Sentiment analysis in Twitter on stock performance in Indonesia

This paper is aimed at investigating the correlation between social media Twitter and stock market performance by looking at industry-level perspective to specific companies incorporated in the. In the context of investment, Twitter has become an alternative source of information to answer the needs of investors to be able to keep abreast of news factually and practically. Therefore, we are interested in applying the same research objectives to the stock of several companies in Indonesia that are incorporated in the IDX30 index. This stock index is the illiquidity sub-category of the Headline Index, which contains companies that are used as a reference to describe the stock market conditions that meet the main criteria of having a high level of liquidity. To support this research, we collected some opinions obtained from Twitter as a source of streaming data using Python programming, and Thomson Reuters to obtain information on company performance in the form of stock prices, volumes, and market capitalization. Research models are built based on the Amihud Illiquidity method and volatility calculation to measure the correlation between sentiment analysis and stock performance.This research shows that sentiment analysis of statements uploaded on Twitter has insignificant correlation to the liquidity and volatility of IDX30 stock in Indonesia. Nevertheless, this research has not been able to separate between tweets which are generated based on user opinion and tweets which are made based on requests from certain market participants to influence the value of shares by spreading biased information to provoke a public reaction.