Investigating a soccer dataset in Python
Posted on by Kevin Foong
Using Python to analyse a dataset on Premier League soccer matches around the metrics of "possession" and "matches won", reveals some interesting findings. This is a previous assignment I did for my Udacity Nanodegree in Data Analysis.
In season 2015/2016 Leicester City won the league at starting odds of 5000-1. Their possession stats (number of matches where they had more possession than the opposition) put them 3rd from bottom, but then we can deduce that perhaps they were also a lethal counter-attacking side.
In season 2013/2014 Southampton, a mid-tier team, led the table in possession but that didn't necessarily translate to matches won.
See my full investigation on Github in a Jupyter Notebook.