Presence Analytics: Making Sense of Human Social Presence within a Learning Environment

The various activities that take place within an observed environment such as a university campus, determine to a large extent, the kind of social interactions exhibited by the users in such environments. Using a big data set of wifi-traces, we attempt to understand the rules that governs these social interactions. We discovered that there are at least two types of social interactions within a university campus: formal such as attending a class and informal such as meeting friends at the cafeteria for coffee. Each of these two types of social interactions is tightly associated with a specific set of locations within the university campus. We also discovered that users tend to restrict their social interactions to a small set of geographical locations, where users revisited the same location to socialise with the same social group. Also, irrespective of the type of the social interactions, users tend to restrict their revisits to geographically nearby locations and only revisit locations that are further afield when they are in the company of their social group. These findings are based on the social groups detected by a new scalable density-based clustering method applied to a large data set of mobile users wifi traces. The results of the large experiments carried out in this research demonstrate how the proposed algorithm can noninvasively detect social groups on the basis of the activity performed at the selected location.