Date of Award

May 2022

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Information Studies

First Advisor

Xiangming Mu

Committee Members

Jin Zhang, Wonchan Choi, Min Wu

Abstract

In December 2019, the coronavirus disease 2019 (Covid-19) was officially reported as an acute respiratory infection, which was first identified in Wuhan, China. On March 11th, 2020, the World Health Organization (WHO) declared that Covid-19 could be characterized as a pandemic. Governments across the world imposed or recommended various non-medical interventions to reduce transmission of Covid-19, such as washing hands, wearing face masks, social distancing, and quarantining as well as lockdown measures including banning large gatherings, issuing stay-at-home orders, closing certain businesses, and imposing travel restrictions. The increase in social, behavioral, and economic issues that the disease has generated has had both negative effects, such as large numbers of new cases and mortality at its peak times and locations, and positive effects, such as scientific discoveries. Given the enormous and lasting impact of the pandemic, the primary aim of this study was to explore the information flow, users’ perceptions and sentiments and topics dominating social media as captured in Twitter in the 60 days before and the 60 days after the declaration of the Covid-19 pandemic and analyze these data using machine learning techniques.Over 68 million tweets for this study were collected from Georgia State University's Panacea Lab. Before being analyzed, the data were prepared and cleansed. After cleansing the ii data, 21,655,284 tweets were used for data analysis and categorized into tweets posted during the 60 days before the declaration of the Covid-19 pandemic (3,406,055) and tweets posted during the 60 days after the declaration (18,249,229). Three machine learning techniques and inferential analysis were applied. The sentiment analysis, and emotional analysis used to understand users' characteristics, a Latent Dirichlet Allocation (LDA) model was employed to uncover discussion topics about the Covid-19 pandemic that users tweeted. Inferential analysis was applied to investigate the differences between sentiment characteristics and discussion topics. The results of the sentiment characteristics and emotional analysis show how the declaration of the Covid-19 pandemic change the users’ characteristics during the Covid-19 pandemic on Twitter. The results of the LDA model show various discussion topics 60 days before and 60 days after the declaration of the Covid-19 pandemic. The inferential analysis showed there are no differences between the two groups. The findings of this study revealed the characteristics of users on Twitter in the early stage of the Covid-19 pandemic. Theoretically, the findings demonstrated the significance of social media for policymakers in monitoring a health crisis. The distinctive implication of this study was to identify users' characteristics on Twitter as a source of information and knowledge relevant to managing a health crisis for policymakers. The methodology employed in this study demonstrated a systematic approach to evaluating Twitter users' behaviors in the early stage of the Covid-19 pandemic. The outcomes of this study can contribute to establishing a fast and low- human-effort surveillance system for monitoring people’s attitudes towards policies and policymakers during a health or other crisis.

Share

COinS