Video Analysis Reliability for Ski and Snowboard Cross Injury Events

644 reviewed situations from video footage of snowboard cross (SBX) and ski cross (SX) events were categorized by parameters like crash type and competitor behavior. The dataset was created by Björn Bruhin and published on figshare in April 2026. It supports a study evaluating the inter- and intra-rater reliability of a standardized video analysis methodology.

Use Cases

• Training machine learning models to classify injury-related events based on the described categories like Crash and Time of No Return.
• Evaluating the reliability of human raters in sports video analysis based on the reported Fleiss' Kappa and Cronbach's Alpha metrics.
• Developing standardized coding schemes for winter sports safety research based on the parameters of crash type, course trajectory, and competitor behavior.
• Benchmarking automated video annotation systems against human reliability scores for categories such as Out of Balance and Contact.

Strengths

• 644 situations were reviewed, providing a substantial sample for reliability analysis.
• A standardized process was established for training four raters, suggesting methodological rigor.
• Reliability was assessed using multiple statistical measures: Fleiss' Kappa, Cronbach's Alpha, and Cohen's Kappa.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The data is contained in a PDF file (6.2 MB), which may require extraction to be used for analysis.

Provenance

Source

figshare

Collection Method

A team of four raters reviewed video footage and categorized situations using a standardized process.

Freshness

Last updated 2026-04-09 05:20:04; freshness should be verified.