Proximity indicators

People’s ability to reach services with limited mobility is one practical way to describe accessibility. In this case, we focus on a proximity indicator for Accessible subway stations: the share of residents who live within a typical walking distance of an ADA-compliant subway station.

“Proximity” indicators - the percent of a population living within walking distance to XXX.

  • Link to indicators
  • Link
  • Link

So how do we calculate that?

1) Identify ADA-accessible subway stations

We start with the MTA Subway Stations dataset and keep only stations flagged as ADA-accessible. These stations are the points where people with mobility needs can actually get into the subway system.

Map of ADA subway stations

2) Define walking distance from each station

Instead of drawing circles, we use the walking-network from OpenStreetMap. With Dijkstra’s shortest-path algorithm, we find all places reachable within 800 meters of walking distance from each station entrance.

Map of 800 m isochrones

3) Combine all station walkable areas

To keep things simple and avoid double-counting, we merge all the individual walkable polygons into a single unioned map of “all areas within 800 m of any ADA station.”

Unioned isochrone map

4) Estimate the population in the walkable area

Who actually lives inside this area?

  • We start with Census Block Groups (BGs) for population totals.
  • Because block groups are large and might cross the walkable area, we use PLUTO residential units to figure out how many people are likely inside that area.
  • If a quarter of the residential units in a block group are inside the walkable area, we assume about a quarter of the population is too.

This gives us a percent of the population within walking distance for every block group.

Map of block groups by % population within walking distance

5) Roll up to larger geographies

Since block groups nest into larger areas, we can add up the results step by step:

  • Census Tracts: Block groups form tracts. By summing BG estimates, we get the percent of each tract population within walking distance.
    Output: 800m_CT_pct_walkable_ADA_subway.geojson

  • Neighborhood Tabulation Areas (NTAs): Census tracts form NTAs. Aggregating tract results gives us the percent of each NTA population within walking distance.
    Output: 800m_NTA2020_pct_walkable_ADA_subway.geojson

  • Community District Tabulation Areas (CDTAs): NTAs form CDTAs. By summing NTA results, we can show the percent of each CDTA population within walking distance.
    Output: 800m_CDTA2020_pct_walkable_ADA_subway.geojson

Map of aggregated tract, NTA, and CDTA results

6) Handle UHF42 neighborhoods

UHF42 neighborhoods are different: they don’t line up neatly with Census geography. To estimate population here, we had to split block groups across UHF42 boundaries:

  • First, we identify the share of each block group’s residential units that fall inside a given UHF42 area and inside the walkable area.
  • Then, we allocate both the block group’s total population and its “walkable” population across those shares.
  • Finally, we sum across block groups to get the percent of each UHF42 population within walking distance.

This ensures people are only counted once, in the right place, even when geographies overlap.

Output: 800m_UHF42_pct_walkable_ADA_subway.geojson

Map of UHF42 results

What this yields

An interpretable proximity indicator:
Percent of residents within a 800 meters walk of an ADA-accessible subway station, reported for Block Groups and aggregated to Tracts, NTAs, CDTAs, and UHF42.

The same method can be reused with different distance cutoffs or with different resources (e.g., audio bus stops, libraries, clinics).

Published on:
January 7, 2025