Collaboration has been defined as a recursive process where two or more people or organisations work together to realise shared goals by sharing knowledge and building consensus. Business Intelligence (BI) applications have so far been oriented to provide timely information unidirectionally to decision-makers. However, things are changing, and many companies are moving towards a model where (at least for some company issues) employees from all areas of a company will be able make important decisions for their areas of responsibility, backed by easily-accessible information and analyses. In this model, simply publishing data does not suffice. Collaborative tools are needed to make it possible for employees to search, rate, comment on, and request enhancements to such shared data. The goal, in the end, is to take to BI what is already going on in the so-called Web 2.0, where online social collaboration platforms already exist, allowing knowledge sharing across the members subscribed to such platforms. However, although there are some BI tools that promise collaborative work, there is still a lack of consensus on how true collaborative BI can be really achieved. Most efforts so far reduce to use and extend already known concepts from the data integration field to the multidimensional setting. However, collaboration can take many forms, the simplest of which would be sharing multidimensional data. However, more sophisticated collaboration forms may be devised, where recommender systems tools could be used, as well as quality scores, rankings, etc. In this topic we envision the problem of not only building a platform or an architecture for collaborative BI, but also the development of models (and metamodels) that can represent the different ways in which this collaboration can be carried out.

Main Advisor at Université Libre de Bruxelles (ULB)
Co-advisor at Technische Universität Dresden (TUD)