A Study On The Effects Of U.S. Events On Sentiment Relating To Minority Groups On Twitter

This is a study on how U.S. national events affect the sentiment relating to different minority groups on Twitter. First, several groups of minorities were identified for examination. Next, the relationship between hashtag usage and Tweet composition relating to various minority groups was investigated. Using the Twitter API and several Python packages, we were able to gather Tweets relating to our areas of interest over the months of June to August 2020. Average Tweet sentiment was determined to see how that sentiment changed before, during, and after selected events. Three metrics were used to draw conclusions from the collected data. These were average sentiment of the Tweets, percentage of Tweets that were “retweets” versus those composed originally, and number of Tweets that were deemed “relevant” by twitter and therefore available for sample. We found that Tweets relating to the Black Lives Matter and related movements decreased in original tweet composition and average sentiment. Overall, this indicated a negative effect on the sentiment relating to Black Lives Matter initiatives during several related events. Tweets relating to the Coronavirus and people of Asian ethnicity did not change in average sentiment and decreased in original tweet composition in relation to the Coronavirus epidemic. Overall, this also indicated a negative effect on the sentiment relating to Asians due to this world event. Finally, Tweets relating to members of the LGBTQ+ community increased average sentiment, and decreased in original tweet composition during the month of June, commonly associated with “Pride Month”. This indicated a positive effect on the sentiment relating to the LGBTQ+ community during this event. We can come to the overarching conclusion that U.S. national events do not have a direct effect on positive or negative sentiment independently, rather that these events cause a change in the sentiment that can be characterized as either more positive or more negative. People are much more likely to retweet Tweets relating to events that elicit visceral reactions from people. This is a common sign of support. Since the original Tweet often has polarized sentiment due to the emotionally charged content, we can conclude that the events of the study will result in high retweet percentages with polarized sentiment.