Interacting with HUD Homelessness Data in Tableau

Image for post
Image for post

One of my favorite quarantine projects has been self-teaching Tableau. A firm believer in learning by doing, I decided to create a full data science project using Tableau Prep for data cleaning and Tableau Desktop for visualization.

I’m always excited to scroll through Jeremy Singer-Vine’s “Data is Plural” newsletter when it lands in my inbox, and ended up in his archive of interesting datasets while looking for inspiration for this project. Annual homelessness assessment data from the U.S. Department of Housing and Urban Development (HUD) immediately jumped out to me as worthy of further exploration. I see homelessness as a particularly pressing issue given its troubling connection to my field of mental health — mental health disorders are far more common among individuals experiencing homelessness than among the general U.S. population. In 2015, HUD found that 25% of homeless individuals were seriously mentally ill (compared to 4.2% of the general population), and that 45% had any degree of mental illness. I decided to use my Tableau project as an opportunity to learn more about this important and highly relevant issue.

Developing the Project

The HUD dataset is impressively thorough, including aggregated count data representing homelessness by state and by year between 2007 and 2019. Variables that I selected for analysis included the total number of persons experiencing homelessness, counts of homeless individuals, counts of chronically homeless individuals, counts of homeless persons in families, counts of homeless family households, counts of homeless Veterans, and whether or not these persons were sheltered or unsheltered.

I was interested in how homelessness has changed during during the years for which data were available, which sub-groups are most and least likely to be sheltered, and how these patterns differ by year and by state.

The full, interactive dashboard that I created to answer these questions consists of an interactive map, line graph, and bar chart.

Image for post
Image for post

The interactive map shows states with higher rates of homelessness per capita in darker colors (calculated using population data from the U.S. Census), provides hover-activated, state-level counts associated with homelessness variables, and includes a filter to select data from a given year.

Image for post
Image for post

The line graph depicts annual counts of total persons experiencing homelessness between 2007 and 2019. When a state is selected in the interactive map, the line graph is filtered to only show data from that state.

Image for post
Image for post

The bar chart shows the percentage of persons experiencing homelessness who are sheltered, broken down into the following categories: overall, individuals, chronically homeless, people in families, and veterans. This chart is filtered when states are selected in the map and when a year is selected using the “Year” sliding filter.

Exploring the Finished Product

I made several interesting observations regarding homelessness and sheltering by clicking through different combinations of states and years. Of particular note was a drastic change in both the number of individuals experiencing homelessness in Louisiana and the percentage who were sheltered between 2008:

Image for post
Image for post

And 2009:

Image for post
Image for post

Overall homelessness more than doubled in this one-year span, and the percentage of sheltered individuals plummeted in all categories except for people in families. While more data would be needed to definitively explain this change, it seems likely that the particularly devastating 2008 hurricane season left many people in Louisiana without a residence. Shelters and related community resources likely did not have the capacity to respond to such a drastic change in homelessness, resulting in these low percentages of sheltering. Fortunately, Louisiana seems to be recovering and 2019 homelessness levels were below those recorded in 2008.

View and interact with the complete dashboard here.

I am currently a data analyst working in psychiatric epidemiology, and I am excited about the intersection of data science and mental health. Views are my own.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store