Chenguang Pan

I am a fifth-year PhD candidate in the Measurement, Evaluation, and Statistics program at Teachers College, Columbia University, where I am advised by Professor Youmi Suk. I expect to graduate in May 2027.

I develop and extend quantitative methods at the intersection of causal inference, machine learning, and generative AI, and I apply them to help schools make better decisions for individual students using the data they already collect. A tour of my research is on the Research page.

Before pursuing my PhD, I spent nearly a decade at New Oriental Education & Technology Group in Beijing, where I was Deputy Manager of the American Graduate Test Department, as well as an instructor and researcher.

News

2026.10 Chenguang released pytlverse, a Python port of the R tlverse packages for TMLE and optimal treatment regimes, in both single-stage and longitudinal settings. In benchmarks against the R tlverse, it ran about 10 times faster while using about one-sixth of the memory, which makes large-scale simulation studies practical.
2026.10 Chenguang’s paper Estimating Optimal Dynamic Treatment Regimes for Personalized Education: A Tutorial and Applications with Machine Learning (with Yuxuan Li, Thomas Harris, and Youmi Suk), accepted by Zeitschrift für Psychologie, is now in press. The preprint is available on arXiv.
2026.09 The syntax reference Chenguang built for HUDM 5001 Programming for Data Science is now open to everyone at PythonTools.

2026.07 Chenguang will be an instructor this fall for a graduate-level course, HUDM 5001 Programming for Data Science, at Teachers College, Columbia University.
2026.06 Chenguang will give two oral presentations at SREE 2026 in Baltimore.
2026.03: 🎉🎉 Excited to share EDM-ARS (Educational Data Mining Automated Research System) — a domain-specific, multi-agent LLM pipeline that turns a dataset + research prompt into a research paper.
2025.12: 🎉🎉 Chenguang was awarded the Morton T. Embree Outstanding Contribution to Student Learning Award in recognition of his service as a TA.
2025.11: 🎉🎉 Chenguang will give oral presentations at both AERA 2026 and NCME 2026.
2025.06: 🎉🎉 Three works that Chenguang was involved in were accepted by SREE 2025, and Chenguang will give an oral and a poster presentation.
2025.06: 🎉🎉 Chenguang was invited to be a reviewer for the 2026 AERA Annual Meeting.
2025.05: 🎉🎉 Chenguang gave an oral and a poster presentation about his in-progress work Designing Realistic and Interpretable Optimal Treatment Regimes for Personalized Education in ACIC 2025, and Foundations of Data Science Workshop organized by Columbia Data Science Institute
2025.05: 🎉🎉 Chenguang served as a poster reviewer for the American Causal Inference Conference (ACIC) in Detroit.
2025.01: 🎉🎉 Chenguang received a sponsored registration for the 2025 AERA Annual Meeting, AERA 2025.
2024.12: 🎉🎉 Chenguang was awarded the Provost’s Grant for Conference presentation, Columbia University, NY, U.S.
2024.11: 🎉🎉 Chenguang’s paper Designing Optimal Dynamic Treatment Regimes Using TMLE for Personalized Math Course-taking Plans (joint work with Youmi Suk) was accepted by the American Educational Research Association (AERA) conference 2025, Denver, CO, U.S.
2024.09: 🎉🎉 Chenguang’s work (as second author) Designing Personalized Math Course-taking Plans in High School Using Optimal Treatment Regimes was accepted by and presented at the Society for Research on Educational Effectiveness (SREE) conference. Baltimore, MD, U.S.
2024.07: 🎉🎉 Chenguang’s independent research paper Examining Algorithmic Fairness in Predicting High School Dropouts was accepted by and presented at the 17th International Conference on Educational Data Mining (EDM) 2024. Additionally, he was awarded the IEDMS Scholarship by EDM. Atlanta, GA, U.S.