The stock market is complex and to some degree unpredictable, each stock is constantly being affected by the economy and the market itself. An example of an unexpected event, also known as a “Black Swan” is the 2008 financial crisis that literally broke the market. The Dotcom bubble is another example in which venture capitalists fueled by market overconfidence and pure speculation started investing in any company with a .com after its name, not knowing what was coming ahead.
These types of events are extremely hard to predict. Using technical or fundamental analysis alone isn’t enough when analyzing the financial market. These traditional methods rely on the assumption that traders make decisions based on rationality and objectivity. Which is not true, humans are not robots, we have feelings and emotions.
This is where the problem lies, the market is affected by psychological factors everyday and a lot of traders are losing money due to that misunderstanding.
Since the advent of social media, the way we connect and interact changed tremendously. One of the main features of social media that distinguishes itself from other means of interaction is its information diffusion speed. Everything is fast, everyone likes and shares, be it their new socks that they love or that new pair of headphones they regret buying. New trends and hypes are made in a matter of a day.
Twitter is one of the biggest social medias out there with approximately 8.000 tweets being sent each second, with the majority of its users using it from their mobile devices. Twitter has a cap of 140 characters per tweet, which makes it a great source of information about what its users are “tweeting” about in a quick concise way without too much fluff.
Which brings us to sentiment analysis, also known as opinion mining. A relatively new sub-field of natural language processing which extracts and quantifies affective states from text. Being part of Web 2.0, interest for it grew following the rise of blogging and social networks like Twitter.
As a study from the Technical University of Munich states:
“Even though short text strings might be a problem, sentiment analysis within microblogging has shown that Twitter can be seen as a valid online indicator of political sentiment. Tweets’ political sentiment demonstrates close correspondence to parties’ and politicians’ political positions, indicating that the content of Twitter messages plausibly reflects the offline political landscape.”
(Source: Tumasjan, Andranik; O.Sprenger, Timm; G.Sandner, Philipp; M.Welpe, Isabell (2010). “Predicting Elections with Twitter: What 140 Characters Reveal about Political Sentiment”. “Proceedings of the Fourth International AAAI Conference on Weblogs and Social Media”) (Full study: https://www.aaai.org/ocs/index.php/ICWSM/ICWSM10/paper/viewFile/1441/1852)
Let’s take the stock market for instance, some studies noted that a significant correlation exists between the rise and fall in stock prices of a company and the sentiments and opinions expressed about that company on Twitter. Since Twitters caps each tweet at 140 characters, identifying, extracting and quantifying sentiments from tweets is fast and easy on resource for computers. Using those sentiments and coupling them with other types of analysis would prove useful only if they relate to each other significantly.
Twitter is a key platform for sentiment analysis, with its sheer amount of information about its users emotions and opinions in form of tweets, it’s certainly not a media to be ignored. Be it for stock traders or marketers, anyone can get something out of it, especially if used with other mainstream analysis techniques.