Helping communities mitigate financial shocks
With inflation now running at a 40-year high (+7.9%) and gas prices up over 50% since this time last year, millions of American families are experiencing significant financial setbacks. Economic shocks have wide-ranging effects, including an increased need to access credit, unstable housing conditions, difficulty in pursuing one’s education, and more.
In response and with support from the Mastercard Center for Inclusive Growth, DataKind has been hard at work on developing our Economic Resilience portfolio, which undertakes highly concentrated applications of machine learning and AI to unlock powerful tools and interventions that can give affected populations a better chance to withstand, recover from, and prevent these shocks. Our Economic Resilience portfolio helps individuals, communities, and policymakers move away from reaction-mode and gives them a better chance to foresee, adapt to, and leverage changing financial conditions to their advantage.
We’re proud to share two distinct bodies of work within our Economic Resilience portfolio.
The first is a US-focused cohort of projects supporting the economic resilience of individuals and families across the United States. Under this body of work we executed a housing insecurity initiative in close partnership with New America’s Future of Land and Housing Program with the support of The Rockefeller Foundation. DataKind built an open source tool that gives local policymakers and relief organizations the ability to make data-driven decisions when directing millions of dollars in federal CARES Act funding to homes and families across the state before foreclosures and evictions happen.
We also recently launched seven projects with support from Google.org (read more below). The goal of these projects is to increase access to economic opportunities across the US and provide the scaffolding necessary to sustain them. These projects are focused on key economic bell-weathers, including financial inclusion for traditionally underrepresented entrepreneurs, post-secondary education, safe and affordable housing, and increasing access to services that promote economic stability and poverty alleviation.
The second body of work, with support from the Conrad N. Hilton Foundation, has an international lens and aims to execute two projects in close partnership with the Global Opportunity Youth Network. Through our collaboration, DataKind aims to generate insights on workforce demand in Mombasa, Kenya to better connect jobs to jobseekers and define hyper-local community zones in Mombasa to better connect young people to essential services.
One powerful example of this work is the prediction of housing loss risk at the neighborhood level.
Every year, an estimated 5 million people in America lose their homes to eviction and mortgage foreclosures. This number is exploding since the COVID-19 pandemic. According to the Urban Institute, one in five US residents are now in rent arrears, and over 10 million are behind with mortgage payments.
These destabilizing events can contribute to job loss, homelessness, adverse health impacts, and downward economic mobility. Yet, we know little about where these events occur and who is most impacted, in large part because cities and counties lack accessible, timely, and quality data on evictions and mortgage foreclosures. Indeed, a 2021 National League of Cities survey shows that 38 percent of rural leaders and 22 percent of city leaders didn’t even know whether evictions had increased or decreased in the previous year.
As part of the package of funding distributed in the CARES Act, Federal emergency rent subsidies were made available in 2021 to populations at risk of losing their homes across America, but for a number of reasons, many eligible citizens and families were unaware of the availability or how to access these funds.
Local relief organizations are, then, left having to canvas every city district door-to-door to identify and inform eligible households of their available benefits.
What if these same organizations could be handed a data-assisted map that shows the neighborhoods where evictions and foreclosure have historically been highest for highly targeted outreach?
Over the course of six months, the DataKind team of data science volunteers built the open source Foreclosure and Eviction Analysis Tool (FEAT) that helps local leaders understand where housing loss is most acute, when during the year housing loss is occurring, and who is most impacted. FEAT provides local policymakers, community-based organizations, and relief organizations with the ability to make data-driven decisions, including, for example, directing millions of dollars in federal CARES Act funding to at-risk homes across the states before foreclosures and evictions can happen. FEAT is now an open source global good available to any housing organization or municipality that needs help directing funds.
To scale our Economic Resilience portfolio work to the next level, DataKind launched a cohort of seven projects.
With support from Google.org, we’re identifying where data science and AI can accelerate each organization’s impact on individual, family, and community economic empowerment.
DataKind and Google.org selected partners across the US in four key thematic areas within economic empowerment: (1) financial inclusion to increase access to credit for traditionally marginalized entrepreneurs; (2) post-secondary education to empower colleges and universities with the data-driven insights to support students’ academic journey to graduation; (3) safe and affordable housing to ensure individuals and families have the foundation upon which to build their resilience; and (4) increasing access to key services that promote economic stability and poverty alleviation.
We launched the execution of these projects in Spring 2022 with a determination to deliver high quality, ethical, and impactful products within six to nine months; tools that would be sustainably integrated into teams and processes to positively impact each partner’s mission. Examples of our work include supporting a student’s pathway to graduating with their bachelor’s degree within six years of entering a university by developing a predictive model to assess individual student risk, and providing local organizations with data-driven insights to pre-empt increased need in their communities and coordinate a targeted response by predicting demand for key services (e.g., health, rental support, food).
Through the execution of this work with our incredibly talented volunteer data scientists, we’ve begun to identify key findings that’ll not only contribute to increased impact for our partners, but also support learnings across the greater ecosystem. For example, our work in post-secondary education has already provided us with insights that students who didn’t decide on a major in their first semester upon transferring to a four-year university were at higher risk of not completing their bachelor’s degree. These students would therefore benefit from individual engagement to ensure support in achieving their academic goals.
As we continue to execute on the work and drive towards sharing the final products with our partners, we look forward to leveraging further data science insights to identify where, across each thematic area, replication and scale with new partners and across new regions could increase data-driven economic empowerment.
Image above courtesy of John Jay College / Arpi Pap.
Join the DataKind Movement
DataKind’s work is possible through the generosity of our volunteers, donors, and corporate sponsors. If you’re moved and can join us in supporting this and our future important work, please become a donor here. Every gift helps DataKind harness the power of data science and AI in the service of humanity.