Dimensions as preferred source of truth for Altmetric research outputs

We heard from customers using Google BigQuery (GBQ) for large-scale analysis of Dimensions datasets that they would like to replicate that experience with Altmetric data, and access Dimensions and Altmetric data at the same time. 

In 2024, we released a customer-facing GBQ offering that includes Altmetric data. As a number of research outputs available in Altmetric are also available as publications in Dimensions, this highlighted the opportunity to increase the consistency between the Dimensions publications dataset and the Altmetric research output dataset, by making Dimensions the preferred source of truth for research output metadata fields in Altmetric. 

1. What’s new?

Historically, CrossRef has been the primary source of truth for Altmetric research output metadata (with the exception of a number of books and chapters, which primarily use information from Springer Nature). Dimensions is now the preferred source of truth for a selection of research output metadata fields in Altmetric. 

As a result:

1. Wherever possible*, new research outputs in Altmetric will be populated with Dimensions metadata

2. Wherever possible*, existing research outputs in Altmetric will be populated with Dimensions metadata. As of the initial release on 26 February 2024, we have aligned 28.1M Altmetric research outputs with Dimensions.

* This alignment project primarily focused on research outputs categorised as articles in Altmetric. See Out of Scope section later in this document for more information.

2. Value and user impact

We believe that increased alignment between Altmetric and Dimensions publications will:

• Enable users of GBQ to merge Altmetric data with Dimensions data, and use apps and insights more effectively

• Reduce the volume of support queries connected to inconsistencies between Altmetric and Dimensions data.

As a result of this alignment project, users may notice changes in the metadata of some research outputs in Altmetric products (e.g. titles, author names, publication dates). These changes are a result of changing the source of the metadata to Dimensions, to increase consistency across datasets.

For any instances of incorrect publication data in Dimensions or Altmetric, please note that:

• Incorrect publication data appearing in Dimensions can be changed in CrossRef. In the majority of cases, corrections in CrossRef will then filter to Dimensions. 

• If research outputs in Altmetric display incorrect data, the Altmetric Support team will be able to check whether this has been sourced from Dimensions. If that is the case, corrections in CrossRef should resolve any issues. 

3. Out of scope

This project primarily focused on:

• Increasing alignment between the Dimensions publications dataset and the Altmetric research outputs dataset

• Aligning research outputs categorised as articles in Altmetric

Therefore, the scope of the project did not include:

• Metadata fields available in the Dimensions publications dataset, that Altmetric store outside of the research outputs datasets:

  • Journal metadata, other than ISSNs (most of the metadata available in Dimensions is stored in the Altmetric journals dataset)

  • Publisher metadata available in Dimensions

  • Author metadata, other than author names (Altmetric does not support additional attributes beyond author names at present)

• Publication types Altmetric and Dimensions currently manage with significantly different logics in order to fulfil key use cases  (books; chapters; preprints)

• Publication types that don’t yet have an equivalent across the two datasets  (Dimensions monographs and proceedings; Altmetric clinical trial study records)

We don’t have immediate plans or a timeline to address these areas. We expect that achieving full alignment between datasets will require further research from both the Altmetric and Dimensions teams, and potentially changes on a larger scale, in order to develop a robust joint approach.