map, so not as incomparable as you assume.

  • cjlane

    “IMO every station that was there for a subset of their dataset … should be marked somehow, even if we can’t necessarily see month of installation.”

    As I note to John, the Southside stations at issue, and the comparable edge Northside stations, show zero rides in 2014, so they are ‘marked’, even tho it is sorta subrosa.

  • skyrefuge

    I’m not as dumb as a DNAinfo reporter, so my rough analysis posted above had already factored in station duration and everything else you mentioned. My “20-30%” number came from only looking at 2015 data, and I made sure to only compare stations that were part of the spring 2015 expansion. The northern edge stations did not go into place substantially earlier (if at all) than the southern edge stations.

  • By older stations, I mean ones that were installed in 2013. 2015 station installation began on the South Side, then reached the West Side, and ended on the North Side.