Bryden et al. Word usage mirrors community structure in the online social network Twitter. EPJ Data Science 2013, 2:3.

This is my summary post. The intent of these summary posts is to provide a brief overview of the methods used and the results. 

Summary of
Word usage mirrors community structure in the online social network Twitter. 
Bryden et al. EPJ Data Science 2013, 2:3. full text here

Methods

Snowball sampling
The Map equation
Modularity
Z-score to rank words within a community
Euclidean distance with a bootstrap to determine word usage difference significance between communities. 

Results

Word frequencies of individuals differ by community, and by extension, to hierarchies of communities. A consequence of this pattern is that communities can be characterized by the word usage of their members; subsequently, the words of an individual member can be used to predict their community. Also the language patterns (word endings, letter-pair frequency, etc.) within a community are significantly different that those outside of the community. This makes sense: a shared language facilitates communication and interaction. They go on to say that words can act as cultural markers, which other individuals and groups use to decide whether or not to make a new association. This can be partially explained through homophily and peer effects.