For researchers
Bulk dataset & DuckDB tier
The full normalized corpus — provisions, ML labels, the 42-subject taxonomy, and the rollup cube — as downloadable Parquet, plus DuckDB-over-R2 for ad-hoc analysis beyond the pre-aggregated API.
What’s coming
- ◆Parquet exports: provision_labels, provision_subjects, index_rollups, and the provisions_labeled fact table
- ◆DuckDB-over-R2 endpoints (/v1/corpus/search | stats | aggregate) for full-corpus queries
- ◆A documented, stable schema with GEOID / OCD-ID / URN / FRBR identifiers
- ◆Versioned snapshots with a changelog, so research is reproducible
- ◆Academic access terms; the legal text is public domain
The underlying data already exists in the engine — this surface is being built next.