Chair of Social Data Science

Profile

The work at Professor Ratkovic's Chair combines work in causal inference, machine learning, and natural language processing.  Research is developed in service to understanding social phenomenon, including legislative speech, trade theory, and program evaluation.  Topics include integrating causal inference and deep learning, developing synthetic experimental studies, and simulating political negotiations using mixtures of agents.  An early contributor to the field of causal inference and machine learning, his award-winning work has evolved to include methods for high-dimensional regression, robust methods for propensity score estimation, recovering heterogeneous treatment effects, and the semi-parametrically efficient estimation of an average causal effect.  Looking forwards, the work at his chair will extend these works into natural language processing and large language models, with a particular emphasis on originality and rigor.