Research doesn’t stand alone and relationships show the connections between research outputs, people, and organizations. We deliver these connections via a relationships API endpoint, which makes the Research Nexus visible.

Relationships show how items are related to each other. A published research article may be supported by data, software, and peer reviews; previous versions might be available as a preprint and there could be online commentary and annotations. Research is reused and cited in other research works, as well as being mentioned in online venues such as blog posts and news articles. The relationships endpoint makes all of these connections available in a single place.

We collect relationships from a number of sources:

  • Metadata deposited by our members
  • Links we discover between existing metadata records, for example, through reference matching.
  • Event Data agents (not yet implemented).
  • Trusted community partners, such as DataCite.

The relationships endpoint is currently released as a beta version, meaning that it may contain undetected data quality issues and bugs, and we will not respond immediately to performance issues. The dataset available currently includes works updated since the beginning of 2024.

How does the relationships endpoint work?

We collect metadata from our members, Event Data, and selected other sources. When we find a connection between two items (a research output, organization, or person), we represent it as a relationship with the following information:

  • Subject (source of the relationship)
  • Object (target of the relationship)
  • Relationship type

We also provide metadata about when the relationship was last updated and which organizations deposited metadata about the relationship.

How to use the relationships endpoint

You can access the results in JSON format from our REST API.

Query filters are available to look for specific items, types of items, types or relationships, time ranges, and source organizations. For a full range of queries see the beta endpoint Swagger documentation.

The relationships endpoint is optimized for specific types of queries. This means that certain combinations of query filters will return results quickly whereas others (particularly when combining more than one filter) will take much longer.

The following cases are currently supported:

Query parameter combinationsExample use cases
relationship-type, object.registration-agency, from-update-time, until-update-timeCollect citations to DOIs from a different registration agency, e.g. DataCite.
relationship-type, asserted-by, object.type, from-update-time, until-update-timeData citations deposited by a Crossref member.
relationship-type, from-update-time, until-update-timeFind all relationships of some type in some timeframe, e.g. citations, preprint to article links, members of Crossref.
object.steward, asserted-by, object.type, from-update-time, until-update-timeFind information added by Crossref or another organization to a member’s metadata records.
subject.steward, asserted-by, object.type, from-update-time, until-update-timeFind information added by Crossref or another organization to a member’s metadata records.
subject.steward, relationship-type, object.type, from-update-time, until-update-timeData citations from a Crossref member’s works (deposited by the member or another organization).

Relationships are provided via a freely available public API. This makes it suitable for post-processing and creating downstream tools such as search interfaces or dashboards. We do not curate the metadata we receive or provide plugins and data analytics.

Note that counts of references are based on our member metadata and differ from other services; see this Community Forum post for further background.

Page owner: Martyn Rittman   |   Last updated 2024-April-11