MINING FACEBOOK PAGE FOR BI-PARTISAN ANALYSIS

Social media, particularly Facebook, has become ubiquitous in everyday life. Almost all news sources have adopted Facebook as a platform for dissemination of news. There are many opinions and studies on the partisanship of journalism. What makes social media interesting is that people do not only consume but also interact with others centered around a news article or post. Depending on the partisan bias of both the provider and the consumer, the interactions, and thus the conversation may vary. This research is a preliminary step towards mining these interactions and conversations pivoted against the topic of “fake news” from CNN and Fox News. We used several techniques of data mining, data analytics, and text analytics to generate summaries and descriptive statistics to explore user behavior. Our findings suggest that CNN follower base is more interactive and gregarious. Additionally, CNN followers’ use of Facebook reactions is more diverse, favoring the “haha” (funny / sarcastic) reaction, while those on Fox News’ inclined more towards “like” and “love” (agreement)