croissant.json

{
  "@context": "https://w3id.org/croissant/v1",
  "@type": "Dataset",

  "name": "NDL Core Corpus",
  "version": "v0.1",

  "description": "The NDL Core Corpus is an experimental, AI-ready aggregation of UK public sector data, developed as a minimum viable prototype (MVP) for the proposed National Data Library (NDL). The dataset is intended to support AI experimentation using UK public sector data, knowledge-base construction for AI agents, retrieval-augmented generation (RAG) pipelines, policy research and evaluation, and prototyping data infrastructure aligned with the Open Data Institute’s Data and AI programme.",

  "creator": {
    "@type": "Organization",
    "name": "The Open Data Institute"
  },

  "publisher": {
    "@type": "Organization",
    "name": "The Open Data Institute"
  },

  "license": {
    "@type": "CreativeWork",
    "name": "Open Government Licence v3.0",
    "url": "https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/"
  },

  "conditionsOfAccess": "Open access. Individual records reference their original public-sector licences in a dedicated license field.",

  "inLanguage": ["en", "cy"],

  "keywords": [
    "national data library",
    "public sector data",
    "ai-ready data",
    "retrieval augmented generation",
    "knowledge base",
    "agentic ai",
    "uk government data",
    "open data"
  ],

  "spatialCoverage": {
    "@type": "Place",
    "name": "United Kingdom"
  },

  "temporalCoverage": "2025-01-01/2025-12-31",

  "distribution": [
    {
      "@type": "FileObject",
      "name": "ndl_core_aggregate.parquet",
      "description": "Full aggregated corpus combining text and structured data from multiple UK public sector sources, standardised and enriched for AI use.",
      "contentUrl": "ndl_core_aggregate.parquet",
      "encodingFormat": "application/parquet"
    }
  ],

  "conformsTo": [
    "https://www.iso.org/iso-8601-date-and-time-format.html",
    "https://w3id.org/croissant"
  ],

  "measurementTechnique": [
    "Web crawling",
    "API ingestion",
    "Text extraction",
    "Data standardisation",
    "Deduplication",
    "Language identification",
    "Semantic classification using EU Data Theme Vocabulary"
  ],

  "isBasedOn": [
    {
      "@type": "CreativeWork",
      "name": "GOV.UK",
      "url": "https://www.gov.uk"
    },
    {
      "@type": "CreativeWork",
      "name": "Hansard",
      "url": "https://hansard.parliament.uk"
    },
    {
      "@type": "CreativeWork",
      "name": "legislation.gov.uk",
      "url": "https://www.legislation.gov.uk"
    },
    {
      "@type": "CreativeWork",
      "name": "data.gov.uk",
      "url": "https://www.data.gov.uk"
    },
    {
      "@type": "CreativeWork",
      "name": "Office for National Statistics",
      "url": "https://www.ons.gov.uk"
    },
    {
      "@type": "CreativeWork",
      "name": "Department for Environment, Food & Rural Affairs",
      "url": "https://environment.data.gov.uk"
    }
  ],

  "variableMeasured": [
    {
      "@type": "PropertyValue",
      "name": "identifier",
      "description": "Unique text identifier (UUID) for each record."
    },
    {
      "@type": "PropertyValue",
      "name": "title",
      "description": "Title of the resource or source filename."
    },
    {
      "@type": "PropertyValue",
      "name": "description",
      "description": "Description or summary of the resource."
    },
    {
      "@type": "PropertyValue",
      "name": "source",
      "description": "Originating source from which the data was crawled or ingested."
    },
    {
      "@type": "PropertyValue",
      "name": "date",
      "description": "Date of creation or publication of the resource."
    },
    {
      "@type": "PropertyValue",
      "name": "collection_time",
      "description": "Timestamp indicating when the data was crawled or ingested."
    },
    {
      "@type": "PropertyValue",
      "name": "open_type",
      "description": "Collection grouping such as Open Government or Open Source."
    },
    {
      "@type": "PropertyValue",
      "name": "license",
      "description": "Licence governing reuse of the individual record."
    },
    {
      "@type": "PropertyValue",
      "name": "tags",
      "description": "Automatically assigned tags using the EU Data Theme Vocabulary."
    },
    {
      "@type": "PropertyValue",
      "name": "language",
      "description": "Automatically identified language of the content."
    },
    {
      "@type": "PropertyValue",
      "name": "format",
      "description": "Data format of the record (e.g. text, parquet)."
    },
    {
      "@type": "PropertyValue",
      "name": "text",
      "description": "Full textual content of the resource, where applicable."
    },
    {
      "@type": "PropertyValue",
      "name": "word_count",
      "description": "Number of space-delimited words in the text."
    },
    {
      "@type": "PropertyValue",
      "name": "token_count",
      "description": "Number of tokens computed for machine learning use."
    },
    {
      "@type": "PropertyValue",
      "name": "data_file",
      "description": "Relative path to an associated structured data file, if present."
    },
    {
      "@type": "PropertyValue",
      "name": "extra_metadata",
      "description": "Sparse, source-specific metadata fields preserved from original datasets."
    }
  ],

  "intendedUse": [
    "Machine learning training",
    "Fine-tuning language models",
    "Retrieval-augmented generation",
    "Knowledge-base construction",
    "Policy research and evaluation",
    "AI agent experimentation"
  ],

  "limitations": "This dataset represents a static snapshot of public sector data collected during 2025. It is not real-time and coverage varies by source. Semantic harmonisation is necessarily partial due to source diversity.",

  "updateFrequency": "Static snapshot",

  "identifier": {
    "@type": "PropertyValue",
    "propertyID": "UUID",
    "description": "Each record is identified using a universally unique identifier."
  }
}