Towards a Formal Model of Social Data

Computational social science (CSS) is an emerging field of research that seeks to apply computational methods and tools to important and interesting social science questions and problems. Situated within CSS, Social data analytics as a research stream aims to collect, archive, retrieve, process, transform, analyse, and report social data from social media platforms such as Facebook and twitter. Formal methods, models and tools for social data are largely limited to graph theoretical approaches informing conceptual developments in relational sociology and methodological developments in social network analysis. As far as we know, there are no integrated modeling approaches to social data across the conceptual, formal and software realms. Social media analytics can be undertaken in two main ways - ”Social Graph Analytics” and ”Social Text Analytics” (Vatrapu, in press/2013). Social graph analytics is concerned with the structure of the relationships emerging from social media use. It focuses on identifying the actors involved, the activities they undertake, and the artifacts they create and interact with. Social text analytics is more concerned with the substantive nature of the interactions, it focuses on the topics discussed and how they are discussed. What keywords appear? What pronouns are used? How far are negative or positive sentiments expressed? In this report, we first present and discuss a conceptual model of social data followed by a formal model based on set theory. Second, we exemplify the semantics of the formal model with real-world social data examples. Third, we briefly present and discuss the Social Data Analytics Tool (SODATO) that realizes the conceptual model in software and provisions social data for computational social science analysis based on the formal model. Finally, we exemplify our approach with help of a case study on big social data of the fast fashion company, H&M. from its Facebook page.