Data analysis revealing behavior patterns

Studying human behavior in virtual environments provides extraordinary opportunities for a quantitative analysis of social phenomena with levels of accuracy that approach those of the natural sciences. ASA researchers employed concepts from physics to social science to analyze the “big data” of a multi-player online game. Similarly to the interactions between particles, the notions of “force”, “acceleration”, and “interaction potential” can be used to describe interactions between people in the virtual world, and can be quantified by recording the number of messages or amount of goods, etc., which people exchange [1]. In a case study of this, ASA researchers analyzed the dynamical features of sequences of actions of players depending on players’ traits (age, gender, etc.); finding that while men and women act similarly when performing cooperative actions, women are slightly faster for non-cooperative actions [2].

Big data analysis can also be used effectively to explore public health issues. For example, ASA researchers used a national data sample of almost two million diabetic patients to show that these individuals have a higher risk of Parkinson’s disease, depression, and schizophrenia. Hypertension was found to be highly sex-sensitive for this group: women were at lower risk during the fertile ages, but higher risk otherwise [3]. These results may be useful to improve screening practices in the general population and this and other possible ways of using big data analysis are being discussed with the decision makers in the Austrian health care system.

Last edited: 16 March 2016