BEN SHABY Colorado State University

Associate Professor

I develop statistical methods and computational strategies, usually either explicitly Bayesian or with a Bayesian feel, with an eye toward environmental, ecological, and geophysical applications. This leads naturally towards thinking about spatial dependence, including dependence in extreme events.

Extreme values and Spatial Dependence

Extreme value theory is a useful tool for understanding specifically spatial dependence in rare events. This is an important idea if one wants to realistically model things like heat waves or rain storms, where extreme events have obvious spatial dependencies. Almost all known routes to achieving this modeling goal present significant computational challenges. I am developing tools to overcome some of these challenges, as well as exploring related modeling strategies. I still do some more traditional spatial statistics based on Gaussian processes as well.


For the past several years I have worked with the Finkbeiner lab at the Gladstone Institute at UCSF, which is trying to better understand how certain neurodegenerative diseases work. Examples include ALS, Parkinson's and Huntington's diseases. The experiments in the lab involve large-scale inference on complex phenotypes, and therefore often require novel statistical modeling.