About the Dataset Schema Generator
Dataset markup is how researchers, governments, newsrooms and companies make their data discoverable in Google Dataset Search, a separate search engine for datasets. Google requires name and description. Recommended properties include creator, license, distribution (a DataDownload with encodingFormat and contentUrl for each file), spatialCoverage, temporalCoverage, keywords, identifier, version and url.
This dataset schema generator builds the JSON-LD in your browser. Write a description that explains what the data measures, how it was collected and its limits, since that is what people read in the search listing. State the licence clearly with a URL, such as a Creative Commons licence, and list every download format you offer. Put the markup on the dataset's landing page rather than on the raw file itself.
How to use it
- Enter the dataset name and a clear, detailed description.
- Add the creator, licence URL and keywords.
- List each download as a distribution with format and file URL.
- Add spatial and temporal coverage, then publish on the landing page.
Frequently asked questions
Where do datasets with this markup appear?
Dataset markup is used by Google Dataset Search, a dedicated search engine for datasets. It is not a rich result in standard web search. Researchers, students and journalists use Dataset Search to find data by topic, format, licence and update date.
What are the required properties for Dataset schema?
Google requires name and description. Everything else is recommended, but creator, license, distribution, spatialCoverage and temporalCoverage make a listing far more useful, because they help people judge quickly whether the data is relevant and whether they are allowed to reuse it.
Can I mark up a dataset that is not free?
Yes. Use isAccessibleForFree set to false and describe how to get access on the landing page. Dataset Search can still list it, and users can see the access conditions before clicking through. Make sure the landing page itself is public and crawlable, and state the licence clearly so people know how they may use the data once they have access.