Combing Social Media for Trading Ideas

Terry Flanagan

The ‘twitterverse’ is a potential source to learn of market-moving events, but making trading decisions on Twitter sentiment can be risky because of the possibility of misleading or false tweets.

In order to sift true indicators from false ones, social media analytics company Contix algorithmically identifies actionable social media alerts from the entire universe of Twitter users, then confirms that the news has not broken first on traditional media, allowing users to easily identify the most differentiated and valuable alerts.

“Our idea was to develop a platform where we would ingest the entire ‘fire hose’ of social media content and then cluster that with traditional media content in order to identify the very earliest mentions of breaking news,” Ryan Bailey, co-founder and CEO of Contix, told Markets Media. “There are citizen journalists. There are reporters that break news on Twitter at the same time as their news organization. When we’ve identified a news alert that is actionable, valuable, and relevant to the financial markets, we send that via an alert to our users who can access that either via web app, desktop app, mobile app or e-mail.”

In its most recent suite of enhancements, Contix has expanded to mine over 500 million social media and over 500 traditional media sources each day. Taking in Twitter and traditional media updates, the platform uses natural language processing and statistical analysis to surface events algorithmically.

Photo via Contix

Photo via Contix

“Our focus is simple,” Bailey said. “We identify breaking news events earlier than other sources, so traders use our product to make sure that they are not missing out on breaking news. If they don’t have our product and they aren’t using social, they are going to miss out on breaking news events. They’re going to be late to trades.”

For example, on the day Twitter’s earnings were broken early, the Contix system identified the first social media post, created an alert based on that post and sent it to users.

“It was clear from that post that they had missed their revenue estimates,” Bailey said. “They were going to come in below expectations and some of our users placed trades based on that alert. The stock subsequently dropped 17% through the close and then further the next day, and our users had the earliest access to that information via our system.”

Social media analytics is still in its early days in terms of its application to algorithmic trading. “Tree years ago when it first got started, there was a growing awareness that social is important, that event detection was important, that it can be used for sentiment, but there’s still not widespread understanding of the value of social,” said Bailey. “A lot of traders and investors need to be educated about the value of social for their investment processes.”

The customers that Contix works with, both institutional and individual, use it to supplement their other breaking news applications. “They might have Bloomberg. They might have another traditional media source that they’ve gotten used to using that provides that certain other fundamental data, and then they will incorporate our product to provide additional breaking news and trade off our alerts on a daily basis,” Baileys said. “We see the value of social and ours is a very easy to use and understand product for deriving the value from social.”

Feature image via Dollar Photo Club

Related articles

  1. Allowing algorithms to control the supply and demand of crypto was challenged by Terra Luna.

  2. CME Boosts European Energy Business

    Infrastructure and sustainability will be one of the biggest opportunities in alternative investments.

  3. The broker's algo development team gains access to high quality, granular Level 3 data.

  4. Trading Europe From ‘Across the Pond’

    Bank aims to gain share from its systematic approach to trading, market structure proficiency and modern techn...

  5. The firm will be rolling out algos for institutions to navigate crypto market structure.