Nathan Kellerman

Statistics Ph.D. student @ UConn

Hello! This Fall, I'll be a first-year Ph.D. student in the Department of Statistics at the University of Connecticut. My research interests involve using applied mathematics and statistics to model social processes through time.

I've recently enjoyed thinking about how mixture models can be used to infer latent political cultures across time and space in US states and Swiss municipalities.

  • Ph.D. in Statistics — University of Connecticut, Storrs CT (2026—)
  • B.A. with Honors in Mathematics — Bowdoin College, Brunswick ME (2026)
Email: [first] [dot] [last] [at] uconn.edu
LinkedIn GitHub

Recent Updates

  • (May 2026) Defended undergraduate thesis, in which I developed a rapid model-based algorithm to infer latent mixture structure in social science count data. This work, which is linked below, was advised by Prof. Jack O'Brien.
  • (April 2026) Awarded Best Overall and Best Visualization prizes at the 2026 ASA DataFest event, hosted at Colby College.

Research Interests

  • Understanding societal and individual behavior mathematically.
  • Spatially modeling social processes and networks.
  • Statistical machine learning, applied Bayesian statistics, high-dimensional social science data.

Publications

  • [PDF] Nathan Kellerman (May 2026). A rapid algorithm for inferring latent mixture structure in replicate social science data. Undergraduate Honors thesis.
  • Chidinma Ezugwu, Nathan Kellerman, Thomas Weighill, Benjamin Yam. (2025+). Size-aware topological analysis with accumulation barcodes.

Talks

  • [PDF] A rapid algorithm for inferring latent mixture structure in replicate social science data. May 2026. Bowdoin College Mathematics Department Honors Thesis Defense.
  • Finite mixture modeling replicate-rich data: quickly approximating Bayesian posteriors for interpretation in the social sciences. April 2026. ASA Symposium on Data Science & Statistics. (accepted to give refereed talk)
  • [PDF] Mixture modeling political cultures (and more?). December 2025. Bowdoin College Mathematics Department Honors Research Mid-Year Talk.
  • Understanding the shape of data with accumulation barcodes. November 2025. UNC Greensboro Regional Mathematics and Statistics Conference.

Teaching Experience

Bowdoin College
  • Teaching assistant. MATH2606: Mathematical Statistics (Spring 2026).
  • Teaching assistant. MATH2208: Ordinary Differential Equations (Fall 2025).

Other Interests

  • Printmaking, tennis, science fiction, swimming, poker.