In this article Zachary Crockett explained how he analyzed Donal Trump’s tweets over the past 7 months, he filtered out retweets, quotes, and article titles, and ended up with a list of 2,500 tweets (44,231 words). He ran a sentiment analysis (computational linguistics and a language processing) and found that 45% of his tweets are negative in sentiment, 27% neutral, and 28% positive. He compiled 60 adjectives most frequently used (provided a table and the number of times it was used), and came up with the following results: Trump’s use of positive words are things like “great” “good” “nice”, however negative adjectives are more complex “phony” “fraudulent” “unethical” “worthless” “hostile”. It’s also interesting to note that 76% of his tweets contain an exclamation point, and 72% of the time Trump tweets “Make America Great Again” it’s written in all caps. He tweets about the media more than he tweets about his policies, around 3.5 more times. Zachary also noticed that 23 of the 25 people Trump most frequently interacts with on Twitter are members of the press (article provided who he twitter interacts with the most).
The article uses a combination of qualitative and qualitative data to present his findings. He uses qualitative methods such as the language processing tools and linguistic searches (sentiment analysis) to come up with his quantitative data (number of times Trump said this or that, etc.) This data is then used to showcase the overall sentiment of Trump based on his Twitter, the article also mentions that Trump runs the account himself. The data is sort of telling us his personality, even if we didn’t already know. If the data was explained in a way that only showed us his negative tweets or his positive tweets then the overall frame of the article could have been very different, it would have seen either as pro trump or not pro trump. The article laid out the data in the way that the author found it and explained it in a way that was neutral, and didn’t seem biased.
No comments:
Post a Comment