Home > Blog > Citing Datasets and APIs in Academic Papers: APA, MLA, Chicago 2026 Guide

Citing Datasets and APIs in Academic Papers: APA, MLA, Chicago 2026 Guide

Key Takeaways

  • Datasets and APIs are increasingly common research sources, but citation rules vary by style guide
  • APA 7th Edition treats datasets as data sources and code/APIs as software — each with distinct formatting
  • MLA 9th Edition doesn’t explicitly cover datasets; you must adapt the standard template
  • Chicago Manual of Style lacks specific dataset/API rules — adapt author-date or notes formats
  • The version number and access date matter more than you’d expect — they’re your reproducibility anchor

What Counts as a “Dataset” or “API” in Academic Writing?

Before you can cite something correctly, you need to know what you’re working with. Not every digital source fits neatly into traditional citation categories.

Datasets are structured collections of raw data — think numerical records, survey responses, sensor readings, or genomic sequences. They’re typically hosted on dedicated repositories (Kaggle, ICPSR, Zenodo, Harvard Dataverse) or published by government agencies (Census Bureau, World Bank, NOAA). Unlike a book or article, a dataset’s primary purpose isn’t argument or analysis — it’s raw information designed to be queried, filtered, and re-analyzed.

APIs (Application Programming Interfaces) are the plumbing that lets programs talk to each other. When you use Python to pull Twitter data via the Twitter API, or access the PubMed API to search medical literature, you’re interacting with an API. The API itself isn’t the content — it’s the tool you used to access the content.

Code libraries and software repositories (like those hosted on GitHub) blur the line between datasets and APIs. A repository might contain both code and data. The key question is: what are you using? If you’re pulling structured data from a repository, it’s a dataset. If you’re writing code that interacts with an API endpoint, it’s a software/API citation.

Here’s the core distinction that trips up students:

Dataset API
What it is A static collection of data A live interface that returns data
Analogy A library of raw numbers A doorbell that summons data
Citation focus Author, title, repository, version Developer, documentation URL, API name, version
Style category Treated as “Data set” or “Software” Treated as “Software” or “Webpage”
Access date Recommended if data changes Required (APIs update frequently)

If you’re not sure which category your source falls into, ask yourself: am I citing the data itself (a dataset) or the tool/interface I used to retrieve data (an API)? Your answer determines the citation format.

APA 7th Edition: Citing Datasets and APIs

APA (American Psychological Association) is the most widely used citation style in social sciences, education, and health research. The 7th Edition (2020) introduced specific guidance for datasets and software, making it the go-to style for data-driven research.

How to Cite a Dataset in APA

The APA format follows this structure:

Creator(s). (Year). Title of dataset (Version) [Data set]. Repository. DOI or URL

Notice three elements that often get missed:

  1. The version number — Even if your dataset has no “version,” APA still wants one. Use “No version” if you’re unsure.
  2. The bracketed format descriptor — [Data set] signals to readers and reviewers that you’re citing raw data, not a traditional publication.
  3. The repository name — This isn’t optional. If the dataset lives on Kaggle, ICPSR, or Zenodo, the repository name is the publisher equivalent.

Example — Dataset from a public repository:

Smith, J. K., & Johnson, M. L. (2024). National health and nutrition examination survey (Version 3.2) [Data set]. National Center for Health Statistics. https://www.cdc.gov/nchs/nhanes

Example — Dataset with DOI:

Garcia, M. R., Lopez, J. A., & Rodriguez, A. C. (2023). Longitudinal educational achievement data (Version 5.0) [Data set]. Harvard Dataverse. https://doi.org/10.7910/DVN/EXAMPLE

Example — Government dataset:

U.S. Census Bureau. (2024). American Community Survey 5-year estimates [Data set]. https://www.census.gov/acs

How to Cite an API or Software in APA

APIs and code libraries fall under the “software” category. The format is:

Developer(s). (Year). Software name (Version) [Computer software]. Repository or URL

Example — GitHub repository:

Mozilla Foundation. (2022). Pdf.js: PDF reader and viewer in JavaScript (Version 3.4.120) [Computer software]. GitHub. https://github.com/mozilla/pdf.js

Example — API with documentation:

Halpern, J. A. (2023). Slow-render: A static site generator for long-form essays (Version 2.4.0) [Computer software]. GitHub. https://github.com/jhalpern/slow-render

Example — API accessed through code:

Wickham, H. (2024). Tidyverse: A collection of R packages for data science (Version 2.0.0) [Computer software]. https://www.tidyverse.org/

In-Text Citation for Datasets and APIs in APA

APA uses parenthetical author-date format:

  • Dataset paraphrase: (Smith & Johnson, 2024)
  • Direct data reference: (Smith & Johnson, 2024, Table 3)
  • Organization as author: (U.S. Census Bureau, 2024)
  • Software/API paraphrase: (Wickham, 2024)
  • Commit-specific reference: (Mwangi, 2024, commit a8f3e91)

Common APA Mistakes (and How to Fix Them)

Mistake Why It’s Wrong The Fix
Missing version number APA requires versioning for reproducibility Always include “(Version X.X.X)” or “(No version)”
Forgetting [Data set] or [Computer software] These descriptors tell readers what type of source they’re looking at Add the bracketed format descriptor after the title
Omitting the repository The repository is the publisher equivalent for datasets Include the full repository name
Using “Retrieved from” before URL APA 7 removed this prefix for online sources Delete “Retrieved from” — just include the URL
Using the creation year instead of version year You’re citing a specific version, not the original publication Use the year of the version you accessed

MLA Style: Citing Data Sources and Code Libraries

MLA (Modern Language Association) 9th Edition doesn’t have explicit guidelines for datasets or APIs. This is one of the biggest gaps in student-facing resources — and it’s why so many students cite datasets incorrectly in humanities and social science papers.

The key principle for MLA: adapt the closest available template. Since MLA doesn’t specify data or software categories, you use the standard “web source” or “book” template with discipline-specific additions.

How to Cite a Dataset in MLA

Author(s). Title of Data Set. Publisher/Repository, Publication date, URL or DOI.

Example — Dataset from a repository:

Bureau of Transportation Statistics. 2020 National Census of Ferry Operators. United States Department of Transportation, 01 March 2022. www.bts.gov/NCFO

Example — Dataset with DOI:

Smith, John K., and Jane L. Doe. National Health and Nutrition Survey. National Center for Health Statistics, 2024. doi.org/10.5281/zenodo.1234567.

Example — Academic repository dataset:

Garcia, Maria R., et al. Longitudinal Educational Achievement Data. Harvard Dataverse, 2023. doi.org/10.7910/DVN/EXAMPLE.

How to Cite an API or Code Library in MLA

Developer. Software name. Version, Year. Platform/Repository. URL.

Example — GitHub repository:

Mozilla Foundation. Pdf.js: PDF reader and viewer in JavaScript. Version 3.4.120, 2022. GitHub. https://github.com/mozilla/pdf.js.

Example — API accessed through code:

Wickham, Hadley. Tidyverse: A Collection of R Packages for Data Science. Version 2.0.0, 2024. https://www.tidyverse.org/.

MLA in-Text Citation

MLA uses author-page format, but datasets and APIs complicate this:

  • Author as corporate body: (U.S. Census Bureau)
  • Author as individual: (Smith 15) — though page numbers don’t apply well to datasets
  • No page numbers: Many students skip in-text citation entirely when citing datasets. This is risky. At minimum, cite the author and year: (Smith, 2024).

What MLA Lacks (and How to Compensate)

MLA doesn’t specify version fields, access dates, or format descriptors for data sources. Here’s how to compensate:

  1. Version: If available, include it after the title (like you would with a book edition)
  2. Access date: MLA recommends “accessed” dates for online sources that may change. Always include “Accessed Day Month Year.”
  3. Repository: If no publisher exists, the repository name fills the publisher slot.

Chicago Style: Citing Data and Software

Chicago Manual of Style (17th Edition) is popular in history, humanities, and some social sciences. The style’s two systems — Notes & Bibliography and Author-Date — require different approaches for datasets and APIs.

Chicago Notes & Bibliography Format

Author(s). Title of Dataset. Version (Place: Publisher, Year). URL or DOI.

Example — Dataset:

Smith, John K., and Jane L. Doe. National Health and Nutrition Survey. Version 3.2 (Washington DC: National Center for Health Statistics, 2024). https://www.cdc.gov/nchs/nhanes.

Example — GitHub repository:

Halpern, James A. Slow-render: A Static Site Generator for Long-form Essays. Version 2.4.0 (n.p.: GitHub, 2023). https://github.com/jhalpern/slow-render.

Chicago Author-Date Format

Author. Title of dataset. Version. Place: Publisher, Year. URL.

Example — Dataset:

Smith, John K., and Jane L. Doe. National Health and Nutrition Survey. Version 3.2. Washington DC: National Center for Health Statistics, 2024. https://www.cdc.gov/nchs/nhanes.

Example — API:

Wickham, Hadley. Tidyverse: A Collection of R Packages for Data Science. Version 2.0.0. n.p., 2024. https://www.tidyverse.org/.

Chicago Short-Note Format

For footnotes or endnotes (the first citation of a source):

  1. John K. Smith and Jane L. Doe, National Health and Nutrition Survey, Version 3.2 (Washington DC: National Center for Health Statistics, 2024), https://www.cdc.gov/nchs/nhanes.

Subsequent citations: Smith and Doe, National Health and Nutrition Survey.

What Chicago Lacks

The Chicago Manual of Style (17th Edition) doesn’t have explicit rules for datasets or APIs. The guidance above adapts existing book and online-source templates. When in doubt, follow your professor’s preference — some instructors accept APA-style citations even in Chicago-style papers for data sources.

Step-by-Step: How to Cite a Dataset from a Repository

Now let’s walk through the practical steps. Here’s exactly what you need to do, dataset by dataset:

Step 1: Identify the Dataset’s Core Elements

Every dataset citation needs:

  • Author/Creator — Individual name or organization
  • Year — Publication or release year of the version you used
  • Title — Dataset name (italicized)
  • Version — Version number (if available)
  • Repository — Where the data is hosted
  • DOI or URL — Permanent link or direct link

Step 2: Match the Format to Your Citation Style

Citation Style Format Category Key Identifier
APA 7th Data set or Computer software [Data set] descriptor required
MLA 9th Web source or Book (adapted) Access date + “accessed” prefix
Chicago Online resource or Book Bibliography/notes system choice matters

Step 3: Write the Citation — Real Examples for Every Repository

Kaggle (data science community):

Kaggle. (2024). World happiness report dataset (Version 4.0) [Data set]. Kaggle. https://kaggle.com/unsup/happiness

GitHub (code and data repositories):

Halpern, J. A. (2023). Slow-render: A static site generator for long-form essays (Version 2.4.0) [Computer software]. GitHub. https://github.com/jhalpern/slow-render

ICPSR (social science data archive):

Inter-university Consortium for Political and Social Research. (2023). American National Election Study (ICPSR 37337) [Data set]. University of Michigan. https://doi.org/10.3886/ICPSR35157.v4

Harvard Dataverse (academic research repository):

Garcia, M. R., Lopez, J. A., & Rodriguez, A. C. (2023). Longitudinal educational achievement data (Version 5.0) [Data set]. Harvard Dataverse. https://doi.org/10.7910/DVN/EXAMPLE

Data.gov (U.S. government data):

U.S. Census Bureau. (2024). American Community Survey 5-year estimates [Data set]. U.S. Census Bureau. https://www.census.gov/acs

World Bank Open Data (economic and development data):

World Bank. (2024). World development indicators [Data set]. World Bank Open Data. https://data.worldbank.org/

NASA Earth Data (environmental and geospatial data):

NASA. (2024). Satellite Earth observation data [Data set]. NASA Earth Data Repository. https://earthdata.nasa.gov/

Step 4: Check Your Reference List Formatting

  • APA: Alphabetize by creator’s last name or organization name. Use hanging indent.
  • MLA: Alphabetize by author’s last name. No hanging indent required (but many instructors prefer it).
  • Chicago: Alphabetize by author/organization. Notes in footnotes/endnotes are numbered, not alphabetized.

Step-by-Step: How to Cite an API or Code Library

APIs and code libraries require a slightly different approach. Here’s the process:

Step 1: Gather the Required Information

For an API or code library, you need:

  • Developer/Author — Individual or organization that created the API
  • API/Library Name — The name as it appears in official documentation
  • Version — The specific version you used (or the current version if you’re unsure)
  • Year — Year of the version or current year if undated
  • Documentation URL — The link to official API documentation (not just the homepage)
  • Access Date — Highly recommended for APIs (they change frequently)

Step 2: Format the Citation

APA 7th Edition:

Wickham, H. (2024). Tidyverse: A collection of R packages for data science (Version 2.0.0) [Computer software]. https://www.tidyverse.org/

MLA 9th Edition:

Wickham, Hadley. Tidyverse: A Collection of R Packages for Data Science. Version 2.0.0, 2024. https://www.tidyverse.org/. Accessed 15 June 2026.

Chicago Notes & Bibliography:

Wickham, Hadley. Tidyverse: A Collection of R Packages for Data Science. Version 2.0.0. n.p., 2024. https://www.tidyverse.org/.

Step 3: Know When to Cite vs. Mention

This is where most students make critical errors:

Cite: You’re using a specific dataset, repository, or library as part of your research methodology. The API or dataset is a source you’re analyzing.

Mention: You’re using an API as a tool (like Google Maps API for plotting data in a figure) or a programming language/framework (Python, R, TensorFlow) for analysis. In this case, mention the tool in your methods section — don’t add it to your reference list.

Rule of thumb: If the API or dataset is part of your data, cite it. If it’s part of your methods (the tool you used to process data), mention it in text only.

Common Mistakes in Data Citation

Even experienced researchers make mistakes with dataset and API citations. Here are the most common errors students encounter:

1. Skipping the Version Number

Datasets and APIs change. A dataset cited without a version number is essentially a citation to a moving target. APA 7th explicitly requires the version field. MLA and Chicago don’t specify it, but including it demonstrates academic rigor.

Wrong:

Smith, J. (2024). National health data [Data set]. National Center for Health Statistics.

Right:

Smith, J. (2024). National health data (Version 3.2) [Data set]. National Center for Health Statistics.

2. Using “Retrieved from” Before the URL

APA 7th removed the “Retrieved from” prefix for most online sources. It’s only required when the source is unlikely to be updated or archived (like a social media profile). For datasets and APIs, the URL alone is sufficient.

Wrong:

Smith, J. (2024). National health data [Data set]. National Center for Health Statistics. Retrieved from https://www.cdc.gov/nchs

Right:

Smith, J. (2024). National health data (Version 3.2) [Data set]. National Center for Health Statistics. https://www.cdc.gov/nchs

3. Citing the Wrong Year

The year in a dataset citation should reflect the version you used — not necessarily the year the dataset was originally published. If you used Version 3.2 released in 2024, cite 2024 even if the original dataset was published in 2020.

4. Forgetting the Repository Name

The repository is the publisher equivalent for datasets. Omitting it means your reader can’t find your source. Every dataset citation needs the repository name (Kaggle, ICPSR, Zenodo, Dataverse, etc.).

5. Using the General API Homepage Instead of Documentation

When citing an API, the URL should point to the documentation page for the specific version you used — not the product’s general homepage. This ensures readers can verify your citation.

Wrong:

https://developers.google.com/youtube

Right:

https://developers.google.com/youtube/v3/

When to Cite vs. When to Mention

Not every digital source in your paper needs a reference entry. Here’s the framework:

Cite (reference list entry required):

  • The raw dataset itself (survey data, sensor readings, genomic sequences)
  • A repository hosting structured data (Kaggle, ICPSR, World Bank)
  • A code library you’re analyzing or comparing
  • An API where you extracted data for analysis

Mention (text only, no reference entry):

  • Programming languages (Python, R, MATLAB)
  • General-purpose analysis tools (SPSS, Stata)
  • APIs used as visualization tools (Google Maps API for a map in your figure)
  • Software used to format or process data (Excel, LibreOffice)
  • Web browsers used to access sources (Chrome, Firefox)

The logic is simple: If the source is your data, cite it. If the source is your tool, mention it.

What Your Professor Should Know About Dataset Citations

Here are three things most professors care about that you should proactively demonstrate:

1. You Can Point to the Exact Version Used

If your professor asks which dataset version you relied on, you should have the version number or commit hash from your citation. This isn’t just academic etiquette — it’s what separates rigorous research from casual data gathering.

2. You Understand Reproducibility

Including version numbers and access dates isn’t pedantry. It’s what allows other researchers to verify your findings using the same data state. If you’re using a dataset that updates regularly, the access date is your reproducibility anchor.

3. You Know the Difference Between Data and Tools

Citing an API as your data source is wrong. Citing Python as your data source is wrong. These are tools — they go in your methods section, not your bibliography. Professors spot this mistake instantly.

Final Checklist: Before You Submit

  1. Identify your source type: Dataset or API/code library?
  2. Pick your citation style: APA, MLA, or Chicago?
  3. Match the format: Use the correct template from this guide
  4. Include the version: Every dataset and API citation needs a version field
  5. Include the repository or documentation URL: Can your reader find your source?
  6. Format the bracketed descriptor: [Data set] or [Computer software]
  7. Check your year: Use the version year, not the creation year
  8. Verify alphabetization: Reference lists should be sorted alphabetically by creator
  9. Use hanging indent: APA and Chicago require it; MLA recommends it
  10. Double-check: Compare your citation against the examples in this guide

Summary

Citing datasets and APIs in academic papers isn’t complicated once you understand the format rules. The key takeaways:

  • APA 7th Edition requires version numbers and bracketed format descriptors ([Data set] or [Computer software])
  • MLA 9th Edition doesn’t have explicit rules — adapt the web source or book template and always include an access date
  • Chicago works with both notes and author-date systems; the main difference is formatting, not content
  • Version numbers are non-negotiable for reproducibility — they’re your citation’s anchor
  • Know when to cite vs. mention — data sources go in references; tools go in the text

When in doubt, ask your professor for the preferred dataset citation format. If they don’t have one, use the templates in this guide. The rules are clear enough once you know where to look.

Need help formatting citations correctly? Our academic writers ensure every source is properly cited across APA, MLA, Chicago, and Harvard styles. Order now — professional formatting, plagiarism-free, delivered on time.

Related Guides

Additional Resources