JISC Open Citations Project web site

The JISC Open Citations Project website at http://opencitations.net exists for several purposes:

and, most importantly,

The JISC Open Citations Project Home Page

For the purposes of this demonstration, we are running an instance of Fuseki, that provides access to the Open Citation Corpus in an underlying TDB quad store, with a local patch to enable query timeouts, as was done for the SPARQL endpoint of CLAROS, a related project in which we have been intimately involved.

The web site and the underlying RDF citation data follow key linked data principles:

  • We don’t use blank nodes
  • All URIs are dereferenceable and content-negotiable.
  • Where possible we use standard identifiers based on DOIs, ISSNs and ISBNs.

For users wishing to jump into the data, we provide tabs to access data about journals and articles, each giving greater detail about the contents as one clicks in.

The Journals page

Each Journal page displays metadata about the journal, and links to useful information and the ability to download the journal metadata in a variety of formats.

Details of one journal

The articles tab displays a subset of about 10% of the entire corpus, for speed of loading the demonstration.

The Articles Page

Each ‘article’ page displays details about the selected article and its citation network, using graphviz, along with a user input form to customise the display of the network, namely for input or output citations, and for different forms of display. The article metadata and the citation diagrams can be downloaded in various formats.

Input citations for Codon, C. (2007). Maturation and degradation of RNA in bacteria. Current Opinion in Microbiology 10(3): 271-278. doi:10.1016/j.mib.2007.05.008.

We also display the RDF properties of each entity, allowing a ‘follow your nose’ style of data discovery. Each page lists the SPARQL queries that were run to generate it, enabling the user to see how the site works and to tweak the queries for their own ends.

Further work is planned to display citations along a time axis, and to develop other tools to permit users to exploit the data.

Functional clustering of CiTO properties

CiTO v2.0 contains just two main object properties, cito:cites and its inverse cito:isCitedBy, each of which as thirty-two sub-properties. Intentionally, these properties are not constrained as to domain or range, thereby maximising their applicability in a wide range of citation contexts. Additionally, CiTO contains one additional generic object property, cito:shareAuthorsWith, that may be used even outside a citation context.

Some have criticised CiTO for having too many properties, making it confusing for potential users, and Martin Fenner chose to use only the ten ‘most popular’ properties in his CiTO plug-in for WordPress, previously mentioned in this blog post. In response, I would point out that each property has a distinct and clearly defined meaning, and that together they provide an appropriate level of expressivity for effective use. Nevertheless, it is clearly difficult to conceptualize all the CiTO properties at one time, if they are being viewed using an ontology editor such as Protégé. I have thus created the following figure that groups CiTO properties by similarity, in the hope that this will facilitate choice of the most appropriate one.

Clustering of CiTO relationships by similarity

As shown in the figure, the CiTO properties and sub-properties (and, consequently, their inverses) may be classified as rhetorical (upper oval) and/or factual (lower oval, dark blue text), with the rhetorical properties being grouped in three sets depending on their connotation: positive (green), neutral or informative (blue) and negative (red). Five properties (in purple, within the overlap of the ovals) have both factual and rhetorical characteristics. The inverse properties are not shown.

Graffoo, a Graphical Framework for OWL Ontologies

Graffoo, a Graphical Framework for OWL Ontologies [1], is a wonderful new open source tool developed by Silvio Peroni that can be used to present the classes, properties and restrictions within OWL ontologies, or sub-sections of them, as clear and easy-to-understand diagrams. Several Graffoo diagrams have been developed to explain SPAR ontologies, or portions of them, and are to be found in the appropriate ontology directories at https://sempublishing.svn.sourceforge.net/svnroot/sempublishing/.

For example, in Figure 1 the diagram, BiRO.png, illustrates the structure of BiRO, the Bibliographic Reference Ontology:

Graffoo diagram of BiRO, the Bibliographic Reference Ontology

The upper part of this diagram shows how a bibliographic record, which in FRBR terminology is a Work, (a) is a set of components (author names, publication year, title, volume number, ISSN, publisher, dates of submission, acceptance and publication, etc.), (b) references a published entity (a FRBR Endeavor), and (c) can form part of a bibliographic collection such as a library catalogue, which is also a set of components, in this case of bibliographic records.

The lower part of the diagram shows how such a bibliographic record can be realized (i.e. expressed in a specific form, forming what in FRBR terminology is an Expression) as a bibliographic reference, such as you might find in an article’s reference list. The bibliographic reference, while referencing the same published entity, differs from a bibliographic record in two important ways.

  • First, a bibliographic reference typically does not include all the components of the full bibliographic record. References typically exclude the ISSN, the name of the publisher, and the submission and acceptance dates; may exclude the publication date and some of the authors’ names if there are many authors; and in certain journal (e.g. Nature) also exclude the title.
  • Second, the components of a bibliographic reference form an ordered list rather than a set, although the precise order of components can vary from journal to journal (e.g. whether or not the publication year immediately follows the authors’ names).

The diagram also shows that, just as bibliographic records may be grouped into bibliographic collections, so bibliographic references may be grouped into bibliographic lists such as reference lists (which, of course, are themselves ordered lists), nicely exposing the symmetry within the BiRO ontology in a manner that would be hard to grasp simply by looking at the class structure within an ontology editor such as Protégé.

The advantages of using such a Grafoo diagram are thus that it displays the logical relationships between elements of an ontology, or a sub-section of an ontology, in a manner that is relatively straightforward to understand, once one has grasped the meaning of the different elements of a Graffoo diagram. These elements are shown and defined in the Graffoo key (Figure 2).

Figure 2: The legend for all possible Graffoo objects

From our preliminary empirical studies, it appears that Graffoo allows us to create representations of OWL ontologies that can be comprehended in detail without the person viewing them having to understand the details of OWL 2 or of any of its linearizations (Turtle, RDF/XML, Manchester Syntax, or OWL/XML).

Graffoo has been developed using the standard library of the yEd diagram editor, a free diagram editor running on Windows, Mac and Linux. The graphml format version of those Graffoo objects is also available here. We commend the use of Graffoo to other ontologists.

[1] Graffoo, the Graphic framework for OWL ontologies. http://www.essepuntato.it/graffoo.

Search OpenEdition Search

You will be redirected to OpenEdition Search