[ibimage==33201==Large==none==self==ibimage_align-center]
The day breaks in New York City and thousands of stock brokers and traders bustle through the narrow, windy roads of the Financial District to get to 11 Wall Street: the New York Stock Exchange. Before the Opening Bell rings, many of these financiers have already started their day with a cup of deli coffee, a street cart bagel and a side of insight from Chicago-based startup, Social Market Analytics.
Over the past two years, Social Market Analytics as taken “Big Data” from Twitter and whittled it into two morning e-mails for traders and investors alike. The technology behind these messages consistently predicts the movement in specific stocks - so accurately that the NYSE Euronext and Markit have adopted the service for their members.
“The idea for Social Market Analytics came to me after repeatedly being approached by several buy-side investors for my opinion of social media’s impact on the capital markets,” CEO and co-founder Joe Gits said.
Gits wasn’t new to big data and the influence it could have on market researchers and traders. Previously, he co-founded Quantitative Analytics, Inc., an integrated database solution that was integrated into Thomson Financial - now Thomson Reuters.
The ability social media had to move markets was known by hedge funds and traders but there was no way of sifting through the 30-million daily tweets about security and security prices and returns. “There was no way to clean and quantify this data in any meaningful way,” Gits said.
So in 2011 Gits partnered with world-class mathematicians to create a system that analyzed and filtered big data from Twitter into statistically powerful intelligence. This “clean” information was directed to both buy-side and sell-side traders to predict market movements.
The Naperville-based company company officially launched in 2012 and in January 2014, reported a total of $410,000 in partial funding.
The Proof
Gits and his team surfed the markets for 760-days and analyzed data from December 1, 2011 to December 31, 2013. During this period, they developed a system that tested day trades from the “Universe of stocks” and scored tweets each day by highest positive social sentiment, an “S-Score > +2”, and by lowest social sentiment, an “S-Score < -2”. “Day trades” are speculations in security - specifically buying and selling of tradeable assets (like cash, bonds, stocks, equity, etc.).
[ibimage==33204==Original==none==self==ibimage_align-center]
Overall, Social Market Analytics found the ability to predict stock behavior based on the positive and negative sentiments of day trades. "Market Sentiments" gauge the overall attitude of investors toward a particular security or stock. Looking at all of these sentiments is considered insight into a markets "crowd psychology". Because Social Market Analytics assessed conversations happening directly in security markets, they were able to make solid predictions on how the index would fluctuate. For instance, S-Scores less than -2 returned 4.38 percent, underperforming the S&P index while S-Scores greater than +2 outperformed the index at 54.78 percent.
The system also revealed information about Sharpe Ratios. Generally speaking, the ratio identifies whether the return on a portfolio is due to smart investment decisions or the result of excess risk. Social Market Analytics discovered that stocks with positive S-scores generated better returns when compared against the S&P 500 (as measured by the returns of SPY, a popular index fund).
The Process
All of this data is processed through three stages of patent-pending technology. The results of this are clean, intelligent data that Social Market Analytics refers to as the “Secret to Quantifying Twitter’s Predictive Power”. The three stages are as follows:
The Extractor:
This process assesses the millions of tweets cast each day and pulls an estimated 30-million that mention financial terms and symbols.
The Evaluator:
This process filters out about 90 percent of those 30-million tweets. The surviving 140-character-or-less messages are analyzed by “Accounts” (measured qualitatively and quantitatively) and “Tweets” (assessed by content).
The Calculator:
“Our algorithm produces a family of seven metrics, called S-factors, that help active traders quickly understand the validity, importance and pervasiveness of that sentiment against historical values,” said Gits.
[ibimage==33198==Original==none==self==ibimage_align-center]
The predictive power comes into play by assessing the overall tone of social media conversations. “Positive S-Scores are associated with favorable changes in investor sentiment and statistically lead to higher prices, while negative levels are associated with unfavorable changes and statistically lead to lower prices,” said Gits.
The Product
Currently, Social Market Analytics users are active day traders. The service is subscription based, offering four tiers of access with prices ranging from free to $99/month. At minimum, a user will receive one e-mail when the market closes that outlines the days highest and lowest S-Scores. The top service features customizable reports that allow a user to track and get up-to-the-minute information on 50 chosen stocks. This quarter Social Market Analytics will begin to offer an “Excel Add-In”.
And while Wall Street hasn’t been quick to adopt social media insights into useable data, it’s now seeing the benefit. Said Gits, “Twitter has become a leading indicator of stock movement both up and down. By filtering for the intentions of professional investors you can create statistically significant actionable intelligence.”
.jpg)
.png)
