Visualizing Homelessness in WNY

The problem of homelessness is something all cities must work to solve but a critical question to ask is "Who are the homeless? Who is the data representing?" This project is an effort to understand and visualize these questions with a focus on homelessness in the Buffalo/Erie County region. The data used was provided by the Homeless Association of Western NY for the year 2015 and only accounts for those individuals who have checked into a homeless shelter. As such, this data is only half of the real story and does not account for those who are temporarily lodging with family or friends, which is a growing trend in the Buffalo/Erie county region.

A quick glance over the data showed us there were approximately 5,000 instances of homelessness for the year 2015. Of these 5,000, the predominant household type was "single individual" and the average length of stay in a shelter was 16.7 days.

In order to map the data, we geocoded the zip code of the individual's last permanent address, which is defined as the place they last paid rent. Keep in mind that while every point within a zip code represents an individual whose last address was in that zip code, the precise location of the point is arbitrary. The dispersion of points was simply an aesthetic rather than a precise and accurate depiction of the exact coordinates of an individual's last address. Due to the nature of the data and individuals involved, precise locations were not provided in order to keep the privacy of individuals. The following three maps visualize where the homeless are coming from specifically within the borders of Erie County and then New York State.

Flow from within Erie County

Percent Homeless at some point in 2015 in Erie County (interactive)

Flow from within New York State

Looking at the maps of individuals from within Erie County, we can see a concentration of homeless coming from the downtown Buffalo region. Additionally, we can zoom into New York City on the New York State map and see that many homeless from NYC are coming from Brooklyn. The next map, which depicts mean income per zip code, adds another layer to the story.

Mean Income per Zip Code in NYS

Comparing the Erie County maps and the New York State flow maps with the mean income per zip code, you can see an overlap between income and where the homeless are coming from within the state. The zip codes from Buffalo not only represent a higher percent of homeless but they also have lower income compared to the rest of Erie County. You can zoom into New York City and see the same thing happening with regards to where the homeless are coming from. This isn't really new information but the interesting part of this data occurs when you look at it from a nation-wide perspective, because New York state isn't the only place WNY homeless are coming from.

Are Homeless Being Bused to Buffalo?

Imagine you are out on the streets with nowhere to go and very few possessions. At some point, you receive word of a new program your city is implementing. If you have relatives outside of the city and want to go live with them, the city will provide you with a bus ticket out. It's a win-win situation! You have an opportunity to be closer to relatives and help, and the city gets to reduce their rates of homelessness. But, what if that wasn't the whole story?

Nationwide Flows of Individuals to Erie County

When looking at this map, it's apparent that while New York State is where many of the homeless are coming from, there are a couple of areas that stand out, particularly areas in Florida and Puerto Rico. Upon further research, we found several news articles detailing the policies these cities have towards homelessness.

Cities that Bus (click on the points for news articles!)

The above map shows several points that have some sort of busing program for the homeless. Each of these points is a location listed as the "zip code of last permanent address" from our data. Though many of the busing programs say they are trying to send homeless to relatives, this isn't always the case. There have been several instances, especially within Florida and Puerto Rico, where the relocation of the homeless wasn't entirely for the good of the person involved. To get a better idea of how we could represent this, we also mapped the length of stay in a shelter for each individual.

Avegrage Length of Stay

We haven't performed any statistical analysis of our data but by simply looking at the map, we can see that certain areas, such as New York City, cities in Florida, and Puerto Rico, appear to have longer stays in the shelter. This could be due to a homeless person being sent or "bused" without actually having a relative in the city. This could be especially true of Puerto Rico, which doesn't have a busing program but a very similar one that involves flying homeless and drug addicts to cities in the U.S. under the guise of receiving rehabilitation.

Future Research

Much of what we've unconvered, especially with regards to Puerto Rico, has shed light on the situation of homelessness in Buffalo. Before we fixed our data, all of the zip codes listed as three digits had been thrown away or discarded. This meant that several people and locations from the data were simply deleted. By reformatting the zip codes to show as five digits, we've managed to bring attention to a key component of the homeless in Western New York. The homeless are already a margionalized group, and by throwing away certain data, an even smaller (but still significant!) group was being marginalized further. By better understanding who the homeless of WNY are, where they are coming from, and how they are getting to Buffalo, people can be better equipped for making decisions and policies on how to reduce homelessness in the region.

For our future research, we're hoping to gather similar data from other cities in WNY to see if a more evident and concrete pattern of busing occurres. In addition, we're looking to also focus in on Puerto Rico and see if Buffalo has become another location for the country's "addiction" program.