Media:
The New Social Power
A look at the use of word choice & sentiment analysis through Twitter

University of Hawaii - West Oahu

Abstract

 

Knowledge is power. An expression that is applicable today in the form of consumption of information on social media. In an inundated world of overwhelming clickbait information, the proficiency of selective word choice, within a specific context, has become increasingly connected to the amount of power a person may have. In this study, knowledge equates to selective word choice and sentiment to convey a message, and power is the influence and engagement of followers on a selective platform. Selective word choice directly correlates to the sentiment of a statement and the power it may yield through its conveyance. The study analyzes data from the social media platform Twitter, a platform based on microblogging and social texting, by gathering sentiments through texts (i.e. tweets) of powerful individuals and media outlets and their selective word choice on the platform.

Twitter Users' 20 Most Liked Tweets & Sentiment Usage (April 2020)

Interactive Node Tree

*click nodes for further stems

The interactive node tree shows users' top 20 tweets and the sentiments used as well as their ratio used per tweet. 

Built using RStudio

packages("collapsibleTree")

The Twitter Sentiment Analysis App

Analysing Trump Tweets Daily

The Twitter Sentiment Analysis App was developed using R and Tableau. Gathering data from a user's Twitter account, the App runs a sentiment analysis by scrubbing the user's tweets and referencing the Syuzhet lexicon. In this example, the App is showing Trump's Twitter usage, showing the amount he tweets by categorized by month in the pie chart, the polarization of sentiments he uses by percent in a radial graph, a word cloud of his 50 most used words, the amount of each sentiment he has used displayed in the bar graph, a filled line chart of how many tweets and retweets he has done within 2020, and a running index of his tweets. These are all organized and updated daily. 

Built using Tableau

Last updated May 12th 2020