University of Virginia · School of Data Science
Associate Professor of Data Science/Quantitative Foundation Endowed Chair
I teach people how to move data from wherever it lives to somewhere it can answer a question — ingest, store, wrangle, visualize. I write about statistical methods for the social sciences, study how language models reason about math, and put students to work on data problems that local government and nonprofits actually have.
Every data project runs these four stages. My master's and Ph.D. courses walk students through all of them, in Python, on real and messy data.
Flat files, JSON, REST APIs, and web scraping — including the parts that break.
Relational, document, and graph schemas: PostgreSQL, MySQL, SQLite, MongoDB.
SQL, MongoDB's query language, and pandas — reshaping until the data fits the question.
matplotlib, seaborn, and plotly, built up into dashboards other people can use.
Written up in full, free and open: Surfing the Data Pipeline with Python
01 / Teaching
I teach the core data engineering sequence in the master's and Ph.D. programs. I directed UVA's online Master's of Data Science from 2022 to 2026, and before joining the School of Data Science I taught the graduate statistics curriculum in the Department of Politics.
Currently teaching
The master's core, residential and online: get data, store it, clean it, show it. Built as an active-learning course with labs, recorded lectures, and a companion textbook I wrote for it.
Developed for the launch of UVA's Ph.D. in data science in 2022. Environments and containers, remote repositories and workflow, ingest, database design, and dashboards.
Master's of Data Science Capstone Year-long team projects with external partners, residential and online — UVA School of Data Science
2020–2022Mathematics for Social Scientists Calculus, linear algebra, and probability for Ph.D. students — UVA Politics
2013–2019Linear Regression and Data Management R, the tidyverse, linear models, robust estimation — UVA Politics
2013–2019Generalized Linear Models Likelihood theory and custom likelihood functions for binary, ordinal, count, and duration outcomes
2013–2019Time Series and Panel Statistics ARIMA, vector autoregression, error correction, fixed and random effects
2013–2019Statistical Measurement Factor analysis, item response theory, scaling, cluster analysis
2013–2019Designing Your Own Maximum Likelihood Models Essex Summer School in Social Science Data Analysis, University of Essex
2015–2019Programming in R U.S. Department of State, Foreign Service Institute
2020Bryan Christ, Advancing Methodology for Artificial Intelligence for Math Reasoning and Education Chair
2025Kimberly Ganczak, Broadening the Political Methodologist's Toolkit: A Population Dynamics Model of Political Science Time Series Data Member
2021Danilo Medeiros, Extremism and Polarization: How Income Inequality Affects Legislative Behavior in Brazil Member
2019Robert Kubinec, Crony Capitalism, Democracy and the Arab Uprisings in North Africa Member
2018Min-Gyu Paik, Going Above and Beyond: Dependence and Military Coalition Participation Member
2016Covering the Housing Affordability Crisis in Charlottesville, VA Best Paper, Policy Track, IEEE SIEDS 2022 — Bozsik, Cheng, Kuncham, Mitchell
2022Investigating the Illicit Trade of Cultural Property with an Automated Data Pipeline Barraza, Landi, Lee, Naranjo-Velasco
2022Longitudinal Classification and Predictive Modeling for Historical CPS Data Johnson, Schmuckler
2022noderank: An R package for differential gene expression analysis Derby, Howlett, Dadlani, Chatfield
2021Measuring Digital Force Applied Toward the Public Through Police Surveillance Technologies Adams, Setser
2021Multi-Output Random Forest Regression to Emulate the Earliest Stages of Planet Formation Hoffman, Sung, Zazzera
2021Detecting Infrastructure Damage using Satellite Imagery and Neural Networks Boano, Tyree, Moore, Sarbanov
2021A LoRaWAN-Based IoT Air Quality Sensor Network for Public Good Howerton, Schenck
202002 / Writing
Two books, both written because the course I was teaching needed one.
Textbooks
Free online textbook
An open, browser-based book covering the whole pipeline: pulling data from files, APIs, and web pages; working with SQL and MongoDB; wrangling with pandas; and building visualizations and dashboards. Used in DS 6001 and open to anyone.
Read the bookSage, 2015
A first course in the math that quantitative social science actually uses — algebra and precalculus through calculus, probability, and linear algebra — written for graduate students who arrived without it and need it by the end of the semester.
View at Sage03 / Research
Three threads: how language models reason about mathematics and how to make them useful for teaching it; the policy and public trust questions facing American public media; and statistical methods for social science data — duration models, panel data, missing data, and measurement.
Selected work
Technology for the Public Good: Mapping a Fragmented Field Across Research and Practice
Policy Window as a Trapdoor: Rethinking the Multiple Streams Framework through the Defunding of U.S. Public Media
EduMath: Generating Standards-Aligned Educational Math Word Problems
An Island of Trust: Public Broadcasting in the United States
Math Neurosurgery: Isolating Language Models' Math Reasoning Abilities Using Only Forward Passes
MathWell: Generating Educational Math Word Problems Using Teacher Annotations
Beyond the Hazard Ratio: Generating Expected Durations from the Cox Proportional Hazards Model
Interpretation and Identification of Within-Unit and Cross-Sectional Variation in Panel Data Models
coxed Expected durations and marginal changes in duration from the Cox proportional hazards model
R · Statami Multiple imputation through iterative equations, with Ben Goodrich and Andrew Gelman
RCommunity
I co-founded Code for Charlottesville in 2019 and captain it today. Volunteers with data, design, and engineering skills take on projects for local government, legal aid, and area nonprofits — records expungement, housing data, public health, and more.
New volunteers are welcome, whatever your experience level.
Visit codeforcville.org04 / Background
I came to data science from quantitative political methodology, by way of a statistics postdoc.
Appointments
Associate Professor, School of Data Science University of Virginia
2024–Director, Online Master's of Data Science University of Virginia School of Data Science
2022–2026Assistant Professor, School of Data Science University of Virginia
2019–2024Assistant Professor, Department of Politics University of Virginia
2013–2019Postdoctoral Fellow, Applied Statistics Center Columbia University, with Andrew Gelman and Jennifer Hill
2011–2013Ph.D., Political Science, University of North Carolina at Chapel Hill New Approaches to Discrete Choice and Time Series Cross Section Methodology for Political Research
2011B.S. Mathematics (with honors) and B.A. Political Science (with distinction), Ohio State University
2005Quantitative Foundation Endowed Chair, School of Data Science
2024–Ph.D. Program Mentorship Award, School of Data Science
2025Ph.D. Program Teaching Award, School of Data Science
2024All-University Teaching Award, University of Virginia
2018Excellence in Graduate Student Mentorship Award, Society for Political Methodology
2015Co-chair, steering committee Public Interest Technology University Network
CurrentSteering committee Aperio, open access academic publishing
CurrentPh.D., Academic Affairs, and Math Curriculum committees UVA School of Data Science
CurrentPublic Service Pathways Steering Committee University of Virginia
CurrentReviewer Public Interest Technology Challenge, New America Foundation
Current05 / Contact
Prospective students, collaborators, and reporters are all welcome to write. Email is the surest way to reach me.
Where to find me
School of Data Science
University of Virginia
Charlottesville, VA