In general, to find information in a triplestore, we have at our disposal a rather tedious “sparql endpoint”, i.e. a window to accommodate a request in Sparql language, possibly with some sample queries, but not more However the user base of a triplestore is not (necessarily) computer scientists. Persée helps to retrieve entire collections of humanities and social sciences publications and data.persee.fr opens to its users all the information relating to these resources. As people interested in humanities and social sciences are not, by training, experts in the Sparql language, we wanted to make our metadata really accessible, searches possible and our user self-sufficient.
This is why we propose Sparklis as an entry point for data.persee.fr.
Sparklis is a program developed by Sébastien Ferré, teacher-researcher at Irisa, which explores the graph of a triplestore (all the data linked together in the database), shows the entities defined as “classes” and proposes, in a drop-down menu, all the properties that can qualify the chosen class.
Example: I am looking for authors. So I choose Give me every *person*. I call this entity “Person”: Give me every person who is the author of something. I can then describe the “thing” this person authored (an article, a review, a bibliographical note, published in such a year, the subject is that keyword). I can then group the authors of the same article, or those who have published in the same journal. And order them according to the number of times they are cited by other works.
Selecting one option after the other, the user builds his request, refines it by grouping properties and cross-checking the results with additional criteria , and phrases his query in plain language simply building sentences. In the background, and in a transparent way for the user, each sentence is being built based on these options and “translated” into a Sparql query until a complete and well-formed query is obtained without the user ever having to activate computer knowledge, language or data structure.
once the query is finalized, if desired, the user can check the Sparql query, store or share a link to his search, download the data corresponding to the result of his query in different operating formats, or view these results directly in a graph using the integrated link to Google Charts.
Persée has entered into a partnership with Irisa to integrate Sparklis into the data.persee.fr website.
Thanks to this tool, and provided that the interface is up and running, any user wishing to search for information in the Persée triplestore can do so, independently.