PID Graph GraphQL Example Disambiguate Researchers

User story

As a researcher, I am looking for more information about another researcher with a common name, but don’t know his/her ORCID ID.

Query strategy

Sometimes we need more information than the name to find the ORCID ID of a particular researcher. This can be additional information such as the affiliation, but a list of publications associated with the researcher can also be very helpful. We therefore query for the researcher by name, and then show not only the given name, family name and affiliations, but also three publications with publication year, publication venue, title and co-authors, including their affiliation. The query in this example returns 648 records on April 13, 2020.

Why GraphQL

This is a rather common problem, and often not possible to address with a single REST API call.

This could be turned into a discovery service for finding ORCID IDs of researchers by building a simple Javascript application using this or similar GraphQL queries.

Use the following query in the GraphQL client at https://api.datacite.org/graphql

{
  people(query: "John AND Smith") {
    totalCount
    nodes {
      id
      givenName
      familyName
      name
      affiliation {
        name
      }
      works(first: 3) {
        nodes {
          id
          publicationYear
          titles {
            title
          }
          creators {
            id
            name
            affiliation {
              id
              name
            }
          }
          subjects {
            subject
          }
        }
      }
    }
  }
}

I tried this with my name and the results were a bit strange. There was a count of 2 but I could only see the first name which was Gabriel Vecchi, which seemed odd.

Very strange indeed. This comes from the ORCID API. Their search web interface gives the expected result: https://orcid.org/orcid-search/search?searchQuery=Frances%20Madden

Let me look into this. Or maybe @TomDemeranville has a suggestion

@maddenfc I checked what the ORCID API is sending us for “Frances AND Madden”. It is the same, result count of two, but only one record for Gabriel Vecchi. If I query for “Frances Madden”, you are not amongst the first 25 hits. Queries by specific fields, e.g. given-names and family-name also don’t help. I guess we need help from @TomDemeranville.

1 Like

I tried this with my name which returned nothing but the total count. I also tried it with the names of some of my colleagues and for some it worked fine, but a few times the results showed names that seemingly didn’t match the query at all, and the one I was searching for was not one of them.

Ok, I will follow-up with ORCID, as this is coming directly from their public API.