Direct answer: To learn how to cite a dataset, identify its creator, publication year or release date, exact title, version or data vintage, repository or publisher, and a stable identifier such as a DOI or URL. Then format those details according to your required citation style and cite the dataset itself—not only a paper that analyzes it.

Dataset citations help readers locate the evidence behind a finding. They also distinguish the specific data release you used from later versions or related research publications. The details matter whether your source is a downloadable research file, a government table, or a customized query from a data portal.

Why cite a dataset?

A dataset is a research source in its own right. Citing it tells readers which evidence you used, gives credit to its creators, and makes your analysis easier to inspect or reproduce. A paper that reports results from a dataset is a separate source: cite that paper when you discuss its interpretation or argument, and cite the dataset when you use the underlying data.

For example, imagine a student compares county-level population figures with public-service access. If the student downloads a government dataset, calculates new percentages, and discusses the results, the dataset should be identified as the source of the underlying figures. The student’s interpretation remains their own. The U.S. Census Bureau makes a similar distinction in its guidance: cite the original data source, and take responsibility for conclusions drawn from your analysis. See the Census Bureau’s guidance on citing its data and research.

What information do you need to cite a dataset?

Before formatting a reference, collect the information provided on the dataset’s repository page or in its recommended citation. Common elements include:

• Creator or responsible organization: the named researcher, research team, agency, or other group responsible for the data.
• Publication or release date: use the date associated with the version you accessed, if supplied.
• Dataset title: copy the official title accurately, including subtitles when relevant.
• Version, edition, or vintage: record the release label or time period represented by the data.
• Resource description: a style may require a bracketed label such as [Data set].
• Repository or publisher: name the archive, data bank, agency, or platform that distributes the dataset.
• Persistent identifier or location: include the dataset DOI when available; otherwise, use an appropriate stable URL.
• Access date or query details: these may be needed for frequently updated sources or a customized query, depending on the style and source.

Not every style requires every element, and their order differs. Do not add information that the source does not provide. If the repository supplies a recommended citation, treat it as a useful starting point, then check it against your assignment’s required style.

How to cite a dataset in 7 steps

1. Confirm that the source is a dataset. Check whether you used a downloadable data file, a published table, a data portal, or an article that reports results. These are related but distinct sources.
2. Open the dataset’s landing page. Prefer the repository record over a search-results page or a temporary download link. The landing page is more likely to identify the creator, release, and citation instructions.
3. Find the recommended citation. Some repositories provide a citation, export tool, or “Cite” button. The Census Bureau, for example, directs users to its citation guidance and offers a citation tool for displayed tables on data.census.gov.
4. Record the exact version you used. Save the release number, edition, vintage, or observation period. A dataset may be revised, so a broad title alone may not point to the same data you analyzed.
5. Choose the persistent identifier. Use the dataset DOI if one is assigned. DataCite’s metadata guidance treats identifiers and descriptive metadata as key to identifying and retrieving research resources. If there is no DOI, use the repository’s stable record page where possible.
6. Apply your required style. Use the applicable APA, MLA, Chicago, or discipline-specific rules. Keep the dataset title and organization name faithful to the repository record; adjust punctuation and order to match the style.
7. Connect the reference to your analysis. Add an in-text citation, note, or source note where readers encounter the data. State which table, variables, time range, or query you used if that detail helps someone understand or reproduce your work.

How do APA, MLA, and Chicago handle dataset citations?

The basic goal is the same across styles: identify the data source precisely. The citation format is not identical, however. The following examples are illustrative templates using fictional details, not citations to real datasets. Replace each sample element with information from your actual source.

Fictional APA-style example: Rivera, M. (2024). Neighborhood transit access, 2018–2023 (Version 2) [Data set]. Example Research Repository. https://doi.org/10.xxxx/example

This sample illustrates the information to collect; the DOI is deliberately a placeholder and must not be used as a real identifier. Purdue University’s APA reference guide provides a dataset example and format. For MLA, consult the MLA Style Center’s guidance on citing published data. For Chicago, first determine which citation system your assignment requires; the Chicago-Style Citation Quick Guide explains the two systems and advises students to check with their instructor.

How to cite a data table, custom query, or changing dataset

Citing a table or specific subset

If your source is a named table within a larger government product, include the table name or number and the product’s vintage when the source provides them. A data portal’s “Cite” tool may generate a suggested citation for the exact table displayed. The Census Bureau’s recommendations, for example, identify the table, full product name, vintage, URL, and access date for certain table citations.

When you create a subset from a large database, make your selection understandable. In the surrounding text or a note, name the query terms, variables, geographic area, and time range that shaped your results. MLA’s guidance on published data likewise recommends identifying query terms for customized data subsets and, where possible, making the resulting table available to readers.

Citing data that changes over time

For a revised or continuously updated dataset, identify the version or date accessed when the source and required style call for it. A DOI can help point to a persistent record, but check whether it identifies a particular release or a general dataset record. The aim is to let a reader distinguish the data you used from a later update.

Citing both a dataset and a related article

Cite both sources when you rely on both. The dataset reference credits and identifies the data; the article reference supports claims about an author’s analysis, methods, or interpretation. One source does not automatically replace the other.

Common dataset citation mistakes

• Citing only a paper that used the data: add the dataset citation when your own analysis uses the underlying data.
• Leaving out the version or vintage: include it when provided and relevant to identifying the release you analyzed.
• Using a search-results or temporary download link: look for a stable repository record or DOI.
• Treating every data portal page as the dataset: identify the specific table, file, or dataset behind the interface.
• Omitting your query details: explain how you selected a subset when that choice affects what readers can interpret or reproduce.
• Copying a generated citation without checking it: verify the creator, title, date, version, identifier, and required style.
• Applying one style’s rules to another: citation systems differ, so follow the style specified for the assignment.

Dataset citation checklist

Before submitting your paper, confirm each relevant item:

• I cited the dataset itself when I used its data.
• The creator or responsible organization is correctly named.
• The dataset title and release date match the repository record.
• I recorded the version, vintage, or time range when available.
• I used a DOI or stable landing-page URL where possible.
• I identified a particular table, file, or custom query where needed.
• The in-text citation, note, or source note points to the full reference.
• The reference follows the style required by my instructor or publisher.

These checks are especially useful when your paper includes calculations, visualizations, or tables derived from external data. If you are also reviewing citations elsewhere in your paper, OakWrites has guidance on citing a government report and citing a source with no author.

Frequently asked questions

Do I cite the dataset or the article that analyzed it?

Cite the dataset when you use its data, and cite the article when you rely on its analysis or claims. Cite both if your discussion draws on both sources.

How do I cite a dataset with no DOI?

Use the dataset’s stable repository or publisher URL if available. Include the creator, title, date, and version details supplied by the source, then format them according to the required citation style.

Do I need to cite a government data table?

Yes, if you use its figures or information. Identify the agency and the specific table or product, and include the vintage or access details required by the applicable style or agency guidance.

Should I include the access date for a dataset?

Include an access date when your citation style, repository, or the changing nature of the data calls for one. For a stable, versioned record, the release or version may be more useful; follow the relevant instructions.

How do I cite a dataset in APA?

APA dataset references generally identify the creator, year, dataset title, version if available, a [Data set] description, publisher, and DOI or URL. Check the current APA guidance and your assignment requirements for the exact formatting.

Key takeaways

• Cite the data source itself, not just a publication that discusses it.
• Record the creator, title, date, version or vintage, repository, and DOI or stable URL.
• For tables and custom queries, identify the specific data you selected.
• Follow the required citation style and verify repository-generated references before submitting.

Accurate referencing makes the evidence behind your argument easier to identify and check. If you want help checking citation formatting or document consistency, explore OakWrites editing and document-support services. You can also review OakWrites pricing or submit a document for a personalized quote.

Sources

• Purdue OWL: Reference List—Electronic Sources
• MLA Style Center: How to cite raw data and published data
• The Chicago Manual of Style: Citation Quick Guide
• U.S. Census Bureau: Citing data, tools, and research
• DataCite: Metadata Schema