Xenophobia Online Project Progress Report 7/26

After preparing the twitter data for April and May, we were able to visualize the data and come up with some questions about the data. Were the spikes in COVID-19 related tweets caused by outbreaks or possibly even COVID-19 related events such as new legislation or hate incidents that occurred that day?

We were also able to visualize the most frequently used sinophobic slur within the COVID-19 related tweets of April and May. Within our word list, ”chinazi” happened to be the most commonly slur rather than “chink” or “chingchong” which was surprising.

Within the past two weeks, we contacted a wide range of organizations who focused on similar xenophobic research for any data we could use to visualize for them as well as mentors. We were able to recruit Dr. Banda, a data engineer at Panacea Lab and associate professor at Georgia State University, as a mentor for Xenophobia Online. After our first meeting with Dr. Banda, we were already able to get quick and insightful feedback for our May/April visualizations. We learned that those visualizations aren’t accurate and that the extremely small percentages didn’t really have any meaning because we only used a very small portion of the dictionary to feed into the NLP for the word list. Once we are able to detect more instances of other sinophobic slurs from our dictionary, we will be able to increase the accuracy of our visualizations by aggregating similar slurs and not leaving out any slur. We will also focus on expanding our dictionary through adding on existing dictionaries and other resources we find.

Our plans for the upcoming weeks are to expand our dictionary, research other methods for data analysis including TF-IDF, NLTK, Emotion Lexicon and Hedonometer so that we can restart the April and May data for better and more accurate visualizations.