Dataset is the Schema.org type for a body of structured data, such as a CSV of benchmark results or a public API's sample data. Google uses Dataset markup to index datasets in Google Dataset Search, a separate search engine for data. It does not produce a rich result in regular web search. Google requires name and description.
Useful properties
- name and description (Google requires a description of reasonable length)
- license: a URL or CreativeWork describing reuse terms
- creator and publisher
- distribution: DataDownload items with contentUrl and encodingFormat
- temporalCoverage and spatialCoverage where relevant
- isAccessibleForFree, keywords and version
{
"@context": "https://schema.org",
"@type": "Dataset",
"name": "Fernhallow open benchmark: API latency by region",
"description": "Monthly median API latency measured from six regions.",
"license": "https://creativecommons.org/licenses/by/4.0/",
"distribution": [{ "@type": "DataDownload",
"encodingFormat": "text/csv",
"contentUrl": "https://data.fernhallow.example/latency.csv" }]
}Why SaaS companies use it
Companies that publish original data, for research reports or as a link-earning asset, can make the underlying files discoverable to researchers and journalists who use Dataset Search. Clear licensing and attribution terms make citation easier. Dataset Search traffic is specialized, so treat it as a discovery channel for data users, not a general traffic source.
Example
Open benchmark data
Fernhallow, an API monitoring service, publishes a monthly latency dataset. The landing page shows a summary, a methodology section and download links, with Dataset markup listing the CSV distribution and a CC BY license. The page is later found and cited in a university networking course.
Common mistakes
Marking up a report as a dataset
A PDF of findings is not a dataset. Dataset markup should describe actual data files or APIs.
Questions
Google documents Dataset markup for Dataset Search. Do not expect a rich result in the main results.
