An Exploration into Social Media Sentiment

Abstract

Background:International and United States-specific media outlets cover the same news, but do not always utilize the same language. The COVID-19 Outbreak is an opportunity to analyze the sentiments being utilized to convey information to the masses. Exploring the words used and in what context can lead to more in-depth knowledge of what is being covered and how it is being explained by the media. Aim:The goal of this project is to analyze tweets from ten major news organizations, both local and abroad, by sentiment. News organizations will then be assessed for their portrayal of the pandemic in a positive or negative light, what sentiments they are using and the frequency, and what words are being commonly written together. This project will also be able to assess the discrepancies between US coverage and that of the world. Approach:Data will be processed through RStudio, utilizing sentiment data found in the NRC Emotion Lexicon and Bing Sentiments. The results will be correlated and graphed to show the variance between news coverage and language in the United States versus coverage during the same time abroad. Custom bigrams will also be created to explore more specific word connections, i.e., “COVID,” “corona,” “pandemic,” etc., nationally and internationally. Results: Tweets will be divided into data frames and then analyzed by word by both sentiment programs. Results for each news organization will be appropriately represented. Additionally, bigrams will be run on any words of significance. Results of the analyzed data and any statistical significance will be released. Conclusion: From the results, conclusions will be drawn regarding the sentiments nation and international news outlets utilize day-to-day.