Sentiment analysis and topic modeling for COVID-19 vaccine discussions

Area of Science:

• Public Health
• Social Media Analysis
• Epidemiology

Background:

• The COVID-19 pandemic and its Delta variant have significantly impacted global daily life.
• Over 20 World Health Organization-approved vaccines offer hope against the pandemic.
• Social media discussions on COVID-19 vaccine effectiveness and safety are widespread but lack systematic analysis.

Purpose of the Study:

• To conduct an in-depth analysis of public opinion on COVID-19 vaccines on Twitter.
• To identify key discussion topics and analyze emotional polarity related to vaccines.
• To provide data support for policymakers regarding public sentiment on vaccination.

Main Methods:

• Utilized Natural Language Processing (NLP) technologies for social media data mining.
• Analyzed Twitter discussions concerning COVID-19 vaccines, considering countries and vaccine brands.
• Examined hot topics and emotional polarity within public discourse.

Main Results:

• The majority of public opinion expressed trust in vaccine effectiveness and a willingness to be vaccinated.
• Negative sentiment was frequently linked to reports of post-vaccination deaths, vaccine shortages, and side effects.
• Analysis revealed distinct discussion patterns across different countries and vaccine brands.

Conclusions:

• NLP analysis of social media provides objective insights into public opinion on COVID-19 vaccines.
• Findings can help clarify misinformation and support evidence-based policymaking.
• Understanding public sentiment is crucial for effective vaccine rollout and pandemic management.