Sports Data Hub (SDH) is pleased to announce the OpenSDH Analytics Research Contest. OpenSDH is a collaborative project between Sports Data Hub, sports industry professionals, and technology companies to build a common sports analytics platform. As part of the analysis portion of the OpenSDH project, we have decided to engage the skills and creativity of the larger analytic community through sponsoring a contest using real professional data. We hope the result will be the further advancement of sports analytics techniques as well as more sports industry opportunities for contestants. In a split from normal contest formats, we are requesting that contestants provide their solution as a short paper (5-10 pages is sufficient) describing their approach along with associated R code. This enables pro teams using the OpenSDH platform to quickly try out contestant solutions on their own data and the contestants end up with a tangible output to help establish professional connections to sports teams.
First prize is a presentation spot at the 2013 New England Symposium on Sports in Statistics (NESSIS) to be held September 21, 2013, as well as $500 + hotel + airfare (for 1 person). Recognition will also be given to the best “Out of the Box” idea. This award recognizes entries that have broken radically new ground and does not have a prize associated with it.
Based on player GPS position and team possession data, the goal of the contest is to learn about various game success measures from team positioning dynamics. The types of game measures of interest: scoring, turnovers created/lost, possession time, and ground gained/lost against an opponent. (Example analysis themes are outlined in the Contest Details document link below)
For the competition, we will be releasing a “small” set of GPS player positioning data for three games from a professional sports league (approximately 1.75 million rows of data). This positioning data is typical of worldwide professional sports teams that utilize newer GPS systems for player tracking.We will also include possession data that represents outcome results.
We have posted several public documents on dropbox that provide further contest details, limitations, background, and FAQs.