Section: New Results

Multi-model Querying

We have proposed a lightweight data integration architecture implemented within Tatooine (see Section 6.1.5); the system was demonstrated on a data journalism use case at the prestigious VLDB conference [9].

A separate effort in the area of multi-model querying considered querying databases of interconnected documents, users and concepts, by means of keywords. In this context, it is important that query results reflect not only the keywords present in documents but also the links between users and documents (so as to return to one user first the results authored in his social neighborhood), links between documents (for instance when a tweet answers another or an article has a link to another), and last but not least semantic information which allows interconnecting and interpreting terms mentioned in text. This research was finalized as part of the PhD of Raphaël Bonaque  [1] and appeared at the EDBT conference 2016 [11].