Funded by the Finnish Foundation for Stock Promotion, Nordea Bank Foundation, and Savings Banks Research Foundation, 2021-2023.
Research group
Senior Research Fellow Kim Holmberg (team leader)
Systems Developer Olli Jalonen
Trainee Aki Raittila
Abstract
Advances in the ability to collect and analyze tremendous amounts of data has seen an increase in efforts to develop methodology to analyze social media conversations in attempts to forecast everything from the results of presidential elections (Liu et al., 2021) to Bitcoin price trends (Cavalli & Amoretti, 2021), and from traffic patterns (Yao & Qian, 2021) to spread of influenza (Wang et al., 2020). In attempts to predict stock market trends machine learning algorithms and neural networks have been developed to analyze and combine vast amounts of numerical data about for instance stock price history and textual data from different news sources (Zhai, Hsu, & Halgamuge, 2007; Nelson, Pereira, & de Oliveira, 2017; Oncharoen & Vateekol, 2018; Awan, 2021). By combining both technical indicators and data from media these approaches have in fact with high accuracy been able to predict whether the price of a specific stock will go up or down in the near future. Another approach in predicting stock markets is to use actual data about consumer behavior. Companies such as Robin Hood provide tools for commission-free investing in stocks and in return they get real-time information about customer behavior and data about investing patterns. However, the decision to invest in certain stocks often precedes by reading other people’s recommendations and participating in online discussions. Thus, identifying and analyzing such discussion can provide even earlier signals about stock market trends and of possible rising interest towards specific stocks. In late 2020 and early 2021 the stock markets saw a sudden and unforeseen rise of certain stocks, as for instance the price of GME (see for instance https://finance.yahoo.com/quote/GME/) rose from around $4 USD in August 2020 to around $20 USD in early January 2021 and peaking at almost $350 USD later in January 2021. The cause of this sudden rise in the stock price of GME was traced back to social media and specifically to a subreddit called WallStreetBets (https://www.reddit.com/r/wallstreetbets/). Members of the subreddit were able to influence stock prices by creating a global movement of people buying certain stocks, causing the stock prices to rise. The aim of this project is to create an early signal system that will identify such social media discussions that may have impact on stock markets.
