Researchers have outperformed the Dow Jones Industrial Average with a trading model built on data culled from Twitter.
The team - led by Vagelis Hristidis, associate professor, Bourns College of Engineering at the University of California Riverside (UCR) - set out in search of a correlation between activity on the microblogging site and changes in prices and trading volumes of stocks.
The researchers obtained the daily closing price and the number of trades from Yahoo! Finance for 150 randomly selected companies in the S&P 500 Index for the first half of 2010. Then, they developed filters to select only relevant tweets for those companies during that time period.
While there have been previous studies into Twitter and trading, these have mostly been concerned with tracking sentiment; whether the tweets are positive or negative.
The latest research concentrated on the volume of tweets and the ways that they are linked to other tweets, topics or users. The team set out to analyses 'connected components' - tweets about specific subjects related to the company, such as a product announcement or financial report.
Put to the test over a four month period in 2010, the model saw a 2.2% loss, outperforming the Dow Jones Industrial Average which was down 4.2% and several other models tracked by the team.
The more connected components related to a stock, the higher the volume of trades, while there was also some, albeit weaker, correlation with price.
Hristidis told UCR Today: "These findings have the potential to have a big impact on market investors. With so much data available from social media, many investors are looking to sort it out and profit from it."
Read the paper:
Download the document now 314.9 kb (PDF File)