Analysis of 2016-17 MLS Season Data Using Poisson Regression with R

To the outside observer, soccer is chaotic with no given pattern or scheme to follow, a random conglomeration of passes and shots that go on for 90 minutes. Yet, what if there was a pattern to the chaos, or a way to describe the events that occur in the game quantifiably. Sports statistics is a critical part of baseball and a variety of other of today’s sports, but we see very little statistics and data analysis done on soccer. Of this research, there has been looks into the effect of possession time on the outcome of a game, the difference in passing 5 minutes before and after a goal occurred, and very little else. In this paper, I present an approach to analyzing the passing schemes using a statistical approach to uncover sometimes-nonobvious insights to the game of soccer. I illustrate the utility of my methods by applying it to data from the 2016-2017 Major League Soccer season collected by americansocceranalysis.com. This data includes passes into each section of the field, passes that lead to shots on goals, goals themselves, assists, touch percentage, and much more. By analyzing this data with the statistics software R and with the use of some descriptive statistics results, boxplots, histograms, etc. we will be able to get a physical representation of the data and make conclusions based on them. I would like to explore the idea of the effect of passing on scoring or how the touch percentage of each player contributes to their performance and the outcome of games. The overall goal though is to find results and hopefully spark an interest into further research and data analysis for the beautiful game