Income, Race, Bikes

Alex Berke

Is the placement of bike-share docks equitable?

This interactive map explores the question visually.

The map shows the addition of bike-share docks, as well as the changes in income and race throughout the years the bike-share program has been in service. You can toggle the display to see only income, or race, or bikes data, or any of their combinations.

The project includes bike-share networks for the cities of New York, Boston, Washington DC, Chicago, and Philadelphia.

About The Data

Income and race data is provided for each year by census tract. It is from the American Community Survey (5-year), which provides yearly population estimates.

The data used for this map is household median income, and household race. 
*Race is displayed on the map as “Percent White,” which is calculated as the percentage of households that report “white only” in a given census tract.

Bike station location data is scraped from the bike-share company’s websites.

For more about the data scraping and processing, see code and documentation.


The question of who bike-share systems serve and have the potential to better serve has been asked and answered before.

For example, a 2019 report found that New York’s Citi Bike system disproportionately serves New Yorkers of higher incomes and those who already have access to other transit options, and disproportionately serves few New Yorkers of color. It reports a median household income of $90,400 in service areas versus $54,700 outside service areas. It also found that the Citi Bike service area was “twice as white” as the rest of the area.

Yet US studies have shown that people of color and people from the lowest-income households are in fact most likely to ride bicycles.

This map explores data throughout the bike-share program years of service because, while equitable placement of docks may be improving, future systems can aim to avoid patterns of servicing privileged neighborhoods first.

Where to optimally place bike-share docks is a complex problem, and different cities have various ways of sourcing new locations.

For example, the New York City department of transportation previously launched an online map to crowdsource suggestions, and there is a similar suggestion map for Boston’s bike-share as well. However, a study published in the Journal of the American Planning Association showed that only 5% of new docks placed in the 2014-2015 New York Citi Bike expansion were located within 100 feet of suggested locations. This does not imply that the NYC’s DOT ignored the suggestions. Optimal dock locations depend on a variety of factors, such as proximity to other transit options, or the viability of removing parking or sidewalk space to make way for docks, as well as how additional docks will complement the placement of existing docks.

In the Boston area, the Bluebikes bike-share system (previously Hubway) is owned by a partnership of multiple adjacent municipalities: Boston, Brookline, Cambridge, Somerville, and recently the city of Everett joined the partnership. The city of Cambridge actively welcomes businesses and property owners to sponsor the installation of docks near them, creating another mechanism to source the location of stations.

Other dockless bike-share and scooter systems are entering the landscapes of these cities as well, such as Ant bicycles, or Lime scooters. They may have the potential to reach the areas that city-owned bike-share systems have not, but they do not openly provide their data (as of 2019). 

The companies that operate the city-owned bike-share systems such as those in New York (Citi Bike) and Boston (Bluebike) are required by contract or local law to compile and provide usage data. We can imagine cities compelling the other bike, scooter, and micromobility providers that operate within them to openly provide their data as well. This data would help city planners, traffic engineers, and researchers better understand the impacts and potential for these new mobility systems on our streets.


This project was updated summer 2022 with contributions from Walter Truitt.

All code, data, and data processing scripts for this project are open source: 

Contributions welcomed.

Thank you to contributions from @stevenleeg.