On July 20, 2016 the Jersey Journal reported the rollout of 15 new Citi Bike stations and 150 new bicycles. The Journal states that of these, nine stations will be concentrated in the Eastern, Downtown neighborhood, with the nearby Bergen-Lafayette neighborhood (in the central West) receiving three stations, and the southernmost Greenville and northernmost Heights neighborhoods acquiring a meager two and one new station respectively. What are the drivers of these station location decisions and what do they mean for the city’s diverse communities?
With just a look at a map of its land uses, it is apparent that Jersey City — in its entirety — is well-suited to a successful bike-share system. From the Southern-most neighborhood of Greenville to the Northern-most neighborhood of The Heights, the city features only high density, residential housing heavily intermixed with commercial spaces, with parks and public transportation interspersed. Herein, are the prerequisite key factors for successful bike share: high population densities in a mixed-use setting with many people walking, biking, and using public transit.
However, despite the prime conditions across Jersey City for bike share, from its initial launch in September of 2015, to the recent addition of 15 new stations, Citi Bike has clearly and strongly favored the Eastern, Downtown neighborhood. A map of stations and lanes even reveals a drastic difference in the city’s bike lane system, with the Downtown neighborhood serving as the only truly (or even partially) well-connected portion of the network.
Data-driven Decisions
The Jersey Journal points to trip and user data indicating that the majority of Citi Bike users live Downtown and the fact that an online survey was provided to garner resident input on where to put stations as the primary justifications for the locations of the 15 new stations. Given the well-documented barriers to bike-share usage and the limitations inherent in online surveys, using these two items alone might serve to increase rather than decrease inequities within the system.
Research indicates that station density (or how close and well-connected by bike lanes one station is to another) and the proximity of stations to potential users’ homes heavily determine whether bike share is widely adopted (1, 2, 3, 4, 5). Some even suggest that in an optimal system, stations are located roughly every 4 blocks (300 m) and stations in relative isolation are essentially un-usable (2). When viewing a map of Citi Bike stations, the Western and particularly Southwestern (Greenville) part of the city clearly suffer from extremely poor bike lane connectivity and low station density. Both prior to and increasingly after the addition of new stations, Downtown is the only region where the bike-share system is optimal and readily usable. If the system is designed — intentionally or no — to be less useful to other neighborhoods, it is no surprise then that Downtown would be were a majority of users live. Trip data might only reflect the inefficiencies of the system instead of where all potential and real users actually live.
A quick perusal of the survey open to residents indicates that the other neighborhoods in Jersey City might in fact find Citi Bike highly useful if it had better bike lane connectivity and station density, as there are dozens of stations requested outside of the Eastern, Downtown area. However, here too this demand might be understated. Online surveys by their very nature are opt-in, meaning that instead of acquiring feedback that represents all of Jersey City residents, only those who purposefully seek out the survey and have access to the internet (and devices that make such survey response easy, unlike mobile devices) will provide input (6). Just as with trip data that is biased by the system’s initial design, here again relying on such responses will likely skew the system towards those who are younger, and have greater affluence and leisure time — i.e. time and ability to seek out and use the survey.
Citi Bike for Whom?
To garner an understanding of the populations favored by the new (and to a certain extent, old) stations, comparison is needed to the City’s population density, household poverty and income levels, and racial and ethnic population densities. When looking at stations relative to population density, the stations and the new ones in particular, unduly favor the Eastern, Downtown neighborhood. Areas of contiguous higher density in the North-, Central- and Southwestern portions of the city either have no stations located within them, or a paltry few in comparison.
When it comes to households falling below the poverty line, both new and old stations do serve areas with higher levels of poverty. However, here again downtown is clearly favored compared to areas with higher poverty levels in the Western portion of the city.
Perhaps more revealing is a comparison of station locations to median household income levels, as well as racial and ethnic population densities. In the former, it is clear that Downtown has more consistent and adjacent block groups of higher median income levels. Despite the addition of the five rather spread-out stations in the Central- and Southwest that are located in lower income areas, the drastically higher addition of nine stations to the east likely reflects a system heavily biased towards higher incomes. The two high income zones on the Southeastern and Southwestern border, are industrial marshland, and a trade port and golf course — as indicated by their lack of roadways — and are therefore disregarded here.
Lastly, when stations are viewed in consideration of the racial and ethnic backgrounds of nearby populations, it is clear that distinct groups are underserved by the Citi Bike system, even with the additional stations. Despite the two new stations to the Southwest (Greenville), this community, consisting mainly of those who identify as Black or African American and non-Hispanic continues to have the lowest station density, making the system far less useful to these residents. Comparatively the Eastern, Downtown area which is heavily White, non-Hispanic and Asian continues to have the densest and most usable portion of the bike-share system. The heavily Hispanic and somewhat White, non-Hispanic Northwest (Heights) and more heavily mixed Central-west (Journal Square) have comparable station densities albeit with clear gaps and nowhere near the density of Downtown.
Conclusion
As it stands, the current Citi Bike system in Jersey City clearly over-proportionally favors the more affluent and White, non-Hispanic Eastern, Downtown neighborhood. The recent addition of 15 new stations likely perpetuates rather than alleviates this bias. However, an overreliance on trip data and the use of an online survey to garner public opinion may have led system administrators to bolster system inequities. Given the city’s high population density, heavily mixed land use, and multiple public transit networks future additional stations can and should capture a high number of potential users in the currently underserved West, and particularly the Southwest. Through opening up public feedback beyond the online survey and using land use, population and public transit geographical information in addition to trip data, the system can better serve all Jersey City residents. In this way, the well-documented benefits of bike share — including greater health, access to jobs and increased local spending (7,1) — can help the entire city prosper. Such a bike-share system would serve as a stellar example for other municipalities across New Jersey, and help advance the NJ Department of Transportation’s “focus on bicycling and walking as elements of active living and healthy life styles, generat[ing] wider support for implementing bicycle and pedestrian improvements” (8).
By Michelle Mayer
Sources:
- Sato, Hitomi, Tomio Miwa, and Takayuki Morikawam. 2015. A study on use and location of community cycle stations. Research in Transportation Economics 53: 13-19.
- García-Palomares, John Carlos, Javier Gutiérrez, and Marta Latorre. Optimizing the location of stations in bike-sharing programs: A GIS approach. Applied Geography 35: 235-246.
- Fishman, Elliot, Simon Washington, and Narelle Haworth. 2014. Bike share’s impact on 12 car use: Evidence from the United States, Great Britain, and Australia. Transportation 13 Research Part D 31: 13-20.
- Lin, Jenn-Rong, Ta-Hui Yang, and Yu-Chung Chang. 2013. A hub location inventory model for bicycle sharing system design: Formulation and solution. Computers & Industrial Engineering 65: 77-86.
- Krizek, Kevin, Gary Barnes, and Kristen Thompson. June 2009. Analyzing the effect of 25 bicycle facilities on commute mode share over time. Journal of Urban Planning and 26 Development 132 (2): 66-73.
- Dillman, Don A., Jolene D. Smyth, and Leah Melani Christian. 2014. Internet, Phone, Mail, and Mixed-Mode Surveys: The Tailored Design. 4th John Wiley: Hoboken, NJ.
- Shaheen, Susan A., Stacey Guzman, and Hua Zhang. 2010. Bikesharing in Europe, the Americas, and Asia: Past, present, and future. Transportation Research Record: Journal of the Transportation Research Board 2143: 159-167.
- New Jersey Department of Transportation. NJ Statewide Bicycle & Pedestrian Master Plan: Phase 2. 2004. 4. https://www.state.nj.us/transportation/commuter/bike/pdf/bikepedmasterplanphase2.pdf