Big data used to solve big problems

Harris researcher Rayid Ghani’s new fellowship encourages data scientists to use their knowledge for social good.

By Nathan Peereboom

Last March, Harris School for Public Policy researcher Rayid Ghani sent a message to a group of his friends from around the world: I’m starting a data science fellowship, and your students have two weeks to apply.

In two weeks, he received nearly 600 applications.

The Eric and Wendy Schmidt Data Science for Social Good Fellowship was designed to apply the practices of big data science to big social problems. The fellowship was conceived by Ghani and Eric Schmidt, Google’s executive chairman, after they met on the 2012 Obama campaign, where Ghani was chief data scientist. Now a researcher at the Computation Institute and the Harris School, and the co-founder of Edgeflip, an analytics start-up focused on helping social good organizations, Ghani is powerfully thrusting data science forward as a way to not just sell widgets, fatten hedge funds, and collect clicks, but as a tool for making the world a better place.

From the 600 applications, 36 fellows were selected for this summer’s fellowship. As the fellowship’s Web site brags, “that’s 6.545% of the applicant pool, for you bean counters.”

The fellows’ goal? Completing 12 projects that involved processing huge data piles for government agencies and nonprofit organizations, with the ultimate goal of making a difference.

The projects included predicting energy loan performance with the Environmental Defense Fund and creating an open-source analytics tool to help the Cook County Land Bank decide which distressed properties to acquire and redevelop as affordable housing.

One particular aim of the program is to incubate productive social relationships across different fields to encourage data scientists to pursue careers in the field of “social good.” Ghani feels that because most data scientists are friends with data scientists who are already embedded in the investment banking, tech, or otherwise corporate world, they are more likely to be employed there.

“People with my kind of backgrounds are really exposed to…a biased part of the world. We know a lot more people from Facebook and LinkedIn and Foursquare and Google and all those Internet companies than…from the nonprofit world or the people who are solving education problems or health care problems or energy problems,” Ghani said.

On many occasions, the social context of the fellowship overlapped with the data-mining skills of the fellows. For example, at one point in the program, a pair of students left their laptops in a taxi. While they didn’t remember the name of the driver or the number of the cab, one of the students had recognized the driver speaking in a Nepalese accent. Knowing that the Nepalese population in Chicago was small, the students then called a Nepalese restaurant and told them about their problem. The next day, the backpack was returned.

Though Ghani and Schmidt consider this year’s $1 million fellowship a success, neither of the two is satisfied. Next year, they’ll be looking for better projects. Ghani hopes that some of next summer’s projects will turn into their own companies or nonprofits. He also plans to expand the fellows program to 50 students for the summer of 2014 to find even more data scientists who are eager to solve social problems.

“[Big data] people care about solving these problems but just don’t know how to do it. And that was a reoccurring theme throughout the summer when we would talk to the students. They kept on saying…this is something I would have wanted to do but didn’t know existed,” Ghani said.