By Rob Mitchum

Building a data science for social good community means welcoming in people from all stages of their education and career. Training undergraduates and grad students at the onset of their life in the workforce offers valuable preparation for their first jobs after school, whether those lie in the non-profit, public, or tech sectors. But it’s also important to bring in fellows who have already applied data science in the world beyond academia, often before returning to school to strengthen their skills in computer science, statistics, or other disciplines.

Since enrolling at the Air Force Academy in 1995, Ryan Kappedal has spent his two decades in military service everywhere from Iraq and Kuwait to Japan and Korea to Alaska and the Pentagon. At each of these stops, Kappedal utilized his data analysis skills for operations research — identifying trends and activities in Iraq to prevent crime and insurgency, developing new methods to reveal underground structures, testing new technologies before they were fielded, and conducting cost analyses for large-scale defense projects. To further develop his skills, Kappedal completed a Ph.D. program in statistics at the University of Washington, complementing his on-the-ground experience with new techniques.

“I got a good view as to what decision-makers care about and how to build a case for them,” said Kappedal, 38. “I also, importantly, got a good view of the dirtier side of data collection in a field environment, and how decisions are implemented at a high level of policy. The engine in between was filled out by academia.”

Currently, Kappedal teaches at the Air Force Institute of Technology near Dayton, Ohio, where he is helping construct a new curriculum on data science and machine learning. He hopes to update the traditional military role of operation research systems analyst (OSRA) — which originated in the quantitative analysis teams of World War II — with the more multidisciplinary role of data scientists, equally comfortable with statistics and computer science.

He’s also interested in broadening his work beyond military research; for example, developing new methods to analyze neuroimaging data and advance the diagnosis and treatment of traumatic brain injury and post-traumatic stress disorder. At DSSG, he hopes to expand his network and draw upon his familiarity with fast-paced team projects.

“I’m really looking forward to energizing and expanding my network of really good talented data scientists,” Kappedal said. “A lot of what we do in the military involves throwing people together and saying you must operate as a team. That’s what a deployment is like, so I’m looking forward to that.”

While data analysis has long been used by the military, the field of urban planning is just starting to use the information and computational tools at its disposal to build better buildings, neighborhoods, and cities. As a city planner in Tel Aviv-Yafo Municipality in Israel, Talia Kaufmann, 32, was disappointed in the untapped potential of these methods for city planning globally.

“I decided I wanted to understand the system from within before I changed it,” Kaufmann, 32, said. “There are a lot of flaws in the process of planning cities, and nothing data-driven. There’s a lot of decisions made daily all around the world without really understanding their implications.”

Kaufmann’s search for a better approach led her to the Massachusetts Institute of Technology, where she completed a master’s degree in urban planning. Her studies and a position with the MIT Media Lab inspired her dissertation work, a “parametric” approach to planning land use in cities that uses building blocks of urban features (restaurants, parks, residences, etc.).

“The way that I’m approaching this building block idea comes from the thought that planning cities is like playing with Legos,” Kaufmann said. “The difference between urban planning and Lego is that Lego combinations are infinite, while in the city there are lots of constraints. So a major question of my thesis was what are the constraints, and how can I constrain possibilities using scenarios that are more likely than others.”

[A video on parametric planning from Kaufmann’s portfolio website.]

In the future, Kaufmann sees herself returning to city planning departments, where she hopes to apply the quantitative strategies she’s developed in academia and her personal projects to reform public practices. Like many DSSG fellows, she sees the challenge and promise of data science to bring real change and additional impact to entrenched fields.

“I think what’s going on in the world, there’s a lot of problems to solve, and a lot of people coming from the realm of the problem that see the issues, but not a lot of openness in the community they are part of to start doing data science,” Kaufmann said. “I’ve come across several people trying to do similar things, where you see how much potential there is and how much you can do, but bridging the worlds is a big challenge.”