What can we learn from holding a competitive global funding challenge that elicited ideas from 108 countries? That was the focus of a recent webinar about the data.org Inclusive Growth and Recovery Challenge.
The Challenge originated from the vision of data.org’s founding partners, the Mastercard Center for Inclusive Growth and The Rockefeller Foundation. Both organizations saw an opportunity to harness the power of data science to tackle the world’s most pressing problems, and worked with data.org to craft a Challenge that would draw ideas and proposals from and for everywhere in the world.
The scale of the data.org Challenge was huge: more than 1,260 individuals and organizations applied for $10 million in funding and technical assistance, and 400 judges from 49 countries spent more than 3,000 hours in technical review to find the projects best aligned with the Challenge’s goals and resources. Technical partner DataKind led the review, working closely with the data.org team.
You can read more about the groundbreaking work of the inspiring round of eight Inclusive Growth and Recovery Challenge awardees by reading the report.
What lessons were learned? And how might those lessons stretch beyond the final awardees to everyone who applied?
The Challenge Encouraged New Partnerships
The Challenge focused on democratizing data and data science to make a positive social impact. With “Inclusive Growth” in the title, a majority of applicants submitted proposals addressing impact through an economic lens. But applicants also partnered with organizations in other areas like peace and justice and on building strong institutions. These partnerships were a positive surprise for Afua Bruce, Chief Program Officer at DataKind. She enjoyed seeing so many new and creative partnerships as she reviewed the applications. “It’s exciting to see that there are so many different ways that the social sector can come together,” she said.
Partnerships Make Better Products
Partnering as grantees to effect sustainable solutions with smallholder farmers in India, The Basel Agency for Sustainable Energy (BASE) & the Federal Laboratories for Materials Science and Technology (EMPA) built a project neither group could have delivered alone. BASE, which is focused on sustainable energy and climate change solutions, and EMPA, a Swiss materials science and research institute, are working with smallholder farmers in India to use data and deliver insights that provide farmers access to more solar-powered refrigeration. Their mobile application, called Your Virtual Cold Chain Assistant, will help farmers better understand their options when it comes to product refrigeration, like where and how to store and sell their produce to make the most profit. The organizations estimate that farmers who use the app will be able to increase their income by up to 30% and will discard 20% less fruit and vegetables. What’s more, moving food through solar-powered cooling systems will cut cooling-related greenhouse gas emissions in half.
“The partnership has been essential because of the complementary skill sets,” said Thomas Motmans, project lead at BASE. “In acknowledging our strengths and weaknesses, we were able to identify and address potential gaps or challenges and work together as one broad team.”
The Challenge Fostered Global Collaboration and Community
As a platform for partnerships, data.org serves as the connective tissue for these kinds of collaborations. But its goals are broader than making project-based connections, said the organization’s Chief Strategy Officer Ginger Zielinskie. data.org is committed to building a global community for shared learning, and this community was able to help applicants as they put together their proposals for the Challenge. There are six criteria for a successful application, including smart problem statements, funding, data sets, data scientists, subject matter experts, and of course, a keen grasp of the people who will use the solutions that are being proposed.
“Some of the applicants wanted support in funding. Some of them wanted to execute work they already envisioned. Some needed support in finding and cleaning data sets. Some needed to know, ‘how do I find a data scientist to help me execute on this?’” she said. “And that’s the role of this community, to help each other address those needs.”
The Challenge Helped Zero-In on the Biggest Problems
Afua Bruce said the most successful applications focused intently on defining the problem before they started to look at how to solve it. Questions to ask included: “How do they define the problem? What are the biggest problems? What are the wicked problems that keep them up at night? And then, what are the tools that need to be in place to solve those problems? What are the data sources that need to be in place to help support that problem? And then from there, how do you actually scope a project?”
To facilitate the application process, data.org provided training webinars to address these needs. “Scoping 101,” for example, helps applicants scope projects once they’ve identified their problems. It’s available along with the other training webinars, here.
“I watched some of them twice because they’re so good and well structured,” said Thijs Defraeye of EMPA, one of the Swiss organizations partnering on the produce refrigeration project. “And you learn a lot from them. It was a big help in scoping the project. The webinars gave us more steps, which let us achieve a higher quality level.”
Zeroing-in on Problems Encouraged Applicants to Listen to the People They Want to Serve
The teams working on the Your Virtual Cold Chain Assistant are interviewing 1,000 stakeholder farmers, one by one, while scoping their project. They asked the farmers what they needed so that they could design something they would actually use and benefit from.
“Doing the research avoids a situation in which we design a very fancy machine learning model that doesn’t actually end up saving food,” said Thomas Motmans. “It’s about consulting with the people who will actually use the product. In our case, asking farmers.”
This close collaboration with the communities served was a hallmark of many of the strongest applications in the Challenge.
The Challenge Encouraged Applicants to Root Out Bias
When organizations are intentional about identifying and mitigating bias in their data sets, their results are more accurate and effective. That’s why the Challenge encourages applicants to screen for bias and to utilize additional or alternative data sources when they do identify bias in their data.
Afua Bruce of DataKind explained how bias presents itself in data, and how organizations can address it. For a project on benefits administration, for example, one might look at eligibility and location. But, she said, “eligibility criteria in benefits data affects people of color, and zip codes also include bias. So should we be using benefits data or zip codes? Is there something else that we can correct for? Or add in? That is the question they should be asking.”
The Challenge Helped Tell the Story of Why Data Science for Social Impact Matters
The Challenge’s primary aim was identifying projects with the best chance of achieving meaningful progress in inclusive growth and recovery. A secondary result, made clear in the Challenge report, is that the broad and innovative applicants tell an effective story of why the opportunity for data science is in the social sector.
By surfacing these applications, the Challenge shows that “data science is about more than helping [people] buy the right turtleneck before they put down [their] phone at night,” said Ginger Zielinskie. The next challenge for partners of the Challenge: “How do we tell that story of what’s going on in ways that resonate and inspire similar efforts around the world?”
You can read more about breadth of the issues being tackled by Inclusive Growth and Recovery Challenge applicants and awardees by reading the report. It shows the range of opportunities that exist to use data science to drive social impact for workers, entrepreneurs, and communities.
Cities and Towns — Leave No Place Behind
Community Lattice will create a platform to predict the cost and risk of brownfield redevelopment projects in the United States to transform a community’s ability to secure redevelopment funding, improve community health, and create economic opportunities.
GiveDirectly and Center for Effective Global Action will develop and implement a new model for humanitarian and development aid that enables cash transfers to be deployed effectively, accurately, and at scale to those who need them most.
To address the digital divide, the University of Chicago will create open-source maps and toolkits that can be used to highlight inequities in broadband access and advocate for more equitable policies and investments.
Using a rich, longitudinal population dataset, Aalborg University will create interactive and actionable maps for policy-makers and urban planners that identify the geographical places in Denmark that are most vulnerable to out-migration and economic instability.
Access to Capital — Leave No Entrepreneur Behind
BASE will use machine learning and physics-based food modelling to enable smallholder farmers in India to access sustainable cooling facilities, thereby reducing food loss and dramatically improving livelihoods.
In Nigeria and Tanzania, Solar Sister will share market insights with its network of women entrepreneurs to grow their renewable energy businesses.
Addressing gender bias in lending algorithms, Women’s World Banking will work with financial service providers around the world to increase credit access for low-income female entrepreneurs.
Jobs of Tomorrow: Leave No Worker Behind
Using data mining and machine learning, Fundación Capital will arm informal workers in Mozambique with essential labor-market insights to increase income and employment opportunities.
The Challenge report shows that mission-driven individuals and organizations of all types, sizes, and maturity that seek to transform communities and realize lasting social change have recognized the potential of data science to achieve their goals. The Challenge clearly highlighted the high and varied nature of the demand for data-driven solutions in service of the social sector. As data.org we will take this learning, and use it as we develop programs, partnerships, and a platform to build the field of data science for social impact alongside this committed global community.