Investigating homelessness in NYC via 311 complaints

Research focus

From October 2019 to September 2024, I want to understand the change in the rate of homelessness in New York City across different boroughs via 311 complaints. Across years, what months and times of day report the most homeless-related complaints? Where are the most homeless-related complaints created? What is the distribution of location types for homeless-related complaints by borough (subways, streets, buildings, parks)? More specifically, I am interested in which agency responds most to homeless-related complaints: New York Police Department (NYPD) or Department of Homeless Services (DHS)? Are there certain boroughs where homeless-related 311 complaints experience more risk of exposure to the NYPD? Over time, has NYPD responsibility for homelessness complaints risen? Is there a difference in the time it takes to resolve homeless-related complaints by the NYPD vs. the DHS? What does “resolving” the complaint mean – is there action taken or not?

The research focus of this project may be of particular relevance to local community organizations within NYC who are focused on homelessness and policing, such as the Coalition for the Homeless. Additionally, this research could be helpful to the NYC Department of Homeless Services in helping to determine the need for more effective handling of homeless-related complaints without the NYPD.

Dataset and variables

The visualizations in this project were created using the 311 Complaint Dataset from the New York City Open Data Portal. This dataset is maintained by the New York City Office of Technology and Innovation, and is updated daily. Every row in the dataset represents a 311 service request (or complaint) submitted by phone, online, or mobile.

For this project’s research focus, I specifically focus on filtering by complaints created from October 2019 to September 2023, that contain the phrasing “homeless” or “encampment” in their complaint type. The reason for the inclusion of “encampment” is explored further in this project. Each row in the filtered dataset that is featured in this research is referred to as a homeless-related complaint.

In order to address my research questions, I will be exploring the unique key, created date, closed date, agency, complaint type, location type, incident zipcode, borough, and resolution description columns. I created the following calculated fields and groups to further my analysis:

fielddescription
resolution time (minutes)a number representing the number of minutes it took to resolve the complaint, measured by the difference of created date and closed date
time of complaint (hour)a number representing the hour in which the complaint was created, parsed from created date
complaint phrasinga string identifying the central phrasing used in the complaint type, either “homeless” or “encampment”
time of complaint (group)a grouping for the time of day for a complaint, based on time of complaint (hour), which can be either:
– Morning (5am – 12pm)
– Afternoon (12pm – 5pm)
– Evening (5pm – 10pm)
– Night (10pm – 5am)
location type (group)a grouping for location type, which can be either:
– Building
– Subway
– Park/Playground
– Bridge, Road, Sidewalk
location type (indoor or outdoor)another grouping for location type, which can be either:
– Indoor
– Outdoor
– Subway
resolution description (group)a grouping for resolution description, which can be either:
– NYPD involved, action taken
– NYPD involved, action attempted
– NYPD involved, no action taken
– NYPD delegated
– DHS involved, action taken
– DHS involved, no action taken
action taken (group)another grouping for resolution description, which can be either:
– Action taken
– No action taken

Visualizations

Over the time period covered by this research, the time series graph above shows that number of homeless-related complaints have generally increased, with the most complaints located in the borough of Manhattan and the least in Staten Island. For each of the boroughs, a seasonal trend can be observed, in which it appears that the number of homeless-related complaints peaks in September or October.

Time of complaints

Across 2020 to 2023, the month of September has the highest number of homeless-related complaints created. The bar chart above ranks each month from most to least complaints, and uses a gradient of the color red to represent the density of complaints.
Across 2020 to 2023, mornings are when the highest number of homeless-related complaints are created, especially in the month of September. The highlight table above uses a gradient of the color blue to represent the density of complaints.

Location of complaints

Focusing specifically on the borough of Manhattan, where the greatest number of of homeless-related complaints are created, the density map above reveals that most complaints are concentrated below 59th Street.
The stacked bar chart above depicts that a large portion of homeless-related complaints created are located at a bridge, road, or sidewalk.
Grouped differently, the majority of homeless-related complaints are located outdoors. For the purposes of this pie chart, subways are a separate category, since subways could be either outdoor or indoor.

DHS versus NYPD

The stacked bar chart above analyzes the breakdown of responsibility for homeless-related complaints per borough. As stated in the annotation on the chart: The majority of all homeless-related complaints in which NYPD is the responding agency are phrased as “Encampment” – which is not the case for DHS, where the word “Homeless” is used in the phrasing for the type of complaint.
In the first iteration of querying data for this research, I noticed that NYPD seemingly stopped being a responding agency to “homeless” related 311 complaints in March of 2021. However, upon further investigation I came across NYPD complaint types labeled as “encampment” instead of “homeless encampment”. I suspect that in dropping the word “homeless” from the complaint type, there is an attempt at obscuring the data that focuses on NYPD and homelessness. In fact, the number of homeless-related complaints under the phrasing of “Encampment” peaked significantly higher than it did when it was phrased as “Homeless” – by almost 2,000 complaints!

Despite a 2020 directive that was meant to lessen NYPD response to homelessness, in response to police brutality concerns amid protests in the wake of George Floyd’s murder in 2020, NYPD responds to homeless-related complaints now more than ever. In the years that have followed the directive, we have seen the rise (and demise) of former NYPD officer and current mayor Eric Adams, who instituted a “de facto sweeps policy” that replaced the previously mentioned directive.

I wanted to further investigate NYPD’s response time to homeless-related complaints in comparison to DHS. Interestingly, the time series above depicting the average resolution time (in minutes) shows that DHS takes significantly longer to resolve complaints. However, I found it unreasonable that the resolution time for NYPD was almost always less than 5 minutes…
Upon further investigation, I looked into the resolution descriptions for homeless-related complaints and grouped based on whether or not action was taken, and by what agency. In comparison to DHS, which resolved the majority of their complaints by taking action, NYPD overwhelmingly resolved complaints without taking any action, which helps to explain their “quick” resolution time.

Questions to think about…

What stake does NYPD have in obscuring their response to homeless-related complaints? Why is NYPD so eager to resolve complaints, despite taking no action? How else can we investigate NYPD’s response to homeless-related complaints? How may the number homeless-related complaints not accurately reflect the prevalence of homeless throughout the city?