Count-based football KPIs

Count-based metrics extracted from the 2016 UEFA Men’s European Football Championship and the 2017 UEFA Women’s Championship. Key metrics by player position and gender were extracted from match action logs and integrated as a single dataset. The resulting dataset of length $n = 4211$ contains 33 variables. The gender target variable is expressed as 1 for male players and 0 for female players. There are 2700 male instances and 1511 female instances. The dataset contains two categorical variables; match period is expressed as 1H for the first half, 2H for the second half, and E1, E2, and P for the two possible overtimes and the penalties respectively, player position in the team formation has the following options: Defender, Midfielder, Forward, Goalkeeper, Substitute Defender, Substitute Midfielder, Substitute Forward and Substitute Goalkeeper. Table \ref{tab:stats} shows the mean value and standard deviation per gender of each of the 30 numerical features of the dataset.