Indonesia COVID-19 Online Media News Sentiment Analysis with Lexicon-based Approach and Emotion Detection

Abstract

Social distancing and isolation are one of the impacts of COVID-19 pandemic, which lead to the increase of internet users across the country especially in suburbs area. Consequently, news regarding COVID-19 reported by the media particularly online media would reach extensive masses. One of the concerns pertaining to this issue is the sentiments and emotions evoked by COVID-19 news, which those sentiments and emotions are crucial in shaping perceptions and attitudes of the public about COVID-19. Therefore, understanding the sentiments and emotions caused by COVID-19 news would help the public more aware in the process of seeking information through online media news. This research used more than 19.000 COVID-19 headlines from known and popular online media starting March 2020 (based on public health authority announcements) to March 2021. NRC Emotion Lexicon used to detect sentiments and emotions from the headlines. The result shown from the analysis stated that 40% of all the headlines evoked negative sentiments. The first seven months were dominated by negative sentiments. Although, at the end of the 2020 positive sentiment started increasing gradually. Sadness, Fear, Trust, and Anticipate were the most dominant emotions evoked by COVID-19 news. The high negative sentiment has no correlation with death-per-million because of COVID-19 in Indonesia.