Creating and Interpreting a Cultural Landscape on Twitter to Understand People and Audiences

Socio-cultural institutions play a large part in shaping our identities as well as signaling to others important parts of our identity. However, the space of such institutions, or "cultural touchstones", is large and the definition is broad; sports teams, places of worship, companies, and celebrities all could fall under the umbrella of socio-cultural touchstones. The New England Patriots, a Hindu temple, Nike, or John Lennon are just a few examples of these touchstones. Making sense of this space and understanding where people fall on it is essential to understand their preferences and identities. This work presents a few main contributions to this understanding. First, methods to create a "cultural landscape" based on influential Twitter users are presented, using the network structure as well as external data about each account. We find that evidence that these influential users self-sort and audience members sort them along dimensions of identity. Race, political orientation, and religion are the most salient dimensions for separation, while state geography, gender, and nationality are less important. We also compare audience sorting and self-sorting of these influential users, and find that audiences separate influential users relatively more on political orientation, religion, and gender, while the influential users self-sort more on race, occupation and LGBTQ status. This finding is a first step in quantifying the difference in how people view themselves versus how others may perceive them. Users and audiences are also placed in the cultural landscape to understand their most salient interests. Finally, audiences are broken down according to location in the cultural landscape to understand them at a more granular level. Both of these methods have significant potential commercial application as they provide granular insight into audience preferences in a a common space.M.Eng