Grammar of social media
We use natural language processing to characterize the unique properties of online communication and investigate whether different topics have identifying patterns of discourse. Investigating these differences provides a better understanding of human language and, more concretely, facilitate the rapid identification of topics by their content. Current methods, such as those in many search engines, identify topics by statistical relationships of words, which only loosely relates to the meaning (semantic content) of words.
Organization of social media
We use graph theory to characterize the structure of online communities and investigate whether different communities have categorically different organizations. We, then, use agent-based modeling to understand how those differences in organization relate to differences in the behavior of individuals online. Investigating these difference provides a better understanding of the organization of human communities.
We test the effects of changes in public health policy on simulated networks of people.