Is Donald Trump’s Presidential Election Victory a Surprise According to Twitter Data?

Abstract

NodeXL Pro is a software developed for network analysis and visualization. NodeXL Pro connects to twitter and extracts tweets about the topics that are set, and makes various analyzes with these tweets. In this study, during the US presidential election held on November 8, 2016, the tweets about the candidates were handled and sentiment analysis was performed on these twitters. When you look at tags, on November 8, 2016, the most popular tags on twitter were; hillaryclinton and trump. Instantaneous; hillaryclinton’s twits number were 24407, compared to trump labeled twets number were 4132. When tweets under both labels were examined, it was seen that the majority of the twitters did not have words with emotional expression. On the other hand, hillaryclinton labeled tweets; 1761 positive emotion words were found and 828 negative emotion words were detected. It is known that Trump had focused on social media throughout the campaign period. Although the instant twet number of the trump tag was less than the hillaryclinton tag, the number of words expressing positive emotion was 5411 and the number of words expressing negative emotion was 1659 in these twets. For Hillary Clinton, the ratio of the number positive emotion words to the number of negative emotion expression words was 2,12, about Trump while the rate of the number of positive emotion words to negative emotion words was 3.26 in tweets. In hillaryclinton-tagged tweets, with the most popular positive words; Proud, love, worked, win and wins, most popular negative words; Hate, collapse, corruption, lies and f..k. In trump-tagged, for the most popular positive; "wins, win, defeat, good, trust, amazing, supporter and work" words, for the most popular negatively; "badly, refuses, lost, f..k, hell, loses and dump" words were the most common words. When word pairs are examined; The hillaryclinton word was used in combination with the most potsword (612 times) and the word with beyonce (603 times). Again, in the twets with hillaryclinton tag positively emotional sentences the "proud" and "same" words had been used together (139 times), "worked" and "toward" words (130 times) . In twitler expressing negative emotion; The words "collapse" and hillaryclintons have been used together (29 times), "corruption" and "looks" (28 times), "lies" and "vote" words (19 times). Trump tagged twets; The trump word was mostly used; with the Donald word (563 times), vote word (198 times) and wins word (169 times). When you look at the tweets that were triggered by the Trump tag and express a positive feeling; Most of the words "trump" and "wins" (169 times), "trump" and "supporters" had been used together (123 times). When you review negative tweets that are trump labeled; The words "refuses" and "allow" (57 times), "hell" and "out" (43 times) were used together. Despite the fact that when trump and hillaryclinton-tagged twits were emotionally analyzed, the number of tweets about Trump was much less than the number of tweets about Clinton. It seems that, the number of positive emotion expression words in tweets about trump were too much in terms of the number of positive emotion words in tweets about Clinton. It is seen that the words that express positive and negative emotions about Trump and Clinton are generally very different from each other.