Retrospective Analysis of Controversial Subtopics on COVID-19 in Japan

TL;DR

This paper presents a framework for identifying polarized COVID-19 subtopics on Twitter in Japan, providing insights for policymakers by detecting controversial issues beyond traditional sentiment analysis.

Contribution

It introduces a novel framework for detecting controversial subtopics and applies it retrospectively to COVID-19 discussions in Japan, revealing key areas of polarization.

Findings

Controversial topics include government, medical, economic, and educational issues.

Controversy score is weakly correlated with traditional indicators like scale and sentiment.

The framework effectively detects real-world controversial issues.

Abstract

For efficient political decision-making in an emergency situation, a thorough recognition and understanding of the polarized topics is crucial. The cost of unmitigated polarization would be extremely high for the society; therefore, it is desirable to identify the polarizing issues before they become serious. With this in mind, we conducted a retrospective analysis of the polarized subtopics of COVID-19 to obtain insights for future policymaking. To this end, we first propose a framework to comprehensively search for controversial subtopics. We then retrospectively analyze subtopics on COVID-19 using the proposed framework, with data obtained via Twitter in Japan. The results show that the proposed framework can effectively detect controversial subtopics that reflect current reality. Controversial subtopics tend to be about the government, medical matters, economy, and education; moreover, the controversy score had a low correlation with the traditional indicators--scale and sentiment of the subtopics--which suggests that the controversy score is a potentially important indicator to be obtained. We also discussed the difference between subtopics that became highly controversial and ones that did not despite their large scale.