An NLP-assisted Bayesian time-series analysis for prevalence of Twitter cyberbullying during the COVID-19 pandemic

This article presents a comprehensive study on the prevalence of Twitter cyberbullying during the COVID-19 pandemic using NLP-assisted Bayesian time-series analysis. The authors begin by defining cyberbullying and its connection to the COVID-19 pandemic, highlighting the increased use of social media during lockdowns. They then describe the methodology, which involves collecting and analyzing tweets using a pre-trained NLP model to filter relevant content. The study spans three years, from 2019 to 2021, and reveals strong weekly and yearly seasonality in cyberbullying trends. The authors also compare these trends with COVID-19 case counts, finding interesting correlations and seasonal patterns. The use of Bayesian time-series modeling allows for a detailed analysis of trends and their dependencies over time. The findings suggest that while cyberbullying may have increased initially during the pandemic, it has since decreased, with significant seasonal variations. The study concludes with a discussion of limitations and suggestions for future research, including the potential for more complex modeling and spatial analysis.