Current Research

Computational Statistics for Social Science

My research develops Bayesian mixture models to find interpretable structures in complex data. While I started in statistical phylogenetics and genetics (viral evolution, malaria genomics, microbial metagenomics), these days I work mainly on social science: political structure in election data, cultural patterns in the archaeological record, and the rhythms of social life online.

Statistical innovations

The Beta-binomial mixture model. How to fit it. How to nail down the number of blocs. Some pieces of pottery. Extensions.

Mapping voting blocs

How to map voting blocs? Blocs are neither left nor right but highly spatial and stable across elections. Maine politics redux.

Local politics

How local is politics? Mapping the demographic structure of voting blocs across the US, including California, Miami, and NYC.

Rotating 3-D plot of Swiss communes on the first three principal coordinates of their referendum votes, 1995–2005, coloured by voting bloc

Politics worldwide

What structures voting blocs within and across countries? Case studies in the Netherlands, Switzerland, Croatia, Botswana and Korea (to start).