MDI Fellow Feature: Nathan Wycoff
Written by Carrie McDonald, MDI Journalism Intern
Nathan Wycoff, Ph.D., a Data Science Fellow, has been at MDI since 2021. He has mainly worked on variational Bayesian methodology applied to forced migration and has recently started working on differentially private synthetic data techniques applied to government data.
Wycoff said that he enjoys the “unique dynamic” at MDI, where he can intensely study statistics within a “world-class cadre of faculty who think about policy issues all the time and have really great questions that we can’t answer yet, but, make us think about things that we wouldn’t if we were just in a regular math department.”
“I got my degrees from an engineering school at Virginia Tech, but I’ve always been very interested in global affairs and government policy,” Wycoff added. “So it was really exciting to get the opportunity to come work on this really important social issue [of forced migration] and to do that with some of the people that are actually dealing with the problems.”
Within MDI’s forced migration team, after the Russian invasion of Ukraine in 2022, Wycoff is particularly proud of his work on research focused on the use of social media and other internet data to try to better understand forced migration patterns. The team found that some data sources, for example, Google searches for “how to leave Ukraine,” were particularly useful predictors of a surge in border crossings. Wycoff hopes that this finding can help the necessary stakeholders plan for refugee arrivals during future displacement crises.
In addition to his work on the forced migration team, Wycoff recently started working with Professor Amy O’Hara, Straus, and Carey on privacy for administrative data. These projects include trying to help local and state governments understand whether their programs, such as educational programs and welfare programs like the Supplemental Nutrition Assistance Program (SNAP), are effective. Wycoff is most excited about a particular project that aims to hopefully help people improve their employment prospects by evaluating the benefits of particular educational programs.
“People right now kind of have to guess whether a particular educational program is going to help them out or not,” Wycoff said. “We’re building this statistical framework that’s going to allow a lot of educational providers and other training providers to evaluate how well their programs are running. […] We can make sure that federal and state money is going to the right programs that are having the most impact.”