Text Data

For a long time, the processing of questions from sports and sports science with informatic methods dealt almost exclusively with numerical data, such as action or position data. However, knowledge about athletes, competitions, and their effects is often also available in textual form, such as countless scouting reports in junior academies of clubs and federations. Nowadays, advances in text recognition and machine learning allow the efficient analysis of large text datasets. Accordingly, so-called “text mining” methods are increasingly used in theory and practice, especially in disciplines that traditionally work a lot with data in text form such as open-ended questionnaires or standardized interviews. In this chapter, examples of three areas are shown, in which text mining was already applied in a promising way in sports: Evaluation of Technological Officiating Aids, Match Predictions and Talent Scouting.