Quantifying socio-economic multiplex networks in a massive multiplayer online game

Understanding human group-behavior and social dynamics poses a unique challenge. One bottleneck in quantitative approaches is the availability and the quality of data. Massive multiplayer online games (MMOG) provide a fascinating way of simultaneously observing thousands of individuals engaged in social activities and pursuing basic economic activities to sustain a `living` in a virtual game-universe. We have compiled a unique dataset of a MMOG which is currently played by 300,000 players. In this game every player has to generate some virtual income through economic activities to `survive`, and is generally engaged and embedded in a multitude of social activities o?ered within the game. We recorded log-?les tracking practically every action of all players over the past 4 years. In previous work we started to test the validity of the game data as a model for non-virtual human societies. In this pro ject we propose to scienti?cally analyze this dataset with state of the art methodology of social network analysis and modern statistics. In particular we want to study structure and dynamics of friendship-, enemy-, trade, aggression- and communication networks and the mutual in?uence of these networks on each other (multiplex networks). For the ?rst time we have the unique possibility to measure the dynamical (!) in?uences of one type of social network on an other. We want to monitor and analyze `economic indices` in the game and relate them with social and economic behavior. We want to test how the `position` (degree, betweenness, clustering, etc.) of players within these networks in?uence their behavior. Further, we quantify group formation patterns and test social-stability hypotheses. Players disclose their gender in the game so that gender speci?c di?erences of `networking` patterns in social and economical behavior will be investigated. The dataset further allows to study economic decision making under quantitatively observable social and economical environments. This might be the ?rst large-scale study of a dynamical complex human socio-economic unit, with practically all social and economic data available from a unique and coherent source

Understanding human group-behavior and social dynamics poses a unique challenge. One bottleneck in quantitative approaches is the availability and the quality of data. Massive multiplayer online games (MMOG) provide a fascinating way of simultaneously observing thousands of individuals engaged in social activities and pursuing basic economic activities to sustain a living in a virtual game-universe. We have compiled a unique dataset of a MMOG which is currently played by more than 300,000 players. In this game every player has to generate some virtual income through economic activities to survive, and is generally engaged and embedded in a multitude of social activities offered within the game. We recorded log-files tracking practically every action of all players over the past 4 years. This data comprises one of the first instances worldwide, that complete information about a human society is available for scientific study. In this project we used this complete data to study a number of facets of human behavior and human interactions. In particular the detailed simultaneous knowledge of various timevarying social networks allowed us to quantify how humans form social bonds, how they organize, how they manage aggression, to what extent their behavior is gender specific, how the build their social relations, and how wealth of players is related to locations within in their high- dimensional social multiplex networks. We believe that studying model societies even though they are not real world societies nevertheless opens a window into a much deeper understanding how human societies are formed and self-organize into stable and resilient units. Moreover the methodology developed in the project might open a completely new perspective and a novel paradigm to a new way of doing computational social science with future big social data sets. The project has generated some public interested and was featured in many international newspaper articles, and has generated several well-cited publications. The work was presented in many international conferences, often as invited plenary talks.