lexgenie/echr-guide-citation-diffs-cleaned

Dataset Card for ECHR Guide Citation Diffs Cleaned

Dataset Summary

This dataset contains cleaned and reaudited citation additions and removals across consecutive versions of European Court of Human Rights guide PDFs.

It is a derived release built from the original lexgenie/echr-guide-citation-diffs dataset, which provides the underlying guide snapshots and citation-diff corpus. This cleaned release repackages the diff layer into analysis-friendly CSV and JSON files and includes the outputs used in our re-audit workflow.

Dataset Structure

Main files:

• cleaned_citation_diffs.csv: flat event-level table, one row per citation addition or removal
• cleaned_diffs_grouped.json: grouped transition records, one item per guide version pair
• cleaned_extracted_citations.json: extracted citation lists per guide snapshot
• comparison_details.json: detailed comparison metadata
• reaudit_summary.md: brief notes on the cleanup and re-audit process

Dataset Statistics

• grouped guide transitions: 103
• flat citation-change rows: 1004

Intended Use

This dataset is intended for:

• studying temporal citation changes in ECHR guides
• identifying which cases are added or removed between guide versions
• serving as the base layer for downstream case linking and guide-update modeling

Provenance

Derived from:

• lexgenie/echr-guide-citation-diffs

This cleaned release should be treated as a downstream processing layer over the original citation-diff corpus, not as a replacement for the underlying PDF archive.