Leo Messi, Barcelona 10-11. So many green numbers, still just 23. He really isn't human. 5.86 successful drib90! pic.twitter.com/VaU5yqgNg8
— Ted Knutson (@mixedknuts) November 13, 2014
Part of the appeal of these charts is that players can be compared using overlays to see where their stats match up and where they differ. You can even overlay a player's previous seasons to see how the his skills or role might have changed over time. While I think these charts offer a lot of insight, I think the radar chart is a bit wanting as a method of conveying information. Comparing overlays in the same category is easy enough, but trying to eyeball differences across categories which are not adjacent on the chart can be tough. A similar method for displaying multi-dimensional data is parallel coordinates. A parallel coordinates graphic treats the different dimensions of the data as individual axes, and maps each data point to its corresponding position on that axis. One benefit of viewing multi-dimensional data this way is that the number of dimensions is not nearly as limited as in a radar chart or alternative visualization method. For n-dimensional data, a larger n quckly becomes an issue for a radar chart. Parallel coordinates charts are able to handle a larger number of dimensions. Using data from Objective Football, I built a graphic which visualizes multi-dimensional player data from the 2013-2014 Premier League season across 178 players from every team. Included are midfielders and forwards who played over 900 minutes. Apart from Age, Minutes, and Shot Percentage, every category is a rate statistic, meaning the number is per 90 minutes played. For stat definitions see Objective Football.