The analysis of social networks on sites like Facebook, Google+, Twitter and others, provides social scientists with an opportunity to examine the different groups that people belong to online and how individuals interact within, and between, those groups. Traditional analysis of a social network gives us information about a variety of features of the network based on the connections between individuals. This analysis however is devoid of the sociolinguistic context that can be used to define the exact nature of the relationships between people and the structure and functioning of groups and sub-groups within the network. Sociolinguistic analysis of communications between members of a social network can be used for a range of purposes, but one use is to define the types of groups that an individual is associated with especially ones related to ethnicity or language.

We describe a theoretical approach to the automated analysis of group features based on an automated sociolinguistic analysis of communications between the members. The types of communication we are interested in include communications with family, friends, colleagues, public individuals (e.g. celebrities, politicians), public companies (e.g. institution/company/government social media profiles) or public ads (e.g. marketing profiles).

The automation of the social network analysis involves web-crawling of public profiles of Facebook followed by a machine-learning approach to build models that can automatically identify the key language and sociolinguistic features defining different types of communication between different group members.

The results of this process are visualised making interpretation of the structure of the social networks amenable to analysis as either a static representation or reflecting changes over time.

The techniques outlined in this process can be scaled to extremely large networks by using Hadoop and other data pipelining techniques.

The importance of this analysis tool is the ability to explore the dynamics of individuals and their social groups. This is important in terms of being able to examine aspects of the digital sociological imagination. The techniques will allow the study of local and home support groups in migrant populations, or even visitors such as international students. More generally it can be applied to any research in which the sociological elements come to the fore.