[Paper Review] Detecting Topic and Sentiment Dynamics Due to COVID-19 Pandemic Using Social Media

Hui Yin, Shuiqiao Yang+1 more|arXiv (Cornell University)|Jul 5, 2020

Misinformation and Its Impacts25 references14 citations

TL;DR

This study proposes a dynamic topic modeling and sentiment analysis framework using 13 million Twitter tweets from April 2020 to track global topic and sentiment shifts during the early COVID-19 pandemic. By combining Dynamic Topic Models (DTM) and VADER sentiment analysis, it reveals that while overall sentiment was positive, topics like 'stay home' showed strong positivity, whereas 'people death' consistently carried negative sentiment, highlighting topic-specific emotional dynamics.

ABSTRACT

The outbreak of the novel Coronavirus Disease (COVID-19) has greatly influenced people's daily lives across the globe. Emergent measures and policies (e.g., lockdown, social distancing) have been taken by governments to combat this highly infectious disease. However, people's mental health is also at risk due to the long-time strict social isolation rules. Hence, monitoring people's mental health across various events and topics will be extremely necessary for policy makers to make the appropriate decisions. On the other hand, social media have been widely used as an outlet for people to publish and share their personal opinions and feelings. The large scale social media posts (e.g., tweets) provide an ideal data source to infer the mental health for people during this pandemic period. In this work, we propose a novel framework to analyze the topic and sentiment dynamics due to COVID-19 from the massive social media posts. Based on a collection of 13 million tweets related to COVID-19 over two weeks, we found that the positive sentiment shows higher ratio than the negative sentiment during the study period. When zooming into the topic-level analysis, we find that different aspects of COVID-19 have been constantly discussed and show comparable sentiment polarities. Some topics like ``stay safe home" are dominated with positive sentiment. The others such as ``people death" are consistently showing negative sentiment. Overall, the proposed framework shows insightful findings based on the analysis of the topic-level sentiment dynamics.

Motivation & Objective

• To monitor real-time topic and sentiment dynamics related to the COVID-19 pandemic using social media data.
• To understand how public sentiment evolves in response to different pandemic-related topics.
• To support policymakers by identifying emotionally charged topics and sentiment shifts during public health crises.
• To develop a scalable framework for dynamic topic and sentiment analysis on large-scale social media data.

Proposed method

• Collected 13,746,822 COVID-19-related tweets from Twitter between April 1 and April 14, 2020.
• Applied the Dynamic Topic Model (DTM) to discover daily evolving topics from the tweet corpus.
• Used the VADER sentiment lexicon to classify sentiment polarity at both tweet and topic levels.
• Ranked topics by volume and analyzed sentiment trends over time for top topics.
• Visualized topic evolution and sentiment dynamics using time-series charts and word clouds.
• Evaluated topic coherence and sentiment distribution across different thematic clusters.

Experimental results

Research questions

• RQ1 How does public sentiment dynamically shift during the early phase of the COVID-19 pandemic?
• RQ2 Which topics related to COVID-19 are most frequently discussed on social media?
• RQ3 How do topics evolve over time in response to pandemic developments?
• RQ4 How do sentiment polarities vary across different pandemic-related topics?

Key findings

• The overall sentiment across the 13 million tweets showed a higher proportion of positive sentiment compared to negative sentiment during the study period.
• Topic 11, associated with 'stay home' and self-isolation, consistently exhibited dominant positive sentiment, reflecting public support for protective measures.
• Topic 49, related to case reports and official data, showed predominantly negative sentiment, indicating public anxiety over rising case numbers.
• Topic 64, focused on deaths due to COVID-19, maintained consistently high levels of negative sentiment, reflecting public frustration or grief over mortality.
• The top three topics—11, 49, and 64—remained in the top 10 daily discussions throughout the two-week period, indicating sustained public concern.
• Despite global positivity, sentiment varied significantly by topic, with 'people death' and 'case reports' showing strong negative sentiment, while 'staying home' was widely perceived positively.