Although analytics is being used in, e.g., the evaluation of players and scouting, it is still challenging to quantify skills and playing styles of players. Such information is important for roster creation and scouting, where teams want to find players that have a playing style that fits within the team, as well as for game preparation to understand the playing style of opponents. In this paper we use player vectors to characterize a player’s playing style. The player vectors contain representations of skills that are computed from game event data. Further, we use fuzzy clustering on the vectors to generate five types of defender playing styles and five types of forward playing styles. For these types, we show the typical skill levels and players with similar styles.