Hey there! My name is

Tracy Morrison.
I build statistical methods.

I’m a PhD student in Statistics working on Bayesian small area estimation, constrained models (benchmarking + inequality constraints), and practical computation.

Get in touch

01. About Me

I’m a PhD student in Statistics interested in Bayesian modeling for small area estimation, uncertainty quantification, and applied work in demography/public policy.

Recently, I’ve been focused on efficient Gibbs sampling for constrained heteroscedastic area-level models and how constraints affect interval coverage.

A few tools I use:

  • R / Stan / Python
  • Bayesian modeling + MCMC
  • Small area estimation
  • Git, reproducible research
Portrait

02. Where I’ve Worked

Research Assistant @ Department / Lab

Aug 2023 — Present

  • Developed constrained Bayesian SAE models with benchmarking + inequality constraints.
  • Implemented efficient samplers and diagnostics for coverage and RMSE.
  • Applied methods to demographic indicators (e.g., fertility rates) and policy-relevant outputs.

Predictive Modeling Intern @ Company

Summer 2026

  • Built statistical models for pricing/forecasting using structured and unstructured data.
  • Delivered reproducible pipelines and documentation for stakeholders.

Teaching Assistant @ University

2024 — 2025

  • Led labs and office hours for regression / Bayesian / applied statistics courses.
  • Created clear explanations and grading rubrics; supported student projects.

Reviewer / Student Leadership @ Community

Ongoing

  • Reviewed submissions and supported inclusive academic communities.
  • Organized reading groups and workshops.

03. Some Things I’ve Built

Add a screenshot: assets/project1.png

Featured Project

CHALM / Constrained SAE Sampler

Efficient Gibbs sampler for constrained heteroscedastic area-level models with benchmarking and inequality constraints, plus coverage diagnostics and runtime improvements.

R • Bayesian SAE • MCMC • Diagnostics

Learn More

Other Noteworthy Projects

↗

Coverage Diagnostics Toolkit

Reusable scripts to evaluate interval coverage under benchmarking and inequality constraints.

R • Simulation • UQ

Spatial SAE Prototype

Explored spatial random effects and compared RMSE and interval scoring across model classes.

Bayesian • Spatial • Evaluation

Applied Demography Case Study

Small area estimation for rates with benchmarking totals and inequality constraints in practice.

SAE • Demography

04. What’s Next?

Get In Touch

If you want to chat about research, collaborations, or internships, my inbox is open. I’ll do my best to respond quickly.

Say Hello