Swiss Court Decisions With Multilingual Text And Citation Links

962,000 court decisions from all Swiss federal courts and 26 cantons. The dataset contains full text, structured metadata in four languages (German, French, Italian, Romansh), and 8.77 million citation links, updated daily.

Use Cases

• Train a multilingual text classification model on full text decisions across DE/FR/IT/RM languages.
• Analyze citation links between 962,000 decisions to map legal precedent networks.
• Build a question answering system using full text and structured metadata from 101 courts.
• Perform temporal analysis of court decisions scraped from official cantonal portals.

Strengths

• 962,000 decisions provide a large-scale corpus for legal NLP tasks.
• Includes 8.77 million citation links enabling network analysis.
• Covers 101 courts across all 26 Swiss cantons for broad geographic representation.
• Updated daily, ensuring current legal data.

Limitations

• Column structure and specific metadata fields are unknown, complicating direct analysis.
• Sample data and file formats are unavailable for preliminary inspection.
• Potential language imbalance across the four languages (DE/FR/IT/RM) is unspecified.

Provenance

Source

Official Swiss federal and cantonal court websites.

Collection Method

Scraped from official court websites and cantonal court portals.

Time Range

null

Freshness

Updated daily.

Geography

Switzerland (all 26 cantons).