Despite having a Mayor who ran on a platform of affordable housing, the homeless population in New York City was at an all-time high this past April after a steady increase over 35 years. It’s obvious to anyone who takes the MTA that the shelter system is broken; The city is dependent on cost-ineffective hotel rooms to host the homeless — 70% of which are families. In 2017 De Blasio unveiled a solution: 90 community-based shelters dispersed throughout the 5 boroughs. Turning this policy into a reality has been a slow burn and has opened up a discourse about shelter footprints, racist planning, and NIMBY advocates. I aim to offer a visualization that examines 9 proposed shelters that have faced severe backlash in community districts with varying degrees of pre-existing shelter housing and their own homeless population.
I was inspired by community-maps I referenced in my tenant organizing days, such as The Citizens Housing and Planning Council (CHPCNYC Making Neighborhoods map and The Association for Neighborhood and Housing Development ‘s Displacement Alert Map.
After reading the City Limits article which contained a bulleted list of addresses and types of the 9 proposed shelters, I knew I wanted to illuminate some reasons why the surrounding population would voice opposition.
I found community district homelessness footprint data prepared by City Limits accessed through the Department of City Planning on July 31, 2019. This footprint data contains five other verticals besides community district ID:
The number of individuals in the shelter system who come from the district; Number of individuals sheltered in the district; Percentage of total shelter population that comes from the district; Share of citywide total shelter population hosted in the district.
This data source was not geocoded, so I searched for a shapefile on community district outlines and found one provided by NYU, created from the Department of City Planning open data.
To create and evaluate the map I used Microsoft Excel and Carto, a free mapping software.
First, I needed to make sure the community district boundary shapefile was accurate and in-line with the community district homelessness footprint data. Unfortunately, the current community district shapefile on the Department of City Planning’s open data portal is, well, wrong. It draws non-existent community districts that are not represented by the footprint data. After a lot of digging, I found the NYU resource that held an outdated, yet more accurate, shapefile. Next, I cleaned up the community footprint data in Excel to match the string values (Bronx 1 vs 201). I also manually looked up the latitude and longitudes via Google Maps for each proposed shelter.
I imported the shapefile first and displayed only the 59 community districts outlined. Next, I used the analysis tool to join the footprint data by the cleaned community IDs. From there, I created a choropleth map weighted by the share of citywide total shelter population hosted in the district in 6 colors. I found this vertical to be the most interesting represented on the map. However, I included a pop-up to provide more supplementary information.
Once I was satisfied with the presentation of the homelessness footprint, I added in the last spreadsheet of the proposed shelters. Carto instantly recognized the latitude and longitude coordinates after uploading, so it was geocoded before adding to the map. I used a contrasting colorway to indicate the sites. I also added a clickable pop-up to describe the type of shelter.
Figure 1 reveals none of the nine contested sites are in districts with currently no sheltered population. It also is interesting a lot of the points are on the borders of districts — perhaps a conscious political decision to not prioritize one over the other. The homeless come from every single district, the lowest contribution being 29 people (Bayside / Little Neck, Queens) and the highest being 2,631 (East New York / Starrett City, Brooklyn). It is apparent in the distribution that there is an unequal proportion of homelessness, not borough-wide (besides Staten Island).
Most of the contested shelter locations are planned for single men, which is not surprising. Most NIMBY pursuits are to keep men, usually of color, out of their neighborhood for “safety.”
Interpretation of Figure 1 led to the construction of Figure 2. I decided to create a second map that instead of showing each community district’s proportion of total sheltered population, shows a percentage of the the amount of people the community district shelters vs. contributes to the shelter system. After conducting a quick ratio column in excel, I was able to easily expose it on the already map structure.
The in-out interpretation also raises questions of fairness, as most of the lowest ratios have the most unequal burden of sheltering, meaning they shelter more individuals than individuals from that district need sheltering.
I would love to incorporate more demographic data to address a lot of the racist NIMBY discourse. I also think an examination of the non-contested sites (both already built and in the process) is necessary.
I’m also skeptical of the branding “contested.” City Limits described the sites as being contested for numerous reasons, “public safety fears, concerns about [the] strain on local services, worries about the safety or suitability of the proposed site for shelter residents, or fears about the impact a shelter might have on other plans for the neighborhood.” However, there is no real measure of contestation, and I believe to contest, you have the information, time, and energy to do so, which is usually not the most vulnerable populations.
Furthermore, I would love to incorporate data on the amount of homeless in varying subway stations, as the subway has become a de-facto shelter in this crisis. Especially since Cuomo and the City are displacing homeless out of the subway after framing them as scapegoats responsible for the disastrous MTA.