A visual analytics approach for understanding egocentric intimacy network evolution and impact propagation in MMORPGs

Massively Multiplayer Online Role-playing Games (MMORPGs) feature a large number of players socially interacting with one another in an immersive gaming environment. A successful MMORPG should engage players and meet their needs to achieve different categories of gratifications. Research on the evolution of player social interaction network and the dynamics of inter-player intimacy could provide insights into players' gratification-oriented behaviors in MMORPGs. Such understanding could in turn guide game designs for better engaging existing players and marketing strategies for attracting newcomers. Conventional dynamic network analysis may help investigate game-based social interactions at the macroscopic level. However, current dynamic network visualization techniques mainly focus on illustrating topological changes of the entire network, which are unsuitable for analyzing player-specific social interactions in the virtual world from an egocentric perspective. In general, game designers and operators find it difficult to analyze the way players with different gratification needs may interact with one another and the consequences on their relationships with direct ties, using a decentralized social graph with complicated time-varying structures. In this paper, we present MMOSeer, a visual analytics system for exploring the evolution of egocentric player intimacy network. MMOSeer focuses on the relationship between a player (ego) and his/her directly-linked friends (alters). We follow a user-centered design process to develop the system with game analysts and apply novel visualization techniques in conjunction with well-established algorithms to depict the evolution of intimacy egocentric network. We also derive a centrality change metric to infer how the impact of changes in an ego's interactive behaviors may propagate through the intimacy network, reshaping the structure of the alters' social circles at both micro and macro levels. Finally, we validate the usability of MMOSeer by discovering different user interaction patterns and the corresponding ego-network structural changes in a real-world gameplay dataset from a commercial MMORPG.