I. Introduction
Working in groups was facilitated with the advent of Web 2.0 technologies whose use for this purpose, becomes more widespread. Their use makes it possible to collect various and varied data (Big Data era) that can be exploited for different purposes such as decision support. If the interest to work in a group is not to prove for the realization of certain tasks, the achievement of intended results and/or the success of the collaboration depends on the composition of the group. The process of developing the latter remains complex. Group composition is usually considered in terms of how individual member characteristics will affect group cohesion or compatibility and subsequently how the group interacts. These charateristics is what we call in this paper: Contextual information/data of the learner, or learner's context. Group composition affects many aspects of the goal achievement, such as how efficiently g roup m embers w ork t ogether and how much relevant knowledge they can share. Thus, when deciding how to compose groups for a project, some tasks are necessary, some are dependent on learner's context such that: (a) Define t he p roject l earning o bjectives (depending o n the learners), (b) decide how to configure t he g roup (depending on some constraints (technical, linked to learner's context, …)), (c) select group members (the characteristics of a group are not the sum of the characteristics of each constituent), (d) develop a contingency plan in case group membership changes during the project; And some are contextual information such that: (e) Identify characteristics of group members relevant to the project learning objectives.