Cincinnati begins using data analytics to predict and prevent evictions
Keeping people in their homes could come down to a math problem. If area homeless prevention agencies got a heads up on who was at risk and what types of bills they were having difficulty paying, they could be proactive and help prevent an eviction before there was even an eviction notice. Cincinnati and other cities are trying it. Maidstone, England, says it has done it the longest and has proof it’s working.
Cincinnati’s program launched in July
With a $2 million, 2-year grant from the City of Cincinnati, Strategies to End Homelessness and its area shelter partners, are beginning to gather lots of data. They are trying to find the most vulnerable, alert them before an eviction notice and offer financial assistance. To identify those at risk, the University of Cincinnati and a third-party agency will enter anonymized data into an algorithm.
Strategies to End Homelessness CEO Kevin Finn has identified forty sources of valuable information including utility shut off notices, and applications for SNAP benefits. “We already have about a half dozen that we’re using. And it’s everything from some historical data of people who previously experienced eviction, previously were homeless and had called the help line for housing resources.”
He’s hoping to help 160 families a year, admitting this is a mindset shift for nonprofits. “[Nonprofits] that are normally used to waiting for somebody to come knock on our door before we considered helping them. Now, we’re going to knock on their door,” he says.
Once a family is identified as being at risk, social workers will receive their name, contact information, and try to help.
Finn anticipates a lot more families needing help than money to go around. Others can go to TenantGuard.org and self-report, entering in their name and contact information. Eventually a chatbot, using artificial intelligence, will suggest sources of financial assistance. That part is not fully operational yet.
The data science company 84.51 got first place at a Hackathon for designing TenantGuard.
Homelessness is up 12-percent nationwide
Cities are having to get creative to prevent evictions and ultimately homelessness. A record number of people in the U.S. are unhoused. The Department of Housing and Urban Development says in 2023 there were more than 650,000 who lacked permanent shelter. (on one night in January) Experts say it’s due to rising rents, expiring COVID money and in some states, a strain on the shelter system from the migrant crisis.
- In 2023 in Greater Cincinnati there were 6,142 people who experienced homelessness. (despite a one year slight uptick that’s a 12-percent decrease since 2019 and the 18-percent who were unsheltered is less than half the national average). Kevin Finn credits proactive measures.
- In Montgomery County homelessness is increasing. In 2023 there were 5,006 people experiencing homelessness for an average of 69 days.
According to the Coalition on Homelessness and Housing in Ohio, for every 100 extremely low-income people trying to find housing, there are only 40 units available. Executive Director Amy Riegel says that creates a system where most are paying more than 50-percent of their income in rent.
She says, “Individuals who we see either perhaps living in a park or in a doorway, or even in a place that’s not fit for human habitation like a car or abandoned home, those numbers are on the rise. We’re seeing more families entering homelessness.”
Riegel is disturbed by hearing more families with young babies are living in a car.
Personal stories
Sandra Johnson and her son were homeless for a while. Brick by Brick interviewed her outside the Freestore Foodbank in Cincinnati. “We stayed with different people, and we were just blessed, and apartments started opening up and we’ve been doing pretty good ever since.”
Eventually her son was able to secure a rent deposit from a nonprofit, her daughter found an income-restricted place in the Cincinnati neighborhood of Pendleton. And Sandra got an apartment with the help of Talbert House.
She encourages people who are experiencing homelessness to keep their head up and reach out to people.
Fifty-one-year-old Kenny Scott lives in a tent under a bridge with his brother. The Cincinnati man says he’s unable to work because of an injury and wishes circumstances were different. “I would love to be on a lake somewhere fishing, you know what I mean? Enjoying my family.”
And these are just two stories of many. Brick by Brick’s Hernz Laguerre Jr. has others. He explains how one woman, facing eviction, could have benefitted from predictive data analytics.
In the best-selling book Evicted, author Matthew Desmond says, “Along with instability, eviction also causes loss. Families not only lose their home, school, and neighborhood but also their possessions: furniture, clothes, books. It takes a good amount of money and time to establish a home. Eviction can erase all that.”
Where Cincinnati got the idea to use predictive data analytics
Maidstone England is 32 miles southeast of London with a population of about 200,000. In 2018 it noticed homelessness had increased 60-percent in five years and reached out to technology partners EY (Ernst & Young) and Xantura with the idea of trying to predict who was at risk of homelessness and help them stay in permanent housing. The program is called OneView.
Head of Housing John Littlemore told Brick by Brick, “We were all a little skeptical. I was skeptical in the beginning that you can use data analytics to prevent homelessness, but I think what’s different about Maidstone is that we’ve been able to prove that yes you can.”
Here’s how it works:
- Data identifies who needs help at least eight weeks before an eviction notice.
- Social workers reach out to those in need.
- If they want help, Maidstone helps them stay housed.
Initially the borough had trouble getting access to data because of privacy concerns. Littlemore says, “Data is quite heavily regulated, quite rightly, so we had to convince people, not only people in our own organization, but also external organizations who we wanted to work with that a) what we were doing was compliant with all that regulation and b) it was the proper use of that information and we would be able to keep it safe.”
He says of the hundreds helped; nobody ever complained about the use of data to identify them.
Xantura reports OneView saved the city nearly $300,000 in administrative costs and about $3M in broader societal savings to the community in its first year, reducing homelessness by 40 percent.
Other Examples
Los Angeles County
CalMatters reports Los Angeles County has served more than 700 clients since 2021 using predictive data analytics and 86-percent are still housed.
The idea began in 2019 when more than 180,000 people in California were unhoused. UCLA’s California Policy Lab tried to see if it could use machine learning, combined with LA County data, to predict homelessness. It started “with a list of 90,000 people who recently used services from the county’s Health Services or Mental Health Services. Using 50 factors, the computer ranks those people from 1 to 90,000 based on their risk of becoming homeless.”
South Bend, Indiana
The City of Southbend, Indiana is also reducing homelessness with predictive data analytics. It worked with Purdue University researchers to build a model using code enforcement, utility bill delinquency, evictions and foreclosures. Some of that data it had to pay for.
Denise Riedl, Chief Innovation Officer for the City of South Bend, told Cities Today that the goal is to provide support for people sooner. “Instead of catching them after they get evicted, or after the house was foreclosed on, or after they experienced some level of housing instability, we can actually move earlier,” she said.
University of Dayton Professor Expanding His Efforts
As Cincinnati gets its program off the ground, there’s a smaller effort underway in Montgomery County. University of Dayton Professor of Electrical and Computer Engineering Raúl Ordóñez started a class called “Engineering Systems for the Common Good.” It produces models using data to solve social problems, like what Ordonez and his students do with robots. Right now, they are focused on homelessness.
Ordonez saw a lot of poverty in his native Ecuador and wants to help prevent it here. He realizes his students can make a difference, using Montgomery County data. By building a mathematical model of how objects and their relationships change over time, he can get a clearer picture of what changes could be made.
“So, what if we spend more money on, say more prevention or on permanent housing? So, what would happen then so that ideally then the county could take these models and then start to study policymaking in a more systematic way,” he asks?
Ordonez needs more data. During his upcoming sabbatical he hopes to gather additional information to make a more complex, larger scale model.
Limitations
There are limiting factors in using predictive data analytics to reduce evictions and ultimately homelessness.
Cost
Cincinnati’s Strategies to End Homelessness says the $2M grant from the City of Cincinnati will be used to help an estimated 160 families per year. Kevin Finn is still trying to figure out the time and cost for staff. There’s also the cost of storing the data in the cloud with a third party. And the grant is only for two years. Finn is fundraising so the program can continue.
Los Angeles County is using $26M in federal COVID funds to pay for its predictive analytics program. That money is expected to run out in 2026.
Some data for South Bend, Indiana costs money, namely eviction and foreclosure information.
Maidstone, England pays technology partners EY and Xantura.
Data
Not everybody is willing to hand over data to nonprofits. Kevin Finn is having trouble getting utility information. It’s one of forty data points he’d like to have. Cincinnati and other cities are keeping this information anonymous until somebody at risk is identified. Maidstone said it even had to convince its non-profit partners to turn over data.
Results
SmartCitiesDive.com reports a data scientist worked with a utility company in Spokane, Washington to study whether anonymized utility bill data could predict people’s financial status. “His model linked that utility data with information provided by Spokane on the use of shelters and assistance programs. It predicted with nearly 75-percent accuracy whose who would eventually experience homelessness.”
Keeping people off the streets
In Maidstone’s last assessment, the city said it was able to prevent 98-percent of the people it helped from becoming homeless. In Los Angeles County 86-percent of people served since 2021 have retained their housing.
Program pays for itself
John Littlemore says Maidstone is saving money by not having people out on the streets. In its first year it saved the city nearly $300,000 in administrative costs and almost $3M in broader societal savings to the community. Also, during its pilot year, over 650 alerts were generated. Of those who were identified as the highest risk, only 0.4-percent became homeless.
Even before Cincinnati began its predictive data analytics program it was saving money. Kevin Finn says helping people who are couch surfing find a permanent home before they’re out on the streets costs $1,600 compared to $4,700 if somebody is already unhoused.
The National Alliance to End Homelessness says the government spends on average $35,578 per year for every person who faces chronic homelessness (including crisis services: jails, hospitalizations and emergency services). The 2017 study says those costs can be cut in half when a person is placed in supportive housing.
Looking ahead
Maidstone, the market leader in using predictive data analytics to prevent homelessness, says other US cities have contacted them for more information. John Littlemore says he’s also spoken about it to the European Union.
Littlemore’s advice to communities doing this? He told Brick by Brick, “Be ambitious. Don’t be afraid to take on these challenges. There are very many clever people out there in the private sector that can help you overcome some of the challenges around that so work collaboratively. Be bold!”
CEO Peg Dierkers of Bethany House, a family shelter in Cincinnati, says, “We hope three years from now we have data where we can really say and market – if you’re in this predicament, please call this list of services and we will find you help We really want to stem the tide.”
Are you enjoying exploring solutions with the Brick by Brick podcast and team? Your feedback is very important to the effort, as we continue to dig deeper into the responses to some of our most pressing social issues, including the housing crisis. Click on the green button below to take a short survey to help reinforce and refine our work. Thank you.
Ann Thompson – Host, Producer
Over the last thirty years in Cincinnati, Ann Thompson has brought a wealth of knowledge and expertise to her reporting. She has reported and anchored for WVXU, WKRC, WCKY, WHIO-TV and Metro Networks and freelanced for NPR, CBS and ABC Radio. Her work has been recognized by the Associated Press and she has won awards from the Association of Women in Communications and the Alliance for Women in Media. She is a former News Director and Operations Manager. Ann has reported from India, Japan, South Korea, Germany and Belgium as part of fellowships. Ann thinks of the Brick by Brick project as “journalism for good.” She serves as host and producer. Ann lives in Anderson Township with her husband Scott. They have two boys. Jake graduated from the Air Force Academy in 2022 and Kurt attends West Point.
