At Control Group, we think a lot about how people move through shared spaces, particularly transit spaces. Understanding customer behavior has implications around advertising, transaction, station design and wayfinding. When we build products and experiences for shared spaces, we consider how they can intuitively meet the needs of customers without any direct interaction on their part. This is what we call a “responsive environment”– one that anticipates needs.
Our “On The Go” digital communication kiosks, for instance, could display real-time arrival information more often on a platform where we detect many passengers waiting for trains; or provide nearby points of interest with real-time arrival information on platforms where we detect that most people are exiting the train and leaving the subway system completely.
Currently, the only way to understand passenger flow at a subway station is via turnstile data. It’s a dataset the MTA has released on its open data portal, and one that hasn’t gotten a lot of attention. This is likely because there are some limitations of the data– it’s only updated every four hours and is aggregated by turnstile bank (usually one bank per entrance). Furthermore, there is often no connection between bank and line/direction, often providing little additional information about passenger behavior.
But we can still explore the potential of turnstile data to inform our efforts. We selected the 86th Street (4/5/6) station that has one turnstile bank per platform with no connection between the two. Since there are “On The Go” kiosks on each of the uptown and downtown platforms, what could looking at turnstile data do for us? Looking at the raw entry/exit data for the uptown and downtown platform at 86th Street, we found that the downtown side has many more entries than exits; on the uptown side we found the reverse:
Broken down into 4 hour blocks over a day, the data looks like this:
Rush hours see more entries and exits consistent with the general pattern; but there is frequent activity at the station throughout the entire day. What does it mean? What insights can we gain from this?
The high number of entries on the downtown side suggests most customers enter the station to wait for a train. The fewer exits from the downtown side tell us that most people do not get off here from stations above 86th Street. On the uptown side, the high number of exits suggests most customers get off here; the low number of entries tells us few board a train at 86th Street to go north. This suggests that our “On The Go” units could show more arrival information on the downtown platform. On the uptown platform, we could show arrival information coupled with information about the area surrounding the station– including special offers or featured destinations.
In the future, technologies like Bluetooth LE and other sensors will provide more granular data in real-time, allowing signage to be more dynamic and customized to the particular place in a station where the sign is deployed. Until then, the turnstile data we have suggests that there are “micro climates” of customer behavior out there that we have yet to understand and will need to be explored if we’re to build a truly responsive city.
Special thanks to the MTA and Chris Wong and Mike Mommsen for their data and tools used in this analysis.