Researchers at the University of California, Berkeley have released the Local Ordinance Corpus for the United States (LOCUS), described by its authors as “a comprehensive corpus and county-harmonized access layer for U.S. municipal and county ordinance codes.”
What LOCUS contains
LOCUS comprises roughly 2.2 million rows of data, with each row representing a specific piece of information tied to a particular local ordinance. The dataset is organized by the functional role of each entry — for example, a rule itself, enforcement of a rule, contextual material about a rule, or procedural aspects — and is also classified by topic. Topics included in the release cover buildings, businesses, zoning, nuisances, and an "other" category.
The authors emphasize that LOCUS-v1 is intended as an access layer rather than a final theory of local legal authority. In practice, they position the corpus as infrastructure for retrieval, comparison, and benchmark construction, not as a replacement for doctrine-sensitive legal analysis.
Why the dataset was needed
According to the paper, although local law is public, it is not practically available as a national research corpus. U.S. local codes are fragmented across commercial vendor platforms that are built for in-browser reading instead of bulk research. Vendors expose different navigation structures, print workflows, dynamically generated PDFs, and jurisdiction indexes. There is no central registry mapping every county or municipality to its hosting platform, and no vendor provides a complete machine-readable index of all jurisdictions it hosts.
Practical implications
Datasets such as LOCUS make the often partially hidden rules that govern much local civic life accessible to AI systems. This accessibility may allow AI to adapt more effectively to hyperlocal tasks and services, and it gives legal-AI researchers a foundation for building benchmarks and comparative studies.
Where to find LOCUS
The release includes coverage metadata to support reproducibility, downstream legal-AI research, and incremental expansion of machine-readable local law. LOCUS-v1 is available under the LocalLaws / LOCUS-v1 collection on HuggingFace, and the accompanying research paper is published as "Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States" on arXiv.
Key points
- The corpus was assembled by researchers at the University of California, Berkeley.
- LOCUS-v1 contains approximately 2.2 million data rows.
- Entries are organized by function and topic (buildings, businesses, zoning, nuisances, other).
- LOCUS is designed as an access layer for retrieval, comparison, and benchmark construction, not as a substitute for doctrinal legal analysis.
- Data and metadata are published via LocalLaws / LOCUS-v1 on HuggingFace, with the research paper available on arXiv.
