We deconstruct the basics of Bayesian hierarchical modeling and then use Fantasy Premier League statistics to construct a model in PyMC that tries to guess the remaining season's results. Along the way, we talk about crucial modeling choices,
Sports analytics is a growing arena for data analysts nowadays, and obviously, the Bayesians are coming for it! We will take a good look at data and figure out what matters for predicting points. No worries if you are not a football fan: we will take a quick look at the rules. In this talk we demystify partial pooling, priors, and other fancy terms statisticians use. Once we have some imperfect but somewhat functional predictive outcomes, we try to understand our results and takeaways. If you want to get started with (or build a bit on) Bayesian statistics or if you are just excited about football, this will be fun.