Baseball4D: A tool for baseball game reconstruction & visualization

While many sports use statistics and video to analyze and improve game play, baseball has led the charge throughout its history. With the advent of new technologies that allow all players and the ball to be tracked across the entire field, it is now possible to bring this understanding to another level. From discrete positions across time, we present techniques to reconstruct entire baseball games and visually explore each play. This provides opportunities to not only derive new metrics for the game, but also allow us to investigate existing measures with targeted visualizations. In addition, our techniques allow users to filter on demand so specific situations can be analyzed both in general and according to those situations. We show that gameplay can be accurately reconstructed from the raw position data and discuss how visualization and statistical methods can combine to better inform baseball analyses.