Dataset features

• rftd contains human-applied semantic tags representing hundreds of thousands of legal and factual issues across more than 20,000 judgments. More detailed statistics are provided with the sample dataset.

• Semantic tags are applied following a strict QA process with auditing and redundancy.

• rftd includes human-readable descriptions of the semantic tags, tests for annotation decisions, and sample source text for each annotation.

• Each case record contains judgment metadata (date, judge(s), online sources, etc.) as well as machine-readable case outcomes.

• rftd is provided in JSON format for easy ingestion by machine learning models.

Please contact us for further information.