Big Data raises a lot of anticompetitive concerns. Those mainly stem from the significant direct, indirect and "learning-by-doing" network effects that characterize this industry but also from the Big Data firms capacity to "nowcast" flourishing markets and nascent competitive threats, which they can cheaply acquire at a very early stage. This thesis adopts two approaches to explore this topic. First of all, through a review of two data-driven mergers, this thesis tries to build a comprehensive view of those concerns and of how competition policy tools might be adapted. Secondly, the "essential facility doctrine" is applied to the Big Data industry. This part argues that the duty to deal access to data could be introduced on specific Big Data market segments as the four criteria set by the European Court of Justice can be fulfilled. Moreover, this part aims at showing that, from a welfare point of view; selling data to downstream competitors could in certain cases induce better-off economic outcomes, particularly in terms of dynamic efficiencies. Finally, this thesis shares the belief that this doctrine could also be used as a pan-European industrial policy.