A snapshot of politics on Facebook in a quantitative perspective – The angry, but not sad, populists


Facebook is full of people constantly expressing their opinions through reactions, comments, friend requests etc. It can be a great challenge for social scientists to extract patterns and find the general tendencies of how people are communicating on the platform.

Of course it would be presumptuous to explain how Facebook is used in its entirety; it’s human behavior after all. Instead, this post is a snapshot of how we can study the largest, most general tendencies of the relationship between news consumption and the political affiliation of users on Facebook.


The method used to provide the results in this post uses all publicly available data from the pages of all politicians and 100 of the most popular media organizations on Facebook in Denmark (300.000.000 data points). An algorithm uses the data from politicians’ pages to create political profiles on users by predicting their party affiliation. The behavior of users on media pages, for example when liking or commenting on news stories, is then observed with respect to their predicted political affiliation.


Using this predicted political affiliation to study reactions to news stories is possible from a myriad of angles. One straightforward question that we can ask is: who are the angriest, who are the saddest? So, among all reactions by all users on all news stories the last year, we look at how users with different party affiliations use the reaction buttons available on Facebook; and an interesting pattern arises. The Danish political party abbreviations can be found here: https://en.wikipedia.org/wiki/List_of_political_parties_in_Denmark


The four figures at the bottom show how much users affiliated with different political parties use the four reaction buttons: ANGRY, SAD, LIKE, HAHA. The level for each party is relative to that party’s overall use of reactions.


We see for example that the Danish populists (O, dark blue) are the angriest, but the least sad ones, whereas most left wing parties have higher scores for both angry and sad.

This little snapshot might seem like an unimportant detail on the surface, but considering the amount of data points behind the calculation it also provides some evidence for the political and ideological use of reaction buttons.



1.0 Average ANGRY reactions per party compared to average total reactions per party



1.1 Average SAD reactions per party compared to average total reactions per party



1.2 Average LIKE reactions per party compared to average total reactions per party



1.3 Average HAHA reactions per party compared to average total reactions per party



Jakob Kristensen

Jakob Kristensen