Bora Jin

Bora Jin

Statistical Science PhD Student

Duke University

About Me

I’m a 3rd year PhD student studying statistics at Duke university under the supervision of Amy H. Herring and David Dunson. I have interests in statistical applications in the environmental and human health sciences. I develop Bayesian methods that provide better interpretation in relation to the reality along with uncertainty quantification. My current work focuses on chemical toxicity, pollutant effects on diseases, and spatial prevalence of hazardous chemicals.


  • Environmental science
  • Spatial statistics
  • Multivariate methods
  • Hierarchical framework
  • Latent variables
  • Bayesian methods


  • PhD in Statistical Science, 2022 (expected)

    Duke University

  • Master's in Statistics, 2017

    Korea University

  • Bachelor's in Statistics, 2015

    Korea University

Current Research


Effects of Short-term Air Pollution on COVID-19

with David Dunson

Sep 2020 – Present
  • Build a Bayesian model to predict effects of PM2.5 with and without wildfires.

  • Include information about dynamics of wildfires and their progression into a Gaussian process (GP) using multiple directed acyclic graphs (DAG).

  • Investigate the short-term effect of PM2.5 on respiratory diseases including COVID-19.


Prediction for Per- and polyfluoroalkyl Substances in relation to Stream Networks

with Amy H. Herring

Sep 2020 – Present
  • Develop a predictive spatial model for per- and polyfluoroalkyl substances (PFAS) in public water system (PWS) along with uncertainty quantification.

  • Identify upstream and downstream information of PWSs in relation to the nearest stream network and construct a corresponding DAG.

  • Fit a multivariate GP using the DAG, incorporating external factors such as known sources or known relationship between PFAS compounds.


Bayesian Matrix Completion for Hypothesis Testing

with David Dunson, Julia E. Rager, David Reif, Stephanie M. Engel, Amy H. Herring

Jul 2019 – Present
  • submitted

  • Adapted Bayesian heteroscedastic nonparametric regression to a multiple hypothesis testing framework.

  • Imposed a generalized latent factor model to form a non-exchangeable prior for testing.

  • Developed a matrix completion method for a latent matrix.

  • Tackled sparsity of the ToxCast data using hierarchical framework.

  • Enabled prediction for non-tested chemical’s activity.

  • Broadened the definition of activity including heteroscedasticity.



Head Teaching Assistant

Duke University

Jan 2021 – Present Durham, NC, USA
Theory and Methods of Statistical Learning and Inference (STA432)

Teaching Assistant

Duke University

Jan 2019 – May 2019 Durham, NC, USA
Statistics (STA250)


International Atomic Energy Agency

Jan 2018 – Jul 2018 Vienna, Austria


United Nations Environment

Feb 2017 – Aug 2017 Geneva, Switzerland