Why should we expose metadata catalogues via SPARQL endpoints?

Chris Wood, Alexandra Kokkinaki, Adam Leadbetter

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

Many scientific disciplines have metadata catalogues managed by a designated responsible organisation. The content of these catalogues can vary by discipline, but may contain information about institutions who contribute data to central data repositories, projects that have been carried out within a particular field, or individual datasets. Within the European oceanographic science community, metadata catalogues have been managed under the SeaDataNet infrastructure. The five catalogues (for datasets, organisations, projects, observing systems, and cruise summary reports) have long been available online through individual web based search interfaces. However, such interfaces implicitly limit how the queries can be conducted, and how results can be viewed. Such limitations can be removed through the development of Application Programming Interfaces (APIs).

We have led the API development for all five catalogues, and have taken a Linked Data approach to the implementation. This approach requires the catalogues to be stored in triplestores, a form of graph database, with querying available using SPARQL, a query language for triplestores analogous to SQL for relational databases.

The query interface is publicly available over HTTP, allowing the whole catalogue to be openly queryable. The true strength of the triplestore approach is the ability to conduct federated queries across the different SPARQL endpoints which simulates a join between tables in a traditional table-based database, but without the need for subsets of the data to be located within the same database. A sixth triplestore, supporting the NERC Vocabulary Server (NVS), underpins the content of the catalogues by providing lists of standardised terms that provide consistency and semantic harmonization across the catalogues. Federated queries can be used to determine the theme of a dataset, described in NVS terms, by simultaneously querying the SeaDataNet dataset triplestore and the NVS triplestore.

In this presentation, we will show the advantages that can be gained by both the catalogue publisher and the end user, the steps needed to setup a new triplestore and the corresponding SPARQL endpoint and the ease by which this software stack can be implemented. We will finish with the lessons learnt in this project, and the future work that will be carried out to further enhance the service to end users.
Original languageEnglish
Title of host publicationGeophysical Research Abstracts
Subtitle of host publicationEGU General Assembly 2018
Number of pages1
Publication statusPublished - 8 Apr 2018
EventEGU General Assembly 2018, Vienna, Austria -
Duration: 8 Apr 201813 Apr 2018


ConferenceEGU General Assembly 2018, Vienna, Austria
Internet address


Dive into the research topics of 'Why should we expose metadata catalogues via SPARQL endpoints?'. Together they form a unique fingerprint.

Cite this