The Barcelona School of Economics Intensive Course on Bayesian Learning and Applications for the Social Sciences introduces modern statistical, Bayesian and machine learning methods for analyzing large and/or complex data. We place particular emphasis on problems with many covariates, estimating treatment effects under high confounding and flexible modeling frameworks that allow one to tackle a wide variety of complex problems. These include latent discrete variable models for unstructured data such as text, non-parametric regression methods to incorporate non-linearities and probabilistic forecasts. We also introduce modern computational methods and programming software that allow one to implement any desired model and showcase how to do so via case studies.

We focus on problems where one wishes to obtain statistically valid inference on parameters, hypotheses, or probabilistic forecasts, rather than on purely point predictions. We discuss Bayesian methodology as a natural probabilistic machine learning framework for such tasks, and their interplay with complementary penalized likelihood and conformal inference methods. 

The program combines a series of lectures by leading researchers in the development of Bayesian methods and their applications, and practical sessions with hands-on data analysis, all of which will be offered simultaneously in R and Python (see requirements below). We emphasize applications in Economics and the Social Sciences, although the presented ideas are also widely applicable to other fields.

CODING KNOWLEDGE REQUIRED

Those with no experience in either R or Python, or who haven’t used either for a while, must also register separately for the course Coding Bootcamp in Python and R, which will be conducted online on Friday, April 29.