Sophymarine/openregistry-snapshots

OpenRegistry Temporal Corporate Data 是一个纵向、时间点公共公司注册状态语料库，由Sophymarine提供。每行数据代表一个公司在特定时间点的状态，带有_retrieved_at时间戳，每周快照累积成时间序列，以便跟踪六个司法管辖区内公司状态的变化（如状态翻转、地址迁移、名称更改、结构转换）。数据来自六个国家注册机构（挪威、法国、澳大利亚、爱尔兰、捷克、新西兰），通过OpenRegistry MCP服务器标准化，提供实时、主要来源的访问。数据集包含公司名称、状态、注册地址、管辖权数据等字段，适用于文本分类和文本检索任务，支持英语、挪威语、法语和捷克语。数据集规模约为30万行每周快照，以Parquet格式提供，并带有时间来源元数据。它适用于生命周期建模、事件检测、面板研究、代理训练信号和回测数据等用例，与实时MCP服务器互补，适用于训练、评估和研究场景。OpenRegistry Temporal Corporate Data is a longitudinal, point-in-time corpus of public company-registry state provided by Sophymarine. Each row represents one company at a specific moment in time, stamped with _retrieved_at. Weekly snapshots accumulate into a time series, allowing tracking of corporate state changes—such as status flips, address moves, name changes, and structural transitions—across six jurisdictions. Data is drawn from six national registries (Norway, France, Australia, Ireland, Czechia, New Zealand) and normalized via the OpenRegistry MCP server, which provides real-time, primary-source access. The dataset includes fields like company name, status, registered address, and jurisdiction data, and is suitable for text classification and text retrieval tasks, supporting English, Norwegian, French, and Czech. It scales to approximately 300,000 rows per weekly snapshot, delivered in Parquet format with temporal provenance metadata. Use cases include lifecycle modeling, event detection, panel research, training signals for agents, and backtest data, complementing the live MCP server for training, evaluation, and research purposes.