When an emergency such as an infectious disease or natural disaster occurs, a negative atmosphere will usually spread throughout society-increasing people's dissatisfaction and anxiety. Because of this, it is rather difficult to thoroughly investigate the actual situation. However, people can post sentimental comments on news sites, allowing for their attitudes either for or against the topics to be better observed. This study extracts the positive, negative, and neutral comments by using sentiment analysis. Then, the social atmosphere is visualized by calculating the approval rating of the comments. This methodology is demonstrated in articles regarding COVID-19. The large volume of comments about two topics, Go To campaigns and PCR tests, were analyzed by using ML-Ask to classify the comments into three categories: negative, positive, and neutral. The results indicate that the social atmosphere about the Go To campaigns tended to be negative.
