One promising application of data analytics in soccer is the identification of groups of players with similar playing styles. This can enhance recruitment processes and assist in tactical decisions, such as substitutions during a match. Previous research has explored various methods to measure a player’s contribution, identify key variables for player positions, and predict market value using machine learning techniques. However, these efforts often focus on traditional player positions or a limited set of gameplay characteristics, lacking a comprehensive analysis of player participation. This paper introduces a novel approach to clustering similar soccer players using supervised clustering and dimensionality reduction. Unlike previous studies, our method aims to refine the well-known positions of players and incorporates a wide range of in-game metrics to provide a holistic view of each player’s playing style.