I wanted to write a quick post to introduce two repos that I created on Github that center on the NHL’s play-by-play data.
1. My code repo that highlights how I crawl and store the raw JSON datasets can be found here
2. I also created a very simple R Shiny App. That repo is here.
In both cases, the codebase has a lot of room to improve. The Shiny app is a poor-man’s version of the icetracker viz that can be found on the NHL website. I really want to refine the Shiny App. It is the first that I have created, so please keep that in mind when offering feedback.
With that said, I have fit a predictive model that estimates “Shot Quality”, which is simply the predicted probability of a shot going into the net. In the future, I plan on writing longer posts to introduce each repo individually and bring forward how the play-by-play datasets can allow us to do more advanced analyses beyond the basic boxscore.
For now, feel free to clone the Shiny Repo and and fire it up in your browser while you watch Game 1.
Here is a screenshot of the super-basic app…..