Seizing the Moment to Advance Data for Social Impact

The coronavirus pandemic has amplified the power of data science for social impact in many ways, from helping to deliver school lunch stipends to parents after schools closed to using technology to make accessing critical safety net programs more streamlined and user-friendly. As demand for safety net programs spiked, nonprofit tech organizations stepped up to not only improve access, but also to analyze user feedback to identify bottlenecks, speed documentation processes and help applicants find other services for which they’re eligible.

A recent webinar hosted by data.org and the Mastercard Center for Inclusive Growth talked with leading data scientists working for social change about how the pandemic has changed their job and the ethical considerations, data-sharing insights and partnerships needed to ensure data can lead to real change. 

Data helps communities respond to COVID-19

Delivering food benefits to out-of-school kids: As schools closed, families who relied on school lunches struggled to feed their children. In response, Congress sought to substitute cash stipends for the subsidized school lunches that nearly 30 million children get each day. Sounds straightforward, but implementation was far from it. “At heart, it is a giant data science problem,” said Tracey Patterson, a senior director at Code for America, which is part of the U.S. Digital Response team providing data experts to governments during the crisis. Code for America has built a digital application for parents that sends a preloaded electronic benefit transfer (EBT) card right to their door. 

The problem: Current data management systems aren’t robust enough to keep track of the changing lives of families, particularly in a pandemic when they might be moving in with relatives or sending children to live with other family members while parents work. Even before the pandemic, a surprisingly large number of kids didn’t have a current address on file at school. Code for America and the Digital Response team are looking ahead to the start of the next school year as a critical data collection moment to avoid this problem in the future.

“We should expect that school won’t look the same this fall,” said Patterson, “so how can we look at the start of school year as critical data collection moment…so we can understand how to continue to serve” these families during a time of crisis.

Getting high-risk homeless individuals out of shelters: During the pandemic, Community Solutions, which helps communities use real-time data to eliminate homelessness, aided organizations in identifying homeless individuals at high risk for COVID-19, such as older individuals or those with underlying conditions, said Roseanne Haggerty, the organization’s president and CEO. That information helped public health agencies move those individuals out of high-risk homeless shelters and into safer quarters.

The power of data helped organize a collective, effective response, Haggerty said. Homelessness has always been life or death for individuals, she said, but COVID-19 has made people realize it’s a public health crisis as well.

The data also allowed communities to identify the racial disparities in COVID-19 cases, and Community Solutions is now turning its attention to disaggregating data by race so communities can continue to better understand the pandemic’s impact on communities of color.

Building capacity to use data for good
As the coronavirus swept the globe, data scientists went into overdrive to help communities, said Ginger Zielinskie, interim executive director of data.org. They are “meeting this moment.”

The webinar participants discussed three action-oriented items that communities, governments and nongovernmental organizations (NGOs) should pay attention to when using data to transform systems.

Responsible data-sharing is critical to solving complex problems: COVID-19 has underscored the importance of a coordinated response across systems, and that requires sharing data. Without shared data, said Haggerty, “you can’t make progress on any complex issues where multiple organizations need to collaborate.” Sharing data across and between organizations allows stronger collaboration, rapid-cycle testing of ideas and interventions and improves the ability to predict and recalibrate situations for better responses. While privacy issues can often stymie data-sharing, they are “totally solvable,” said Haggerty, citing their use of HIPPA compliance releases. Organizations can advance responsible data-sharing by grounding efforts in a core set of principles that guide the ethical collection, management and use of data.

Transparency and multiple insights are key to ground-truthing data: Data is most effective when it’s built on good information—and that means multiple voices and community insights must be at the table from the beginning.

Being transparent and developing people’s comfort level with data ensure they can contribute their expertise into the data science conversations, said Trooper Sanders, CEO of Benefits Data Trust, which helps families access public benefits. More broadly, a multidisciplinary approach injects a sophistication into how data is used. “Some of my favorite partners are lawyers and anthropologists,” he said. Varied voices ensure organizations are not data-driven, but data-informed “because data exists in a broader ecosystem.”

“Collaboration also helps us uncover the truth that the data tells,” said Patterson. In Louisiana, the state SNAP offices realized that the statistics weren’t telling the full story of that food security program. They hired SNAP clients to be data advisers to help them understand the data they were seeing, Patterson said. Collaboration like that, she said, “helps us grow smarter and more understanding and recognizing how we can be more transformative with data.” 

Ethical considerations should be front and center: How the data will be used, who it impacts and how and other ethical considerations should always be part of every conversation. Data scientists should always ask first, “what harm can be done if we fail, and if we succeed ,” said Afua Bruce, chief program officer at DataKind, a global nonprofit that uses data to help organizations build capacity for social good.

Data should never mask the humanity underneath, said Patterson. All involved should remember that the data point is a person, she said. The data tells a story about people’s lives, often in crisis. Understanding the qualitative story is equally important, she said. “If we mask the data points by not understanding the humanity underneath them, then it will tell a different story.”

In the end, the COVID-19 crisis has raised the question of how the country can begin to make data an essential common good for solving the complex problems we must grapple with, said Haggerty. “How do we make this a mainstream conversation?”