The Digital Governance and Analytics Lab's Global Political-Unit URL Census

The largest independent, verified, provenance-tagged, cross-national dataset of official local-government and political-unit websites: 755,700 administrative units across 187 ISO-3166 country files, 148,175 URL-bearing rows in measured-quality layers (14,176+ independently verified at 92-100% audited per-stratum precision; ~130,000 inherited with per-URL quality flags; 3,826 recovered in a recall-repair campaign anchored by independent blind gold standards). Uniform 55-column schema: identity, national-code crosswalk (87.2% of URL-bearing; INSEE/GNIS/AGS/ISTAT/LAU/OKTMO/PSGC and 18+ schemes) + GeoNames interlink, machine-legibility signals (sitemaps, robots/AI-crawler policy, llms.txt, structured data), security posture (HTTPS/TLS/certificate issuer/HSTS), governance-page and e-service depth, and hosting/sovereignty layers. Pipeline recall is measured and reported per country (blind-anchor audited); absence is evidence: France, Algeria, and Morocco floors are corroborated by the states' own registries. Derived from OpenStreetMap (\u00a9 OpenStreetMap contributors, ODbL) and Wikidata, enriched and verified by HTTP fetch. Includes codebook, Gebru et al. datasheet, Croissant ML metadata, per-country quality matrix, and SHA-256 manifest. Git tag: v1.0.0-dataset.