vocabulari.se is a search tool addressed to researchers to explore a social network of research objects, help them choose the wording of their abstracts or the tagging of their papers, and discover their research interests in new ways. This tool is grounded in “science studies” , which show that research objects have a social, even political, life of their own… let us play with it! And open the black box of the ubiquitous yet poor tagging system as the ultimate linking device of words…

vocabulari.se searches for terms related to a given term (i.e. a research object, treated as a Mendeley tag or a Wikipedia entry) according to three qualities of relationships:
– Unexpected relationships (i.e. not frequent yet effective) that give originality to research
– Controversial relationships (i.e most discussed on Wikipedia) that widen the audience ;
– Aggregating relationships (i.e. most multidisciplinary Mendeley associations) that bridge the gap with other disciplines’ interests.

This tool is a mash-up of Mendeley and Wikipedia APIs. Our algorithms start from an ensemble of related tags obtained through co-occurrence analysis in Mendeley data. Then we sort the results through the slope of frequency vs. readership relationship (algorithm 1), through the size of Wikipedia discussion pages resulting from querying two related tags (algorithm 2), through the multidisciplinarity of Mendeley publications readership (algorithm 3).