Understanding some fields of the Metadata Retrieval service via API

I need to consume the Crossref Metadata Retrieval service via API, but I don’t understand some fields of the JSON response. Given the JSON response for the request https://0-api-crossref-org.lib.rivier.edu/members/4419:

  1. How do I interpret the field coverage-type? For example, what the number 0.7772722166048759 means in coverage-type.all.journal-article.abstracts?
  2. What’s the difference between all, backfile and current?

Hi @wisentini ,

Thanks for your message, and welcome to the Community Forum.

  1. How do I interpret the field coverage-type? For example, what the number 0.7772722166048759 means in coverage-type.all.journal-article.abstracts?

This means that for all journal articles - both backfile and current - registered with Crossref, the percentage of those journal articles that include an abstract in the metadata is 77.7%. Many of these coverage values in the API map directly to values in our Participation Reports dashboard, like in this example:

From the REST API:

From Participation Reports:

  1. What’s the difference between all, backfile and current?
    This distinction also maps to our content registration fees.

“Current” includes the current calendar year + the previous two calendar years, and “Backfile” is anything older than that. For records registered with Crossref in 2024, current is anything registered with publication/award dates of 2024, 2023 or 2022. Backfile is anything registered with publication/award dates of 2021 or before.

My best,
Isaac

Oh, so let’s say the member “Universidade Federal de Sao Paulo” has a total of 1131 published journal articles. If its coverage-type.current.journal-article.abstracts value is equal to 0.93, I can assume that this member has around 1052 (0.93 * 1131) journal articles with an abstract?

Let’s use accurate counts just to be clear.

As of 2024-January-21, Universidade Federal de Santa Maria has registered a total of 1978 journal articles with a publication year of 2022, 2023, and 2024 in the metadata record. Remember, that’s how we define current. You can see that here: https://0-api-crossref-org.lib.rivier.edu/prefixes/10.5902/works?filter=type:journal-article,from-pub-date:2022-01-01,until-pub-date:2024-01-21

Now, of those 1978 journal articles with a publication year of 2022, 2023, and 2024, 1933 journal articles have an abstract registered with Crossref: https://0-api-crossref-org.lib.rivier.edu/prefixes/10.5902/works?filter=type:journal-article,from-pub-date:2022-01-01,until-pub-date:2024-01-21,has-abstract:true

That value is equal to 0.9772497472194135.

My best,
Isaac

Perfetct! One more question, though: What’s the difference between the fields coverage and coverage-type?

I still don’t get it how the coverage field is calculated and how it differs from coverage-type (besides not specifying the content type).

The coverage parameter calculates the percentage of the elements underneath it across the entire number of works across the member in question. So, for all journal content (backfile and current), all book content (backfile and current), etc.

The coverage-type parameter breaks the counts out into different content types and then separates those content types into backfile and current segments, so you can see the percentages based on backfile journal-article-level DOIs versus current journal-article-level DOIs versus journal-issue-level DOIs (again, backfile and current). If this member had more content types (books and book chapters, conference papers, datasets, etc.), you’d also see breakdowns by those other content types. Here’s a good example of that: https://0-api-crossref-org.lib.rivier.edu/members/186

My best,
Isaac