Sophymarine/wikidata-alignment
收藏资源简介：
该数据集是一个预计算的映射表，用于将Wikidata中的公司条目（QID）与它们在OpenRegistry平台上的规范URL进行对齐。它基于Wikidata中至少包含一个受支持的国家公司注册标识符（如法国的SIREN、英国的Companies House编号等）的条目，通过查询Wikidata查询服务（截至2026年5月24日）整合了17个国家公司标识符属性的结果而构建。数据集提供每一行的Wikidata QID、国家代码、注册机构属性、原始注册标识符、OpenRegistry URL以及英文和本地语言标签（可能为空）。其主要用途包括实体链接（将Wikidata公司QID解析为实时注册记录）、维基百科/Wikidata参考丰富、KYB（了解你的业务）和供应链尽职调查管道、知识图谱回填用于AI代理，以及覆盖分析（分析各司法管辖区社区验证的公司条目数量）。数据集覆盖17个国家，总计658,297行，其中法国（SIREN）数据最多（462,518行）。OpenRegistry URL是稳定的规范地址，可通过追加扩展名（如.md、.jsonld、.ttl）获取不同格式的记录。需要注意的是，该数据集并非完整的公司列表（仅包含Wikidata社区认为值得注意的实体），不包含OpenRegistry的底层数据，不包含仅基于OpenCorporates的条目，且是快照数据而非实时馈送。数据集采用CC0 1.0 Universal许可证发布。
This dataset is a precomputed mapping that aligns Wikidata items for companies (QIDs) with their canonical OpenRegistry URLs. It is built from Wikidata items that have at least one supported national company-registry identifier statement (e.g., SIREN for France, Companies House number for the UK), by querying the Wikidata Query Service (as of 2026-05-24) and combining results from 17 national company-identifier properties. For each row, it provides the Wikidata QID, country code, registry property, native registry identifier, OpenRegistry URL, and English and local labels (which may be empty). Its primary uses include entity-linking (resolving a Wikidata company QID to a live registry record), Wikipedia/Wikidata reference enrichment, KYB (Know Your Business) and supply-chain due-diligence pipelines, knowledge-graph backfills for AI agents, and coverage analysis (assessing the number of community-validated company items per jurisdiction). The dataset covers 17 countries with a total of 658,297 rows, with France (SIREN) having the most rows (462,518). The OpenRegistry URLs are stable canonical addresses that can return records in HTML, Markdown, JSON-LD, or Turtle formats by appending extensions (e.g., .md, .jsonld, .ttl). Note that this dataset is not a complete list of companies (only entities deemed notable by the Wikidata community are included), does not contain OpenRegistrys underlying data, excludes items with only OpenCorporates statements, and is a snapshot rather than a real-time feed. It is published under the CC0 1.0 Universal license.
