How does it work?

Tracking and collating attention

Today the Altmetric database contains over 256 million mentions of over 24 million research outputs (including journal articles, datasets, images, white papers, reports, and more), and is constantly growing.

To be able to track the online attention for a specific piece of research, there are 3 things we need:

• An output (journal article, dataset, etc.)
• An identifier attached to the output (DOI, RePEc, etc.)
• Mentions in a source we track

Once we’ve picked up a mention of the research, we collate it together with any other online attention we’ve seen for that item and display it via the Altmetric details page, along with it’s own unique donut and automatically calculated Altmetric Attention Score.

You can find more information on the sources we monitor here.

Types of identifiers we track as standard

The unique identifier is really important for us to be able to match the attention to the research – otherwise we don’t know which mentions belong to which research. Identifiers our system is designed to automatically recognize extend far beyond journal articles, and include:

• PubMedID – typically associated with health sciences research
• arXiv ID – Physics, Mathematics & Computer Sciences
• ADS ID – Astrophysics data system
• SSRN ID – Social Sciences outputs
• RePEC ID – economics research
• Handle.net identifiers – often used in institutional repositories
• URN (Uniform Resource Name) identifiers
• ISBNs – books hosted on publisher domains and google books
• DOIs – assigned to individual articles at the point of publication, and by platforms such as figshare and Dryad to other outputs (including datasets, images, and more)

One output: many versions – disambiguation

These identifiers also help us to recognize different versions of the same research output. For example, a journal article might be originally made available on a publisher platform and given a DOI, and then later hosted on PubMed or an institutional repository and given another unique identifier there.

Our system cross-checks these to match them together, ensuring that the details page always displays a collated record of attention for all versions of the research item.