For this example, we will delve into the world of sports; in particular, the National Hockey League (NHL). Much work has been done on baseball (think of the book and movie, Moneyball) and football; both are American and games that people around the world play with their feet. For my money, there is no better spectator sport than hockey. Perhaps that is an artifact of growing up on the frozen prairie of North Dakota. Nonetheless, we can consider this analysis as our effort to start a MoneyPuck movement.
In this analysis, we will look at the statistics for 30 NHL teams in a data set I've compiled from www.nhl.com and www.puckalytics.com. The goal is to build a model that predicts the total points for a team from an input feature space developed using PCA in order to provide us with some insight on what it takes to be a top professional team. We will learn a model from the 2015-16 season...
