Network Theoretic Approaches for Understanding and Analyzing Social Media Based News Article Propagation

Abstract

Characteristically, propagation of news on the Internet is a rather complex scenario. Its comprehensive understanding requires a consideration of diverse facets such as audience, problem domain, channel and type of news being propagated. My dissertation focuses on the understanding of propagation of a specific type of news- news articles, on a particular subset of the Internet, the social media. While a number of studies have looked into the phenomenon of propagation in social media, fewer of these have examined the propagation of content, specifically news articles, published by news provider websites. My dissertation presents a set of network theory based methodologies to extract and analyze various implicit propagation networks formed as a result of news article sharing on Twitter. These methodologies cover aspects related to users' article sharing behavior, influence of the news provider's social media accounts, role of followers and similarities between propagation networks of news providers. Furthermore, it also includes useful inferences derived about the news article propagation phenomenon by using a population sized data sampled from Twitter over a nine-month period. It expands on the inferences from my published works and the challenges identified in the area of news article consumption and distribution on the Internet. My dissertation intends to provide important guidelines for researchers and organizations studying social media related phenomena to derive insights about customer behavior. From the perspective of online news consumption and distribution, my study has important implications for the audience's preference of news content delivery. It also facilitates news providers to gauge their performance on social media and for news editors to help develop editorial policies tailored for an online consumer base. Finally, my dissertation presents an extensive set of network based models and methodologies that can enrich the applied network science discipline.