Check the paper's data availability statement for a dataset DOI, then resolve that DOI through DataCite to confirm the record before you cite the dataset separately from the article. The data availability statement is the paper's own pointer to where its underlying data live, so start there before working from the figure caption.
- Read the data availability statement. Look for a DOI, repository name, or accession identifier tied to the data behind the figure. A DOI is the most direct lead because it identifies the object itself.
- Resolve the DOI. A DOI name resolves through the Handle System to current information supplied for the identified object, commonly a landing page maintained by the registrant. Follow the DOI to that landing page and confirm the title, creators, and repository match the dataset you expect.
- Confirm the record through DataCite. DataCite's REST API exposes DOI metadata and supports queries and filters for works, repositories, providers, clients, and related entities. Use the metadata to verify that the DOI describes the dataset rather than the article or an unrelated record.
- Cite the dataset separately when the journal requires it. Citing Medicine provides structured examples and rules for citing journal articles, databases, datasets, and other publication types, so use its dataset format for the data citation and keep the article citation distinct.
If the data availability statement names a repository but gives no DOI, search that repository for the dataset by title, creator, or accession number. If the DOI resolves to a landing page that does not match the figure's described data, treat the link as unresolved and check the statement again for a second identifier or contact the corresponding author before citing.