Principal component analysis identifies different representative match load profiles in international women’s field hockey based on playing positions. [El análisis de componentes principales identifica diferentes perfiles de rendimiento en función de las posiciones en partidos internacionales de hockey hierba femenino]

Reducing Big Data to Principal Components for Position-Specific Futsal Training

Since training/competition loads must be quickly assessed and interpreted to inform exercise prescription, big data should be simplified through multivariate data analysis. Our aim in the present research was to highlight which variables from big data analyses provided the most relevant information for describing the behavior of top-level futsal players in their different playing positions (i.e., goalkeeper, defenders, wingers, and forwards). We collected data from four top-level Spanish teams that participated in the final rounds of a national tournament. Through principal component analysis (PCA) we grouped 6-9 variables in 3-4 PCs that explained 62-81% of total variance, depending on playing positions. The most relevant variables explaining goalkeepers' performance were accelerations per minute, maximum acceleration (m/s²), 5-8 impacts per minute, and < 3 takeoffs per minute. Defenders' behavior was best explained by absolute distance covered from 6-12 km/h (m/min) and from 18-21 km/h (m/min), from 5-8 landings per minute, and > 8 landings per minute. Wingers' and pivots' performances were mainly explained by accelerations and decelerations, together with a high level of aerobic endurance (especially for wingers). These findings allow for individualized training and game analysis.