ANALYZE SENTIMENT TWITTER OPINION FOR PUBLIC FIGURE USING NATURAL LANGUAGE PROCESSING TEXTBLOB

Abstract

Social media is an a platform that places for peoples to share everything. From text, photos, videos, image, voices, etc. Therefore, as a user is given the power to express anything to the platform. One of the social media apps that is on the rise besides Instagram and TikTok is Twitter. This media platform for sharing and expressing opinions is currently in great demand by Indonesians. Survey shows, that the average global daily Twitter users in 2019 increased by 21% year-on-year. But for Indonesia, this increase is 3.5 times above the global figure. Behind the increasing number of Twitter users in Indonesia, more and more users are tweeting and expressing their opinions[4]. Starting from sentences. Positive, neutral, to the worst are negative sentences. Indicates the wider and more information is spread and In Indonesia, reading habits are still very low, from the existing data taken from the PISA survey released by the OECD in 2019, Indonesia's literacy rate is in the bottom 10 of 70 countries. From that case, people can use it to find out the percentage of positive sentiment and negative sentiment towards public figure from tweet in Twitter to identify and dig deeper into the motivation behind the user's message to see if it includes complaints, suggestions, opinions, questions or even appreciation of the public figure. To solve this problem, author have proposed this final project to make an application tracing public sentiment base on Twitter using natural language processing method based on artificial intelligence that is TextBlob library and tweet that refer to analyse sentiment of public opinion to public figure. This web-based application analyze sentiment will be accommodate to help marketing analysis, product reviews, product feedback, and community service to know the stigma and have beneficial from the data. Sentiment analyze application is more focused on how the public figure seen from Twitter users as correspondents for giving opinions to the public figure. This application provides an analyst one hundred tweet in the past one week refer to the name of that public

figure, shows the sentiment word refer from the public figure, percentage in each sentiment, decided the attitude of the public figure and data statistic in the form of pie chart and bar chart. The data of sentiment can be store in database and someday can be accessed again as a history page.