In the locked down midwinter I was introduced to Yuri!!! on Ice, at which point it was only a matter of time before a statistical analysis of international figure skating data appeared here.

It turns out that figure skating is very statistically interesting, especially since some recent rule changes. To make matters even better, the format fits quite closely with my experience of analysing football players, and there is a fantastic online resource called skatingscores where you can easily find all the data you need.

My aim with this series of posts is to illustrate all aspects of the kind of data analysis workflow that I like, from getting an idea in the first place, to fetching some data, putting it in a convenient format, modelling it, analysing and visualising the results and then drawing some interesting conclusions. Ideally the steps will roughly agree with this paper about Bayesian workflow.

The overall idea might be familiar if you’ve seen the amazing website FC Python. What you find here will likely be a bit more focused on my preoccupations (so figure skating, Stan, the pandas transform method, maybe a little bit of philosophy).

You can find all the code I use here, and here are all the posts: