Although Bayesian methods in statistics are fundamentally sound, have a direct epistemic interpretation, and provide the researcher with optimal tools for inference, they are not yet widely adopted in practice. My research in the philosophy of science and statistics addresses some of the challenges to the widespread adoption of Bayesian statistics. Beyond this epistemological focus, my research also has a naturalistic metaphysical dimension, which I describe with the slogan “Statistics as Second Metaphysics.” Drawing insights from the history of physics and contemporary science, I explore how statistical reasoning underpins claims about what exists in our best scientific theories, and how these claims constrain the future development of scientific theories.
Here’s what I am working on:
Other Work
In my applied statistical research, I explore ways of streamlining Bayesian inferences with large datasets from both a conceptual and a computational perspective. On the conceptual side, I have examined the viability of using novel and computationally efficient techniques to fit Bayesian Gaussian Predictive Process Models on large spatial datasets. On the computational side, I have investigated the Informed Sub-Sampling MCMC (ISS-MCMC) algorithm, which is a scalable version of the Metropolis-Hastings algorithm for big data.
Repositories for some of the projects and my courses are available on GitHub.