Ús d'anàlisi multivariant per a la definició de noves posicions de joc al basquetbol professional

Abstract

[en] Data and sports have been side to side for years. Anyone who watches sports during the weekend, and not only in professional teams, will most likely find people gathering data from those games: points or goals scored, fouls, the time a player has spent on court... Nowadays, mainly powerful teams are following a trend consisting on deeply analyzing these data with advanced methods. This project is an introductory example to these kind of studies. Using data from Liga EBA, the fourth tier in the Spanish basketball competition, and wellknown multivariate analysis methods such as CPA or k-means clustering, this project has as its main objective the classification of players depending on their statistical performance, escaping the classical guard - forward - center division. What this thesis looks for is a totally objective classification based in no more than statistics, following the footsteps of what M. Alagappan and his coworkers did in the NBA.