Over the last decade, Information Technology (IT) has improved dramatically. This caused change in almost every profession in every industry. In this blog post, I would like to go more in-depth into the changes IT caused in stock prediction, and then particularly about the influence sentiment analysis has and how IT has enabled this technique to be more accurate. Sentimental analysis can make use of various techniques to analyze the people’s thoughts and opinions and relate this to the stock price movement. Examples of these techniques include (but are not limited to): text mining, natural language processing and artificial intelligence and then in more particular machine learning (Hughes, 2017). Obtaining relevant data for this analysis has become easier. This is because of the rise in popularity of Social Media platforms. A new method emerged to analyze the connection between the public emotion and the stock trend by analyzing the data on these networks. In this blog post, one particular social media platform is chosen, namely ‘Twitter’.
Sentimental analysis is important in the field of behavioral finance. It is believed that the only factor that matters is the human emotion, being the sole driver of the stock price (Blog.twitter.com, 2017). By using one of the above mentioned techniques, the words that make up sentences can be translated into data that is understandable by computers. Each word is given a sentimental value between the two spectrums positive and negative. Advanced techniques are needed to increase the chance of correctly understanding phrases: combinations of different words.
Twitter is a growing social media platform and often used in a sentimental analysis. The reason for this is because there is a vast amount of real-time market conversation on a daily basis (Blog.twitter.com, 2017). To the question: “Is there really any correlation between Twitter and the stock market price?” there is the following answer (example): In 2013, the Associated Press’ Twitter account was hacked and posted the fake statement that Barack Obama had been injured in an explosion at the White House (Matthews, 2017). According to the Financial Times, this single tweet caused a shock in the stock market. The S&P 500 declined with one percent equaling $130 billion in stock value in just a matter of seconds. So to come back to the previous question whether there is a correlation between Twitter and the stock market price: definitely yes.
So for the investors and traders among us, the question is of course: “Is it wise to rely on sentimental analysis that is based on just Twitter data? “. Out of personal opinion, I am convinced that sentimental analysis based on twitter output is reasonably indicative for predicting the stock price. Of course, this also strongly depends on the technique and model that is used for the sentimental analysis. However, by incorporating multiple techniques to eventually come to your conclusion is more likely to be correct.
Hughes, L. (2017). Machine Learning With Heart: How Sentiment Analysis Can Help Your Customers. [online] Digitalist Magazine. Available at: https://www.digitalistmag.com/customer-experience/2017/10/18/machine-learning-sentiment-analysis-can-help-customers-05429002 [Accessed 22 Oct. 2017].
Blog.twitter.com. (2017). Twitter Data and the Financial Markets. [online] Available at: https://blog.twitter.com/official/en_us/topics/insights/2016/twitter-data-and-the-financial-markets.html [Accessed 22 Oct. 2017].
Matthews, C. (2017). How Does One Fake Tweet Cause a Stock Market Crash? | TIME.com. [online] TIME.com. Available at: http://business.time.com/2013/04/24/how-does-one-fake-tweet-cause-a-stock-market-crash/ [Accessed 22 Oct. 2017].
Interesting post, My bachelor thesis was about the relationship between investor sentiment and stock returns. I used the proxy of consumer confidence as a proxy for investor sentiment. Consumer confidence is measured by consumer surveys but could also be measured by tweets.
In my research, I found that there is a significant positive relationship between consumer confidence and stock market returns. I think it would be interesting to do this research again with the investor sentiment being measured by tweets. I think this would definitely show a strong relation, Although it must be said that one tweet of Trump has a different effect on the stock market than when I would tweet something, this is something that must be balanced out correctly.
Hey Joey, yes – the White House example clearly shows that Twitter post can cause a very immediate and relatively large reaction since Twitter is a frequently used and also well-trusted medium of exchange. However, I think your question was more about whether Twitter can predict longer-lasting trends than just immediate reactions. My answer is no – the reason is that the actors, who influence the market are not the ones that share each and every of their sentiments on Twitter. Why has there been such a drastic reaction to the White House posting – well, the information has been completely new and nobody was aware of it (because it was fake). Basically every piece of information on Twitter that is real will already be known by the traders with the big money (hedge fund managers, insurance companies, you name them) and will already be incorporated in the market – so eventually, technical analysis (extrapolating historical stock market data) will be the better alternative after all. Of course, there is one exception to proof the case – Donald Trump. As everyone knows, this guy is unpredictable and he shares his thoughts on Twitter before anyone else has access to them. In conclusion, being the first one to trade on Donny Ts posts may give you a good profit – for stock market data, the big guys and technical traders will always be a step ahead of you. If you wanna trade on sentimental analysis after all, it may be worth it to go with Bitcoins instead – this market is hardly covered, there is no institutional money and the price is driven by actors that are active on social media (and current returns in cryptocurrencies are superhigh anyway).