Archive of Our Own ^beta

AO3 Ship Stats Dataset

Last updated: January 2026

This dataset includes all of the data collected over the course of the AO3 Ship Stats project. This project has been collecting data about the most-used relationship tags on AO3 regularly since 2013.

Contents

The data consists of the following categories:

• The full pairing tag
• The most-used fandom for that pairing tag, as of the most recent time it appeared in the dataset
• The most-used category for that pairing tag, as of the most recent time it appeared in the dataset
• The number of public works recorded in that pairing tag for each AO3 Ship Stats dataset, listed by date

Notes

If you want to change how this spreadsheet is sorted or filtered, you can do so by selecting the "Data" menu and choosing the "Create Filter View" option.

In order to be included in this dataset, a tag must have been canonical and had over 251+ public (unlocked) works at the time of the dataset's creation. Please note that datasets from 2013 and 2014 (marked by *) were gathered by hand, and had a higher threshold number of works. Tags which have been renamed have not been connected, and may have multiple entries in the spreadsheet.

A small number of tags which should have been included were missed from one or more datasets due to technical errors. This was a particular issue in the datasets from Jan 2021-Jan 2022, due to a methodology issue which was later corrected. Excluding this period, detectable error rates average less than 0.5%.

The data-gathering program can take multiple days to run. The date given for each dataset is the date on which the process completed. This data may not exactly match the data published in the AO3 Ship Stats lists, as that data is updated directly before publishing to ensure that it is as accurate as possible. Some cases of missing data may have been corrected during processing in the published lists, but this has not been included in this dataset.

Interpretation

The following are a few examples of situations which may occur in the data, and possible explanations for them.

A tag which previously appeared in the dataset stops appearing:

• If the number of works was close to 250, the pairing may have had works deleted or privated until it had fewer than 250 works, and thus no longer qualified to appear in the dataset.
• The pairing tag may have been renamed, and continues to appear in the dataset under an alternate name. Please check for other tags in the fandom with the same meaning, but renamed tags may be listed under a different fandom if multiple fandom tags are in use or if the fandom tag has been renamed at any point.
• (Rare) The pairing tag may have been de-canonised by wranglers due to changes in wrangling policy.

A tag which previously did not appear in the dataset appears with a large number of works:

• A new pairing which gained popularity quickly can gain thousands of works between datapoints. This can also apply to older pairings which had fewer than 250 works before a sudden large gain in popularity.
• The pairing tag previously had a different name, and has been renamed. If this is the case, both names will appear in the dataset. Please check for other tags in the fandom with the same meaning, but renamed tags may be listed under a different fandom if multiple fandom tags are in use or if the fandom tag has been renamed at any point.
• (Rare) A pairing tag which previously had use but was not canonical may have been canonised by wranglers due to changes in wrangling policy.

A tag which previously appeared in the dataset is missing for a period before reappearing:

• If the number of works was close to 250, the pairing may have had works deleted or privated until it was below the 250 work threshold, before new works were posted and it re-entered the dataset.
• The pairing may have been missing from one or more editions of the dataset due to technical errors. This is most common during the period Jan 2021-Jan 2022 due to a methodology issue during that time, but can occur at any point in the data.
• (Rare) The pairing may have been renamed temporarily before the name was changed back.

The number of works for a pairing increases significantly:

• The pairing may have suddenly gained popularity, especially due to new releases happening in canon or other major fandom events.
• A large number of back-dated works may have been posted to the archive, either by individual users or as part of the Open Doors project.
• If the number of works drops again after a single year, this may have been due to spam-posting of works which were later removed by AO3 volunteers.
• (Rare) Changes to wrangling policy may change which tags are considered synonymous or sub-tags, which can lead to significant changes in work count.

The number of works for a pairing decreases significantly:

• The pairing may have had a large number of works deleted or privated, either due to negative events happening within the fandom or due to other concerns such as AI scraping (particularly from 2022 onwards).
• If the number of works rose briefly before dropping again, this may have been due to spam-posting of works which were later removed by AO3 volunteers.
• (Rare) Changes to wrangling policy may change which tags are considered synonymous or sub-tags, which can lead to significant changes in work count.

License

The AO3 Ship Stats Dataset is licensed under CC BY-NC 4.0. This means that you are free to use and adapt this data for your own purposes provided you:

• Do not use this material for commercial purposes
• Give credit to this dataset as a source

You can cite either the spreadsheet itself or this AO3 summary post.