The Travelling Soccer Player Problem (TSPP): How to Predict Player Movement Based on Historical Data

Theoretical (Analytical):

Practical (Implementation):

Literature Work:


  • Modern soccer clubs can be regarded as corporate entities, with the soccer team and its successful operation at the center.
  • The game analysis department is directly connected with the coaching team and employs video analysts. The task of these experts is to identify strengths and weaknesses of their own team and of opponents, both in retrospective of historic matches, and in anticipation of upcoming matches.
  • Their findings are used to adjust the training and thereby raise the team’s awareness for dangerous situations, preparing for matches.

Problem Statement

  • Analysts cannot inspect all available past matches manually
  • How can we - during a "live" match - predict what a player will do next? (e.g., where will the player move?)


  • Work with a large soccer dataset containing 60 soccer matches of two teams
  • Use the available data to define appropriate predictors to anticipate future player movement for a given situation
  • How can we visualize our prediction?
  • (Can we detect the player type based on his/her movement?)
  • Think about ways to measure and display the uncertainty of your prediction


  • Basic knowledge about data mining and information visualization
  • Good programming skills in Java


  • Scope: Bachelor/Master
  • 6 Months Project, 3/6 Months Thesis