Understanding the perception of COVID-19 policies by mining a multilanguage Twitter dataset

TL;DR

This study analyzes multilingual Twitter data related to COVID-19 to understand public discourse, misinformation spread, and policy perception over time, providing insights for future pandemic response and a valuable dataset for researchers.

Contribution

It introduces a multilingual COVID-19 Twitter dataset and applies NLP and network analysis to explore discourse dynamics and misinformation transmission during the pandemic.

Findings

Identified common public responses to COVID-19 over time

Mapped misinformation spread patterns on Twitter

Provided a multilingual dataset for future research

Abstract

The objective of this work is to explore popular discourse about the COVID-19 pandemic and policies implemented to manage it. Using Natural Language Processing, Text Mining, and Network Analysis to analyze corpus of tweets that relate to the COVID-19 pandemic, we identify common responses to the pandemic and how these responses differ across time. Moreover, insights as to how information and misinformation were transmitted via Twitter, starting at the early stages of this pandemic, are presented. Finally, this work introduces a dataset of tweets collected from all over the world, in multiple languages, dating back to January 22nd, when the total cases of reported COVID-19 were below 600 worldwide. The insights presented in this work could help inform decision makers in the face of future pandemics, and the dataset introduced can be used to acquire valuable knowledge to help mitigate the COVID-19 pandemic.