BIG DATA CUP 2027

2027 Big Data Cup

Data Sets

Stathletes, Rotman School of Management, and the University of Toronto Sports Analytics Student Group (UTSPAN) have formed a dynamic three-way partnership to host The Big Data Cup, combining industry expertise, academic insight, and student innovation to tackle real-world challenges in sports analytics. The Big Data Cup invites participants to explore hockey through data.

The 2027 dataset brings together 3D goalie movement and puck tracking from a shooting session, opening up new ways to study shot difficulty, goaltending, and tracking data quality. It combines goalie joint positions with puck position and velocity on a shared time axis. The current processed sample contains:

• 81 shot sequences from one goalie in the August 6, 2026 session, each running from the start of the pass until the goalie stands up after the shot
• 19 goalie joints, tracked at approximately 60 frames per second
• Puck tracking at 60 Hz

Each shot is provided as a JSON file containing aligned goalie and puck tracking, plus session and shot metadata.

Access the data here.

three research areas

This year’s Big Data Cup suggests three research areas. The challenge is open-ended: develop your own research question, introduce a new metric, or build a method that helps others work with the data. Suggested questions will be updated ahead of the data release and may be refined.

Please identify in the title which area you will be focusing on:

1. Shot Difficulty

What makes a shot difficult to save?

• Examine puck speed, trajectory, elevation, and flight time.
• Relate the goalie’s starting position and depth to the movement required for a save.
• Identify areas of the net that may be difficult to defend.
• Explore outcome models where suitable shot outcome labels are available.

2. Goalie Fatigue and Movement Consistency

How does goalie movement change over a shooting session?

• Examine changes in reach, movement speed, stance, and positioning across shots.
• Compare movement consistency throughout the session.
• Account for shot location, sequence, and tracking quality when interpreting changes.

Changes in movement alone do not establish fatigue. Consider alternative explanations, including shot mix, adaptation, and tracking drift.

3. New Hockey Metrics

What useful aspect of goaltending or shot difficulty could a new metric capture?

• Goalie readiness: pre-shot stance, stability, and positioning.
• Movement efficiency: goalie movement relative to a feasible interception movement.

Define what your metric measures, explain its intended use, and demonstrate how you assessed its reliability.

2027 Timeline and Key Dates

• Data release: To be announced.
• Submission deadline: To be announced.
• Results: To be announced.

Getting Started

1. Read the data dictionary and inspect one shot file, including timestamps, units, joint definitions, and missing values.
2. Plot the puck trajectory and goalie pose. Identify gaps and unusual values before calculating derived measures.
3. Choose a focused research question and establish a simple baseline.
4. Explain your method, findings, validation checks, and limitations.

The SPL repository will provide animation utilities alongside the release. Start with the sample and documentation linked above.

Submission Rules and Guidelines

Registration, eligibility, team sizes, submission requirements, judging criteria, and prizes will be announced here. Data-use terms and guidance on AI tools will be provided before the competition opens.

Questions

Participant support and contact details will be added before the data release.