An analysis of NHL player data using modern statistical learning techniques.
Ten seasons of data were collected from the NHL's publicly available database at nhl.com/stats for the 2009-2010 season through the 2018-2019 season. Each observation corresponds to a set of statistics measured for a single hockey player in a single season. All in all, there are 8,853 observations over more than 100 variables.
See data/features.pdf for a description of the features recorded.
See src/init.R for the initial data cleaning.