Visual Analytics for Comment Analysis (in Korean)

Abstract

A comment section, such as for an online news article, video, or social media post, is commonly a place to share personal opinions. However, when there are too many comments, it is challenging to gain new insights and estimate the distribution of the public opinions simply by reading them. Selecting and promoting high-quality comments can mitigate those problems, but the required resources for choosing them manually are too high for common practice. In this talk, I will introduce my research on visual analytics for comment analysis. The ability to see an overview of the comments and to create custom rankings supports moderators, editors, and commenters themselves. I also propose a method to aid semantic analysis where users can build custom dictionaries for unique concepts and use them to analyze comments. In the future, assessing public opinions will play an important role in protecting the digital democracy against organized attempts to manipulate public opinions.