Are social media posts effective on stock values?

Abstract

With the widespread use of the internet in recent years, individuals take into account other people’s comments in their decision-making processes about a good or service. Similarly, companies use online data to provide better service to their customers and to see their own strengths and weaknesses. In this process, sentiment analysis has become one of the most preferred methods for interpreting big data in recent years. The purpose of this study is to examine whether written posts on Twitter, one of the widely used social media platforms, have an impact on stock prices. In this sense, sentiment analysis of the written expressions shared on the Twitter platform with the hashtag #TSLA between January 1, 2016 and April 10, 2021 was performed with the BERT model, one of the deep learning methods, and it was determined whether the posts had positive or negative meanings. The prediction of whether the TSLA stock value decreases or increases depending on whether the comments are positive or negative on the day following the day of the post was performed with Neural Prophet. As a result, it is concluded that social media posts have an impact on individuals’ decisions to buy and sell stocks.