Trading Twitter | Using Twitter to Predict the Stock Market | The Twitter Fund
Predicting the direction of the stock market is an extremely difficult endeavour. In fact, no method or formula has ever consistently predicted the market’s direction. Many people have claimed to have discovered a magic formula or a secret indicator, such as the Superbowl indicator, or sell in May and go away, only to later determine their hot streak was attributable to little more than just plain luck. However, perhaps that is all about to change because of....
Huh? Twitter Can Predict the Stock Market?
According to a study published by a group at the Indiana University School of Informatics and Computing, lead by computational social scientist Johan Bollen, the movement of the Dow Jones Industrial Average (DJIA) can be predicted by analyzing the millions of daily tweets submitted to the social media site Twitter.
The idea to use social media as a predictive instrument is not new. In fact other studies have shown that blogs and Twitter feeds can be used to analyze book sales and predict movie ticket sales at the box office. Bollen writes, “ We know from psychological research that emotions, in addition to information, play a significant role in human decision-making. Behavioral finance has provided further proof that financial decisions are significantly driven by emotion and mood.”
How Does the Twitter Indicator Work?
In their research Bollen and Mao used two tools. First, OpinionFinder, a freely available program that analyzes and rates the general level of positive or negative feelings in social data based on words with an emotional tone. Only tweets that contain explicit statements that express the author’s mood such as “I am feeling”, “I am,” and “makes me” were included in the analysis. Second, G-POMS, a combination of Google gathered data with a Profile of Mood States, takes the positive or negative groups of data gathered by OpinionFinder and routes them into groups of different mood dimensions such as calm, alert,sure, vital, kind, and happy. The group analyzed 9.8 million tweets submitted to Twitter from Feb 28, 2008 to Dec 19, 2008. Each mood dimension was compared to the DJIA and examined for any trends. The study concluded that the mood “calm” was positively correlated to the direction of the DJIA. Let’s take a look at the actual data:
Data from Bollen's Paper
Remeber that the red “Calm score” is lagged by 3 days. You can see that when the Twitterverse was projecting an increasinly “calm” mood state, 3 to 5 days later the DJIA typically rallied. When the calmness of tweets lowered, two to six days later the Dow would also decline. Therefore, the team believes they have discovered an oddly correlative relationship between public mood and the DJIA. In fact, they conclude, “We find an accuracy of 87.6% in predicting the daily up and down changes in the closing values of the DJIA”.
A stock market indicator that is accurate 90% of the time! Holy sh*t!!!! KA-CHING!
Naturally, Bollen and company made a lot of friends very quickly.
The Twitter Fund
Would put your money into a fund that invests solely on the basis of public mood determined by Twitter?
The “Twitter Fund”
The Twitter Indicator recently caught the attention of twenty-eight year old start up hedge fund manager Paul Hawtin. Hawtin manages the soon to be open Derwent Capital fund or "Twitter fund" near London. The fund plans to scan Twitter's millions of daily tweets and look for threads that display public sentiment that can be used to predict the market using the strategy discussed above. In fact, Hawtin believes he can generate 15 to 20 percent returns.
Only time will tell if Bollen and Hawtin can live up to the hype they've generated. Of course the Twitter Indicator has its skeptics, including myself. For now I'll stick to buy and hold, dividend, and options investing, but I'll be sure to keep a close eye on the Twitter Fund because maybe, just maybe, these guys have stumbled across a magical key to unlocking the predictive power of the stock market!
Let me know what you think about the Twitter Indicator and the fund in the comments section below. And don't forget to follow me on Twitter @Enni82.
- Hedge Fund Bets $40 Million That Twitter Can Predict The Stock Market
Last October, Johan Bollen and Huina Mao, professors of informatics and computing at Indiana University-Bloomington, caused a stir in the business world when they said Twitter could be used to predict the Dow Jones. Paul Hawtin, a 28-year-old hedge f
- Twitter mood predicts the stock market
Twitter mood predicts the stock market. Johan Bollen,Huina Mao,Xiao-Jun Zeng
Follow your friends and share you life online with Twitter.
- Johan Bollen Interview on Twitter, Stock Market (Oct. 20) - Video - Bloomberg
Johan Bollen, an associate professor of informatics and computing at Indiana University, talks with Bloomberg's Lisa Murphy about his research into sentiment on Twitter Inc.'s social-networking site and how that correlates to stock market performance
- Your Tweets Could Be Worth Millions - CNBC
Starting in February, a group of very bold hedge fund managers are launching a multi-million dollar hedge fund whose strategy relies on one very unusual market indicator: your Twitter account.
- \'Astounding\' Trading Accuracy Using Twitter: Fund Manager - CNBC
Kate Kelly goes one-on-one with the two key players behind an ambitious new UK hedge fund launching in early April, which is promising its investors outsized returns using millions of random tweets to predict changes in the stock market.
- Hedge Fund to Predict Markets Using Twitter - CNBC
Last fall, Indiana University informatics professor Johan Bollen stumbled upon an astonishing connection: That the social network Twitter could predict swings in the Dow Jones Industrial Average with 87 percent accuracy.
- How Tweet It Is!: Library Acquires Entire Twitter Archive Library of Congress Blog
The library of Congress has EVERY tweet ever posted saved. This data can all be accessed to discover more about human psychology and determine if the trading Twitter trends is really relevant.