Intended audience

This document is aimed at database maintainers.

In an era where accountability is central to science, counting reuse of scientific output must be done carefully. Where scientific output was restricted to articles and books in the past, now the situation is much more complex. More prominent nowadays is scientific output collected in databases, and focus is given to being able to cite data, in addition to conference abstracts and posters earlier.

This specification is about exposing what published literature is being references in databases. In particular, it details how a database can report how much data is extracted from a paricular paper, and outlines how this information can be retrieved.

Introduction

Altmetrics is a term generally used to refer to metrics that count the amount of impact of research output. For publications it can count the number of times people have bookmarked it at services like Mendeley, Zotero, and CiteULike, but it can also measure the number of pages views the publication has received.

This specification defines another altmetric: the amount of data from this paper that is found in databases. To expose such data, this specification defined a application programming interface (API) to query for such data. This idea leverages from the work by Andra Waagmeester on CitedIn [[Willighagen2013]], which uses this approach, but does not formalize an API.

The API is composed of two parts: one is the query interface that defines how the database can be queried; the second part is the format of the answer returned by the database. The implementation of the API can be done in various methods: first, there is a REST-like approach; the second is via a SPARQL end point.

REST-like API

Databases implementing this specification must provide an API that uses crafted URIs to specify the queries.

Queries

Data by DOI

This query asks for all entries that cite a particular reference identified by its DOI:

      http://example.org/entries/byDOI/$doi
    

Return format

The query results are provided as RDF as N-TRIPLES using the CiTO predicate cito:citesAsDataSource. For example:

      <http://example.org/entry/substance28490> <http://purl.org/spar/cito/citesAsDataSource> <http://dx.doi.org/10.1126/science.1157784> .
    

SPARQL API

Databases implementing this query type must provide an API that uses crafted URIs to specify the queries.

Queries

Data by DOI

This query asks for all entries that cite a particular reference identified by its DOI:

      SELECT * WHERE {
        ?entry <http://purl.org/spar/cito/citesAsDataSource> <$doi> .
        OPTIONAL { ?entry a ?type . }
      }
    

Return format

Any format specified by the SPARQL specification is allowed.