Skip to content

open-starlab/UFATrack

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 

Repository files navigation

UFATrack Dataset

UFATrack is a professional Ultimate Frisbee dataset released with synchronized video and tracking to support reproducible research and practitioner use.

Source / Scope

  • Match: Oakland Spiders vs Salt Lake Shred (score: OAK 4–9 SLC)
  • Date: June 27, 2025
  • Segment: 2nd quarter
  • Effective play included: ~9.5 minutes (excluding pulls and offense–defense transitions)
  • Possessions: 20 possession sequences
  • Filming: the left and right halves of the court were recorded separately
  • Actual play videos: video folder

annotated

Sampling & Coordinates

  • Coordinate system: metric (x, y) on a standard pitch of 109.73 m × 48.77 m
  • Per frame: positions of all 14 on-field players + the disc
  • Video: 30 fps, tracking: 10 fps

Disc position definition

  • Disc position is obtained by annotating the disc holder
  • When no player is in possession, disc location is linearly interpolated between consecutive possession events

Raw UFATrack columns (overview)

Each row = one object (player/disc) at one frame.

  • frame, id: frame number and object ID (class=="disc" indicates the disc row).
  • x, y: position (meters).
  • vx, vy, ax, ay: velocity/acceleration components.
  • v_mag, a_mag: speed/acceleration magnitude.
  • v_angle, a_angle: velocity/acceleration direction angle.
  • diff_v_a_angle, diff_v_angle, diff_a_angle: derived angle-difference features.
  • class: role label (offense / defense / disc).
  • holder: True for the disc-holding player at that frame.
  • closest, selected, prev_holder, def_selected: helper fields for analysis/preprocessing.

Generate Event/Tracking Data via OpenSTARLab PreProcessing

You can convert the raw UFATrack file into standardized Event data and Tracking data using the OpenSTARLab preprocessing pipeline. The preprocessing module merges and aligns raw inputs, normalizes coordinates, attaches context (team/period labels), and produces an analysis-ready space-data representation that can be used directly by downstream evaluation modules (e.g., SpaceEval).

tracking

Repository

OpenSTARLab PreProcessing (GitHub): here

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors