Welcome! This is the academic website of Jack O’Brien. I’m an applied statistician, data scientist, and a professor of mathematics at Bowdoin College. In addition to teaching, I spend most of my time thinking about how mixture models and related approaches can be used to analyze social science data (pottery sherds, election results, tweets) to create remarkably simple portraits of complex social phenomena.
Experience
- Associate Professor, Bowdoin College (2018-present)
- Senior Statistician, Wellcome Trust Centre for Human Genetics, University of Oxford (2010-2012)
- Research Associate in Statistical Genetics, University of Bristol (2008-2009)
Education
- BA, Pomona College (Mathematics)
- PhD, UCLA (Biomathematics)
Recent work
- (Dec 2025) Check out my recent(ish) paper using mixture models to map political ideology using election data: O’Brien. 2025. “A Bayesian mixture model captures temporal and spatial structure of voting blocs within longitudinal referendum data”. Currently in press in Political Science Research and Methods.
- (Dec 2025) My two honors students, Avery Cutler and Nathan Kellerman, made absolutely amazing mid-year presentations on their respective projects. Check out their projects here: [Avery’s on modeling eviction records in Boston] and Nathan’s work on a new EM algorithm for overdispersed count data.
- (July 2025) Very happy to see Assistant Professor of Physics Fe McBride’s paper, “A radio-quiet AGN as a candidate counterpart to neutrino event IceCube–200615A” (on which I made a very small assist) published in The Monthly Notices of the Royal Astronomical Society.