The 2017 election cycle has been like no other—especially when it comes to responses to the candidates on social media such as Twitter. Donald Trump has taken to Twitter more than any other candidate in history. And Hillary Clinton has criticized him for being baited by tweets. Responses to the Republican and Democratic presidental nominiees have been polarized to an unprecedented degree. Tassie Gniady, Cyberinfrastructure for Digital Humanities, has been analysing the sentiment of the election tweets using custom text analysis tools. To do this, each word in each tweet obtained from Twitter was assigned a value: positive words received a +1 score, negative words received a -1 score. These scores were then summed to give the tweet a total score. On November 2, 2016, as shown below, Hillary Clinton was seen more favorably than Donald Trump across the positive spectrum. Donald Trump was seen more negatively (especially at -4). However, this was not always the case.
On May 5, 2016, just a few days after Trump won the Indiana primary, Trump is seen both extremely positively (+3) and extremely negatively (-3) in comparison to Hillary Clinton, whose only clear dominance appears at -1. However, in July, Trump is seen more positively than Clinton (who, at the time, had not yet been endorsed by Bernie Sanders). By October, the visualization portrays a tightening race; this is somewhat surprising given that this was not long after the 1st presidential general debate in which Clinton was the decided winner.Sentiment analysis of election tweets in early May (first), July (middle), and October (last) 2016.