Clustering NHL Players

####Clustering NHL Players

You can access the interactive similarity map here

The goal of this project was to determine the similarities between different NHL forwards, and to label groups of similar players by role.

The players were clustered relative to the rates of nine different stats, like goals per 60 minutes of ice time. The clusters were labeled and the data was projected into two dimensions and plotted. The distance between players corresponds to the similarity in playstyles, so exceptional and unique players like Crosby and Datsyuk are isolated in the fringes of the map.

Project Details

Date: Aug 18, 2015

Author: Jake Popham

Categories: project

Tagged: nhl, clustering, visualizaton, interactive



Related Works.

Clustering NHL Players

Analyzing the NHL Draft

Identifying Users Across Multiple Devices

Search Result Relevance

Liberty Mutual Property Inspection