Overview

What is a subway desert?

While there is no standard definition of a subway or transit desert, for this project, I defined it as an urban area within New York City located more than one mile from a subway station. The one-mile distance was selected as the walkshed threshold, consistent with common definitions used in transit planning. However, roughly 1.2 million people live more than a mile from a subway station, creating significant friction for regular subway use. To put these numbers into perspective, I created a map in Tableau highlighting MTA “deserts,” with the goal of helping the city identify opportunities for new routes or lines to better cover underserved areas.

Process

What is a subway desert?

While there is no standard definition of a subway or transit desert, for this project, I defined it as an urban area within New York City located more than one mile from a subway station. The one-mile distance was selected as the walkshed threshold, consistent with common definitions used in transit planning. That said, a one-mile radius is relatively generous. Comparable research sometimes uses shorter distances, 0.3 or 0.5 miles, which would result in much larger transit deserts. If we assume one mile equals roughly to a 15-minute walk, it is debatable whether this distance is acceptable for many potential riders. Further research would be needed to determine a realistic threshold.

The Data

To generate a map that visualizes subway deserts in Tableau, several datasets were required. First, a shapefile of New York City Neighborhood Tabulation Areas (NTAs) was used to define neighborhoods. Additional datasets included subway station locations and census population counts per neighborhood. By connecting these datasets, the project becomes feasible. For transparency and reproducibility, the data sources are listed below:

1. Subway station locations with coordinates (data.ny)
2. NYC neighborhood boundaries as polygons (NYC DCP)
3. Population counts per neighborhood (NYC OpenData / Census)

Let’s Build it
While identifying the correct datasets was more challenging than expected, building the map itself was relatively straightforward. The NTA shapefile was joined with the population CSV using NTACode = NTA Code, allowing me to determine how many people live in each neighborhood. The subway station data was not joined, as it sits as a separate layer on top of the neighborhoods. The one-mile radius was created using the BUFFER() calculated field, which allowed the map to take shape quickly.

The map was styled in dark mode to help viewers focus on the key elements. Neighborhoods are colored along a red gradient, where light orange represents lower population density and dark red represents higher density. This makes it easier to identify which subway deserts are most affected. Buffers around subway stations are shown in black with 60% opacity, illustrating which neighborhoods are well served (those fully covered by buffers) and which remain underserved (those still visible outside the buffers).

The Map

https://public.tableau.com/app/profile/federico.sarno/viz/NYC-MTA-Deserts/Sheet1?publish=yes Key

Findings
Subway deserts are primarily concentrated in the outer boroughs. The most populated subway deserts are located in Queens, Brooklyn, and the Bronx, with none in Manhattan. Staten Island’s North Shore is almost entirely a subway desert by design, as the borough is served by only one line (the SIR).

Some neighborhoods are only marginally underserved, while others are significantly affected. For example, Flatlands (Brooklyn) sits at 1.03 miles from the nearest station, just beyond the threshold, whereas Bayside Hills (Queens) is nearly three miles away.

Reflections
There are several important considerations when interpreting this visualization. First, neighborhoods are classified as subway deserts if they are more than one mile from a station. However, due to data limitations, some neighborhoods fall into a gray area where only portions seat beyond the threshold. This makes it difficult to definitively categorize them as deserts or not. Second, using a one-mile radius simplifies real-world travel. Because of varying street grids, actual walking distances may differ significantly from straight-line distances. Some areas within the radius may, in practice, require longer travel times. Finally, this visualization focuses solely on the subway system and does not account for other public transportation options, such as buses or the LIRR, which may help fill some of the identified gaps.

If I were to extend this project, I would incorporate bus stops to broaden the scope from subway deserts to overall transit deserts. This would provide a more accurate picture of which neighborhoods are underserved by public transportation as a whole. Additionally, I would use OSM-based street-network walking distances instead of straight-line buffers to better reflect real travel conditions. Finally, I would explore socioeconomic overlays to analyze whether transit access correlates with income or race, for a much stronger and more meaningful analysis.

Transit access shapes where people can live, work, and spend time. I built this visualization in the hope that it will help future city planners create a better subway system.