Do I need to pay for access?

We provide full access to Altmetric data, in three formats (database snapshot in JSON format, API, and Explorer), free of charge, for a limited time to scientometric researchers who are using it for non-commercial, ongoing research projects with a view to published results. Such use is agreed to by accepting the terms and conditions we set out in the Research Data Access Program application form. Please note access to this program is not granted for non research project work such as tracking you own research but this can be obtained with our free bookmarklet.

In limited cases where unique data must be pulled manually from our database, we provide that service at-cost. Please contact the Altmetric team for more information.

Commercial use of Altmetric data is subject to cost. Please contact the Altmetric team for more information.

How should I acknowledge Altmetric for sharing its data?

Please include credit to Altmetric in the acknowledgements section of your manuscript, presentation, or other research publication using a version of the following language:

The authors wish to thank Altmetric for providing this study’s data free of charge for research purposes.

If you are allowed to cite data in the references section of your article or book, here’s a suggested citation:

Altmetric. (Year of dataset). Altmetric database in JSON format. Unpublished raw data.

Note that journals and disciplinary norms dictate varying data citation standards. As such, you may wish to adjust the above suggested citation accordingly.

How should I describe Altmetric’s data in my research articles, presentations, etc?

Follow our Data Documentation Guide when describing the qualities of Altmetric’s data.

Can I share the portion of Altmetric data that I’ve used in my research when publishing a paper?

That depends. We consider data sharing requests on a case-by-case basis, due to the proprietary nature of some of the data. Researchers must receive written approval before releasing any data for purposes other than private peer review. Please contact [email protected] for more information.

What formats does this data come in?

Altmetric makes its data available to researchers in two ways.

  1. Access to Altmetric Details Page - Counts Only API, in JSON format.

  2. Access to the Altmetric Explorer web application, where data is exportable in CSV format.

  3. Access to the database snapshot (updated annually) in JSON format.

Is there documentation for this data? What does [this JSON field] mean?

Most of our JSON fields are documented on the Altmetric API site. The full “counts only” API call is also documented in detail in this spreadsheet (please contact Support for permission to access).  

JSON-related questions are best answered by searching Q&A site StackOverflow.

How can I get access to Altmetric’s data for my team or collaborators?

All researchers who wish to access Altmetric data must complete the Altmetric Research Data Access Program application prior to using our data. Failure to complete an application prior to giving others access to our data may result in the revocation of your data access privileges.

Full information about the RDAP can be found here:

When did you begin tracking attention for [data source]?

See this Altmetric Support article.

How do you classify X users into “scientist”, “science communicator”, and other roles?

See this Altmetric Support article.

Can you share exactly how [particular data type] is collected?

Check out our existing documentation on the Altmetric Support site.

Which identifiers does Altmetric track? Does Altmetric only track DOIs?

This Altmetric Support article contains a list of the identifiers we track--much more than just DOIs. However, many researchers use databases like Web of Science to generate large lists of articles and their related DOIs and PMIDs, which are then used to query the Altmetric Explorer or Altmetric Details Page API.

Does Altmetric’s data include the full-text of X Posts?

It does not (due to third-party data use restrictions), but we do return posts IDs for all X posts mentioning a research output. You can then use those post IDs to query the X API for the posts’ full text. More information on X data can be found here.

Can I search Altmetric for news articles that include a certain keyword or topic that also mention research outputs?

No. The Altmetric team can run these kinds of searches for you on our back end, but they are resource intensive and take time. For that reason, we charge a fee to run these searches. Contact [email protected] to request a quote.

How can I get more detailed information about [data source] (e.g. X user handles, Wikipedia edit dates, etc) using the Altmetric Details Page API?

That information isn’t available via the Altmetric Details Page - Counts Only API, but we do make some of it available in Altmetric Explorer.

If you need this data, you can contact us (by emailing [email protected]) to request Altmetric Explorer access (useful if you want to retrieve detailed mention data for under 25,000 outputs).

In the case of different identifiers for the same scholarly object, does track and aggregate all of these identifiers?

Yes, in theory. If an article has attention relating to its PMID, and other attention relating to its DOI, we would combine them. Same with arXiv IDs, Handles, and other scholarly IDs that we track.


But in practice, we need to know the link between the identifiers: so the Pubmed record has to have a DOI field, or the arXiv record needs to be linked to a DOI somehow (we get a dump of DOI -> arXiv ID links each month from ADS). For Handles, the link between Handles and DOIs normally comes from the institutional repository record.


For some pairs (like PMIDs and DOIs) the quality of mapping is very good, for some (arXiv to DOI, ISBN to DOI) it's OK, for others (Handles to DOI, Handles to PMID) it's poor.


For most sources, we get the attention first, then figure out what paper it is and combine it later. Normally we try wherever possible to capture everything and filter out anything non-scholarly (without preferring any one ID over another)


Mendeley is different because it's the only source we have to query specifically. We make one call to Mendeley per identifier for the articles we know about in this order: DOI, PMID, arXiv ID, stopping if any return a result.

How "real time" are the counts from Mendeley, X, Facebook, and Wikipedia? Are they captured at the exact time these activities occurred?

It varies. We try to display our data in as “real time” a manner as possible. 


The "update frequency" column in the table at the bottom of this page that might help.


Some of the "real time" feeds are real time from our suppliers - e.g. for news we use a third party source of news, and there may be a delay between an article being published and the source picking it up.