By Mallory Sheff, Portfolio Manager, Economic Resilience Impact Practice, DataKind
From facing an unexpected bill to losing your job, to being repeatedly refused loans to start your own business, economic shocks can come in many forms and have wide-ranging effects on the independence, opportunity, and life outcomes of individuals and communities. The ability to withstand, or recover from, these shocks and access opportunities is known as economic resilience. Beyond resisting and recovering from economic challenges, resilience also ensures individuals and communities can foresee, adapt to, and leverage changing conditions to their advantage.
To better understand the key factors that bolster economic resilience, DataKind launched in 2019 a deep-dive initiative to find where data science and AI could amplify this work. When the COVID-19 pandemic grew in 2020 and left waves of economic disruption in its wake, strengthening economic resilience became increasingly urgent. Exploring how DataKind could best make an impact in this space required us to identify gaps in economic resilience by creating deep partnerships with nonprofits, social impact organizations, and subject matter experts, learning from their insight, and applying the six components of a successful data science for good project across an entire system.
DataKind kicked off this work using our Impact Practice model through which we identify common pain points across a sector, ideate on how data science can be a catalytic tool to overcome these obstacles, and then pilot, replicate, and scale solutions. Through each phase, we employ design principles to ensure adaptability and sustainability of our work and remain committed to exploring challenges with a diverse, equitable, and inclusive lens.
In launching the Economic Resilience Impact Practice, we first engaged with practitioners and subject matter experts working at the frontlines with communities most impacted by the current economic crisis. This enabled us to better understand the areas most relevant to economic resilience, the various challenges in this space, and how data science and AI could be a tool to overcome the most salient obstacles.
Over the course of 12 months, DataKind engaged with over 35 organizations spanning 14 countries through Discovery Days, one-on-one conversations, and systems mapping to better understand the economic resilience landscape and the pain points in this system that could be unblocked using data science, machine learning, or AI. By iterating with partners and subject matter experts, we dove deeper into the recurring challenges organizations were experiencing. This allowed us to refine our thematic concentration such that the solutions explored and executed on were relevant to a larger ecosystem rather than to a standalone organization.
Below, we share strategic principles as we continue the exploration and development of our Economic Resilience Impact Practice. This provides us with guidelines as to how we can leverage both our expertise in data science and our partnerships with key social actors to impact the lives of those who can most benefit from our support.
1. Supporting economic resilience is most achievable through a combination of key determinants that ladder up to an individual’s, family’s, or community’s ability to thrive, especially in the face of economic shocks
DataKind’s landscape analysis reinforced the notion that there are various pathways to achieving economic resilience, stability, and empowerment. Financial inclusion, post-secondary education, and safe and affordable housing were the components that were most recurrent throughout our conversations.
While we’ve started by exploring these thematic areas, we understand that the challenges in economic resilience are intricate, interrelated, and intractable. DataKind looks forward to updating our insights and knowledge as we continue to learn about and engage more deeply in this field.
2. DataKind’s Impact Practice model allows us to surface solutions with the greatest potential to scale
DataKind’s Impact Practice model is central in ensuring that the work we do within each thematic area has an impact beyond single, bespoke partnerships. This long-term investment provides the opportunity for sustainable change, and requires engagement across strategic phases of work.
For example, DataKind partnered with New America’s Future of Land and Housing (FLH) program to generate census-tract level insights on evictions and mortgage foreclosures in the U.S. to identify who was most at risk of losing their homes and where. Our initial engagement to test the feasibility of applying data science for impact in this space, resulted in the publication of our Displaced in America report, which was directly utilized by local policymakers to identify where CARES funding should be distributed for highest impact. The work was then replicated in 5 additional counties and published in our Displaced in the Sun Belt report, thus providing us with the opportunity to refine and strengthen the work done in the first phase.
Together with New America’s FLH team, DataKind is now consolidating insights generated from the first two phases of work to scale the work nationally by developing a toolkit such that local leaders can leverage and analyze their eviction and mortgage foreclosure data themselves to drive decision-making and action in support of local communities.
3. While data science provides opportunities to bolster economic resilience, data maturity and homogeneity are key factors in ensuring the sustainability and scale of successful solution
During our discovery, we confirmed that organizations working in these topical areas have invested in their data infrastructure (from data collection to data utilization) and therefore have data to leverage for data science, machine learning, or AI projects. Many organizations have also invested in human resources, such as data science practitioners and leadership, to bolster their capacity to implement and sustain these solutions, further develop a data culture, and engage with tools to strengthen their data journey.
Data homogeneity was also a key factor in understanding our capacity to have an impact at scale. For example, microfinance institutions and Community Development Financial Institutions (CDFIs), are organizations that support the financial inclusion of traditionally marginalized communities. Given their work in the financial sector, they collect similar client and operational data such as credit scores, client bank balance, amount of loan provided, loan period, collateral, etc. A tool or algorithm leveraging this data to predict client loan repayment could be designed and tested with a small group of organizations with the aim to then generalize its utility across the sector.
DataKind is launching projects with diverse partners in support of economic resilience and empowerment, even as new learnings, insights, and approaches are continuously being identified and explored. Our commitment to ongoing discovery provides us with not only an increasingly comprehensive landscape of the sector, but also the possibility to uncover new opportunities and partnerships where data science, machine learning, and AI could have an impact in support of individual and community economic resilience.
Mallory Sheff leads DataKind’s portfolio of work to support the economic resilience and wellbeing of individuals and communities around the world. She guides social impact organizations and pro bono experts to successfully design and execute projects that, taken together, ladder up to scalable solutions.
Header image above courtesy of iStock/Bill Oxford.
As always, thank you for your support of this critical work!