How Georgia Tech students get community-mindedDavid Terraso and Marc Schultz
The culture at technological institutes—especially those, like Georgia Tech, regularly ranking among the top engineering schools nation- and worldwide—often tend to drive students toward maximum financial reward: the world-class technology company, the high-stakes startup. That’s why Georgia Tech’s Data Science for Social Good program is so extraordinary: Having just completed its third iteration, the summer-long learning experience works like an internship, putting students to work for one of three organizations serving communities, different each year, including nonprofits like Trees Atlanta and government agencies like the Atlanta Regional Commission. Given a pressing organizational challenge, students develop new tools to overcome them, based on information they’ve gathered, analyzed, and prepped for use.
“Students get to see how to apply data science skills to real problems which have real social impact,” said Assistant Professor Bistra Dilkina, who has served as program co-director for two summers. “It also gives them a taste of alternative career paths—showing them they can use their skills for venues other than Facebook or Google.”
One of the challenges tackled by the 2016 cohort came from New American Pathways (New AP), a refugee resettlement nonprofit that helped 483 people start new lives in Atlanta last year. Facing a residency crunch, New AP needed to find new housing options for its families beyond the neighborhoods they had counted on previously. Determining the best apartments to contract with, however, means conducting research into an extensive range of factors that go beyond the willingness of landlords to work with the refugee community.
The city of Clarkston, as described by New AP CEO Paedia Mixon, has been an ideal place for resettlement for a number of reasons: “Small and walkable, with many apartments close to schools, faith centers, and transit lines, a reliable thrift store, a grocery that’s remade itself as a multi-ethnic food store,” and apartment owners excited to work with resettlement agencies as part of local revitalization efforts. The success of Clarkston, however, inevitably led to a new challenge: As the community became more established, apartments that once turned over regularly have remained occupied.
To discover new areas like Clarkston, and make them easy to identify and examine by New AP resettlement staffers, students set out to create a mapping tool that visualizes the assets most important to immigrants becoming self-sufficient in their new country. The first phase of the project was a data mining effort, using information from the Atlanta Regional Commission, the U.S. Census Bureau, Google Places, and GreatSchools, a national nonprofit that compiles detailed reports on education options for kids. Other dimensions required primary research, said Unaiza Ahsan, a PhD student who worked on the project: “Online listing of rental properties are not reliable for prices,” along with the possibility that certain apartment managers aren’t willing to rent to refugees (despite the fact, said Mixon, that refugees are very reliable tenants), while others don’t fit the nonprofit’s livability standards. “What the community looks like is also important.”
Good science, great results
Previously, Trees Atlanta was the subject of a Data Science for Social Good project, which equipped them with two powerful new systems: a Planting Visualization Tool, which maps 15 years of the nonprofit’s primary work, color-coding them based on genus, species, and date planted; and the Urban Tree Canopy Parcel Stratification tool, which details plant coverage across the city, broken down by council district, neighborhood, and individual lot. “Rather than having to drive around aimlessly looking for the right parcels of land,” said NeighborWoods Project Manager Alex Beasley, “I can drive virtually.” That allows him to prioritize neighborhood needs, and plan solutions, in a fraction of the time; he then brings that information directly to residents in a new, neighborhood-specific Canopy Conversations series to generate buy-in from locals. “These tools help us find these locations for planting trees, and demonstrate to neighborhoods why they need them.”
One reason the Data Science for Social Good class is so successful, said Dilkina, is that it’s set up like a 10-week paid internship, ensuring that each student dedicates 40 hours a week to their project. “Georgia Tech was the second university in the nation to use this model, which was started by the University of Chicago several years back with the help of a big sponsor, the Schmidt Family Foundation,” said Dilkina. “So many nonprofits are dying for help understanding data, but they don’t have the people, time, or resources to do it. Every year we have multiple requests from nonprofits, and we don’t have enough teams to address that need.”
Not that they want high demand to discourage any nonprofit thinking about a challenge that could use assistance from a dedicated team of Georgia Tech students. At a recent presentation, PhD student Ahsan made a sincere call for organizations to contact them at dssg-atl.io/get-involved, and for attendees to keep up with them through their blog.
“We were thrilled when we got the call from Dilkina,” said Mixon. The result of their partnership is an interactive map of the metro area embedded with multiple layers of information: on apartments, affordability, crime, job density, retail accessibility, and services like mass transit, faith centers, schools, hospitals, and government agencies. “Being able to sort this data in a visual way is incredible. Our next idea is to make it updateable, so we can share it with other resettlement agencies.”
David Terraso is communications director and Marc Schultz is contributing editor at GCN.