SENTIMENT ANALYSIS OF TWITTER MESSAGES IN COVID-19 PROCESS

Abstract

Sharing changes in people's daily lives through social media channels ensures that valuable information is accumulated and important inferences can be made with this information. The dose of the reaction to any information or sharing via social media, ensure to easily access and decide detailed information about people. Analyzing people's views, evaluations, attitudes, and emotions from the language in which they write, the workspace, which is also described as image analysis, is defined as sentiment analysis. Opinions of people are also important in important social events. The emotional states of individuals also vary during the Covid-19 process, which has influenced the whole world since January 2020. In this study, it was tried to measure the effects of applications on the people with emotion analysis method in the corona virüs process. Since 11 March 2020, highlights spoken to the corona virus in Turkey, which publishes announcements of people and official institutions as text or video via the Twitter social media channels have been identified. According to the sentiment analysis method, an overview was made by classifying the positive and negative comments for the topics collected under 5 headings, and then, whether there was a visible change in the weekly period regarding these issues was analyzed. For the opinions grouped as positive and negative on Twitter data, logistic regression analysis method, which is accepted in the literature and which can be obtained fast results, has been used on approximately 2,000,000 tweets.