July 06, 2020
By Emily F. Keller
A Neighborhood-Based Approach to Understanding Housing Insecurity and Eviction Patterns in the Wake of COVID-19

(Data visualization from the Baltimore Eviction Map)

Interdisciplinary research about neighborhood dynamics in the Cascadia region is growing in both scope and geography to address new risks of housing insecurity across the United States in relation to COVID-19. Research led by Dr. Timothy A. Thomas, a former UW researcher and co-lead of the Cascadia Urban Analytics Cooperative (CUAC) Neighborhood Change Project, is expanding to urban areas in California, Maryland and around the country to address the dynamic circumstances resulting from the pandemic and its impact on neighborhoods.

Thomas is the newly appointed Research Director for the University of California (UC) Berkeley’s Urban Displacement Project (UDP), and lead researcher for The Evictions Study, an ongoing collaboration between UW and UC Berkeley. Updates from both entities are detailed below.

== Collaborative Grant Addresses Housing Insecurity in Relation to COVID-19 ==

In April 2020, the UDP received a grant from the C3.ai Digital Transformation Institute (C3.ai DTI) to investigate and address housing insecurity and displacement risks related to COVID-19, in collaboration with the Eviction Lab at Princeton University and the Data-Intensive Development Lab (DIDL) at UC Berkeley. The project, “Housing Precarity, Eviction, and Inequality in the Wake of COVID-19”, is led by Karen Chapple, Professor and Chair of the Department of City and Regional Planning at UC Berkeley; Joshua Blumenstock, Assistant Professor at the UC Berkeley School of Information and Director of the DIDL; and Matthew Desmond, Maurice P. During Professor of Sociology at Princeton University.

The housing precarity project will utilize cloud computing and machine learning to predict and identify neighborhoods with a high risk of evictions due to circumstances surrounding the COVID-19 pandemic, such as employment and displacement factors. A team of interdisciplinary researchers will develop both a system to track evictions in real-time and a machine learning model to analyze and predict housing insecurity risk by area. These new data sources will support researchers and local policy-makers in making data-driven interventions in their communities.

The project was one of three selected for the first round of awards totaling $1 million from the newly created C3.ai DTI, a group of research universities and companies dedicated to accelerating the benefits of artificial intelligence for business, government and society. As of July, the awards have grown to $5.4 million, providing funding for a total of 26 projects that address COVID-19 across various disciplines such as medicine, urban planning, public policy and computer science.

== Researchers Work with Advocates to Publish the ‘Baltimore Eviction Map’ ==

In May 2020, researchers for The Evictions Study expanded their work from Washington State to the east coast, releasing the Baltimore Eviction Map in collaboration with local attorneys and advocates. Thomas worked with fellow researchers Ian Kennedy, Alex Ramiller, Ott Toomet and Jose Hernandez.

The study provides data visualizations showing the estimated risk of scheduled evictions and eviction removals by neighborhood, along with demographic, income and rental data for context. Results are based on a combination of U.S. Census data and a sample of 9,349 scheduled evictions between July 2018 and June 2019, provided by the Baltimore City Sheriff’s Department, in which 66% of cases led to a physical removal. Determinations of tenant race are based on a Bayesian prediction model developed by Kosuke Imai and Kabir Khanna using surname and geolocation.

Data analysis for the Baltimore Eviction Map generated the following preliminary results:

  • Three female-headed households were evicted for every two male-headed households.
  • Three times the number of Black tenants were evicted, compared with white tenants.
  • Evictions of Black female-headed households far outnumbered evictions of white male-headed households (2,996 vs. 775).
  • Evictions removed 7.3% of all Black male-headed households and 5.4% of all Black female-headed households from their homes.

The study’s neighborhood maps and findings on racial disparity in evictions were recently highlighted by a group of housing activists who are pushing the city of Baltimore to offer free lawyers to tenants facing eviction, which they say will save tens of millions of dollars in costs related to transitional housing and government services, according to the Baltimore Sun.