Going to the store to purchase a newspaper, grabbing a pen, and identifying the latest trends in the market is pretty much an extinct practice at this point. This is no secret especially to the young millennial investors of todays world. They currently play a very pivotal point in who the next big distributors of investment market information will be.
Twitter’s Mood-based Research Algorithms
In this MIT Technology Review article derived from information within the Cornell University Library of Computational Engineering, Finance, and Science; the analysis of the research on a database of 9.7 million tweets from 2.7 million twitter accounts states how/why tweets can be used to predict stock market movements up to 6 days in advance. One algorithm, known as the Google-Profile of Mood States (GPOMS) records the level of 6 emotional states in twitter activity: happiness, kindness, alertness, sureness, vitality, and calmness. The researchers found that there was a correlation between the calmness indicator and the Dow Jones Industrial Average with an accuracy of 87.6% in predicting the daily up and down changes in the closing values between 2 and 6 days in advance.
There are multiple theories as to why this could be accurate. First is the “Informational Theory”, the theory that because the information published on Twitter is new, it has not yet been incorporated into stock prices. Thus, it is extremely rational to expect for this information to effect future stock prices. An example of this taking place would be November 11, 2013, when a Canadian newspaper leaked that the 4.7 billion dollar buyout of Blackberry had collapsed. Users of Dataminr, a data-analysis technology service were alerted within seconds about this information and were able to short the stock a full 180 seconds before Wall Street was able to (because the information was leaked from a small party, rather than being officially released by big-name news sources).
The second is the “sentimental theory”, the price of an asset deviating from its current market value based on its intrinsic value. At around noon on March 30th, 2015 Elon Musk tweeted this out and within minutes Tesla’s capitalization increased by approximately 1 Billion dollars. A somewhat similar story happened on August 13th, 2013 when Carl Icahn’s tweet boosted Apple’s stock price by 3% in a matter of minutes.
Founded in 2009 by three Yale roommates, Dataminr was ranked the 13th most disruptive private company by CNBC in 2014. The company specializes in identifying potential breaking news while still in its early stages. The software does this by implementing machine learning to cross-reference tweets with over 30 data sets pertaining to things like location, stock movement, and patent data to identify clusters of similar tweets that the software deems an usual pattern.
With over 75 major financial firms as clients at the end of 2015, this company is one of the very few that are approved by Twitter to have full access to Twitter’s entire stream of tweets. It uses powerful algorithms to detects patterns and can quickly identify if news is real or fake in real-time. It also determines which tweets are most appropriate for users before being notified.
Determinr began revealing reports of Volkswagen’s emission scandal three days before it’s stock plummeted 30%. It alerted all oil & gas traders about the death of the king of Saudi Arabia approximately 4 hours before the crude prices spiked on the news. These are just some of the countless examples of when this software has alerted its users on information ahead of when the implications of the news has effected markets.
Watson Financial Services
IBM’s Watson Financial Services product has also shown promise for the future. See my last blog for an explanation of how the product works. In a recent study, the software was able to predict overdrafts a week ahead of time with 94% accuracy as well as an 87% accuracy for overdrafts 4 weeks ahead of time. The software also is currently able to accurately predict cancellations and non-renewals in insurance companies with a 78% accuracy.
Stocktwits is a social media site that allows users to view and share information amongst the financial and investment community. Users can subscribe to other users, put specific stocks on their watch list, see what stocks are “trending” (biggest gainers and losers of the day) and many other features. Regardless of credibility or credentials, any user can post on the website as long as the post is under 140 characters, similar to twitter.
Despite the reputation of being the “me, me, me” generation, millennials love to share – on social media anyways. They’ve all been raised in a way that it is second nature to them to share countless bits of information whether it be a picture on Instagram, or an opinion on Facebook or Twitter to their hundreds, sometimes even thousands of the peers via social media. This allows millennial investors another platform to do this as they please.
Millenials want quick and easy. They’re famous for not caring about in-depth research or spending a lot of time trying to find answers so it makes sense for them to gravitate towards this platform with nothing but quick bits of information from users whether they’re credible or not as well as being able to know the hottest stocks of the day by maneuvering through the app for 15 seconds. They’re so used to having everything at their fingertips that spending extended amounts of time to find information is something that they both dread and are simply just not accustomed to.
With over 300,000 investors publishing content to the website to an audience of over 40 million people, the app continues to grow.