Title: On Voting Preferences

Abstract:  The central problem I want to talk about is how to measure voter preferences in a ranked choice election.  Just from voters' rankings, can we "learn" whether there were distinct blocs of voters, who preferred distinct slates of candidates?  To study this, we came up with some interesting ways to measure differences between rankings and between ranked objects, and we developed some exact optimization (IP) approaches.  Much of this should be applicable to other studies of preference.  (There's a much bigger modeling literature on consumer preferences and web rankings than on voting, for some reason!)  This is based on joint work with Kris Tapp and with our own David Shmoys.

Bio: Moon Duchin’s research focuses on discrete geometry and randomized models, with applications to the study of voting and democracy. She is a member of the faculty at the Brooks School and the Department of Mathematics, affiliated with the Center for Data Science for Enterprise and Society. Duchin was hired as part of the Provost Office’s Radical Collaboration initiative. A prominent voice on fair redistricting, Duchin has developed mathematical models to analyze the potential and actual outcomes of changes to policy and voting districts. She has served as an expert in redistricting litigation in Wisconsin, North Carolina, Alabama, South Carolina, Pennsylvania, Texas, and Georgia. Recently, her work has turned to the study of alternative systems of election. Her research has been recognized with a National Science Foundation Faculty Early Career Development Award (NSF CAREER), a Guggenheim Fellowship, and a Radcliffe Fellowship. She was a Sloan Professor at the Simons-Laufer Mathematical Sciences Research Institute (SLMath) for their recent program on Algorithms, Fairness, and Equity.  She is a fellow of the American Mathematical Society. Professor Duchin received her Ph.D. in Mathematics from the University of Chicago. Her undergraduate degree in Mathematics and Women’s Studies is from Harvard University.

Panopto Webcast