Sentiment analysis on Twitter applied to Donald Trump's #impeachment

Abstract

This article presents conclusions around the sentiments expressed on Twitter concerning Donald Trump's impeachment. A mixed approach, of a descriptive type, enabled us to obtain results on emotional aspects associated with users' discourse from the United States in the public sphere. The perspective employed was a transdisciplinary strategy to understand the relevant feelings of tweets in the context of their production. In this case study, data-mining methodologies and their typologies were applied: Text, graph and multimedia, as well as an analysis of multimodal discourse for the extraction and evaluation of affective values concerning the trial. We took into consideration all supports, codes and practices of discursive production in microblogging: Written and hypertext texts, social links and reactions, multimedia content and users' identities in a systemic relation of meanings. The publications of virtual communities in social networks about political processes are important to study in electoral situations and media events with a high degree of polarisation.