MULTILAYER ANALYSIS OF ONLINE SOCIAL INTERACTIONS

The development of online platforms and our capacity to collect data of individuals from these has drastically changed the way scientist explore human society. Social systems can be understood as a superposition of various social networks, where nodes and edges represent individuals and their social relations, respectively. While these networks have been mainly studied as aggregates, only recently new theoretical frameworks have allowed to understand their multidimensional nature. In this work, we analyze a multirrelational social system from a massive multiplayer online game, consisting on three networks of different types of one-to-one interactions (friendship, communication and trade). Firstly, we study their structure separately using the usual single-layer approach. We find that two of them have a very similar structure while the third one is more denser and presents an slightly different connectivity pattern. Nevertheless, this is not maintained in time, which suggests that the networks are very dynamic and vary a lot depending on what is going on inside the game. For the second part, we implement measurements that take into account the multidimensional nature of the system and give insights on how players behave differently depending on the interaction under consideration. The results obtained with these measures indicate that players behave similarly in all the different layers, but that they do so with different people, meaning, for example, that they tend to avoid becoming friends with people they trade with or that they do send messages to people which are not necessary their friends.