By Luce Liu, Sherry Xie, Amy Zhu
We're three UBC students who are interested in exploring patterns concerning truthfulness and user engagement amongst political Facebook posts. The dataset explored here consists of 2282 Facebook posts sourced from nine American political pages over the course of a week in September 2016. Six of the pages are hyperpartisan, and three are considered mainstream.
The truthfulness ratings were provided by Buzzfeed News, as part of their analysis on the topic which you can read here. You can explore the dataset here.
Does non-factual content get more engagement? How is engagement distributed across the political spectrum?
Each bubble represents a post, where the size of the bubble indicates the combined number of likes, reactions, and shares for that post. Click on a bubble to see the original post.
* Posts that were satirical or opinion-driven, or that otherwise lacked a factual claim were rated "no factual content" (Buzzfeed News)
What percentage of a Page's posts fall into each category? How do the Pages rank against each other for each category of content?
Compare post engagement by truthfulness rating between Pages. Each circle represents a post. Does false content get more engagement for some Pages?
Engagement count
What is the breakdown of post engagement across the political spectrum? Which audience engages most with mostly false content? Does opinion-based/satirical content get more engagement amongst certain political categories?
* Posts that were satirical or opinion-driven, or that otherwise lacked a factual claim were rated "no factual content" (Buzzfeed News)
How much engagement is typical for posts of different ratings? Which ratings and formats produce more highly-engaged-with posts?
Each coloured curve is half a violin chart; stronger curves represent more posts.
Engagement count
Post format