Case Study: Toxicity Analysis of a Twitter Thread: Analyzing Replies to President Trump’s Tweet about Contracting the Novel Coronavirus
Conclusion First In this tutorial, we conducted an exploratory toxicity analysis of a Twitter dataset consisting of 298,172 replies to a Donald Trump’s tweet announcing his positive COVID-19 diagnosis. We found that most of the toxicity scores produced by Perspective API are highly correlated. We also found that tweets that are blocked by Twitter have,…