This research examines the relationship that exists between performance and external variables with the player’s market value set by the crowdsourcing platform Transfermarkt.com. Sports media journalists, fans, sports agents, and managers have been using this rating as an official information source in the last decade. The present research contributes to current models by including external variables, which represent the role of the media and social media through Twitter. Also, this study compares the performance of three models; first, a performance (only talent based variables) model; second, a non- broadcast (including non-weighted variables) model; and third, a broadcast (including weighted variables) model.

The analysis takes the English Premier League (EPL) 2015/2016 season as a sample. Tweets from twenty teams and seven sports media organizations are used to create the external variables. The results illustrate that a regression model with weighted variables is a better predictor of a player’s market value than a model with non-weighted variables. The public recognition from teams of opposing players proves to be a significant variable that influences market value. This contrasts with the influence of the media that is not a significant factor.

Language

English

Publish year

2017

Tools or skills

exploratory data analysis, text mining, statistics, R, writing, web scraping, API use, sentiment analysis, social media analysis

Publisher or Organization

University of Tilburg –– The Netherlands