Jargon and Graph Modularity on Twitter

The language of conversation is just as dependent upon word choice as it is on who is taking part. Twitter provides an excellent test-bed in which to conduct experiments not only on language usage but on who is using what language with whom. To find communities, we combine large scale graph analytical techniques with known socio-linguistic methods. In this article we leverage both curated vocabularies and naive mathematical graph analyses to determine if community structure on Twitter corroborates with modern socio-linguistic theory. The results reported indicate that, based on networks constructed from user to user communication and communities identified using the Clauset-Newman greedy modularity algorithm we find that more prolific users of these curated vocabularies are concentrated in distinct network communities.