By, Lilian Jin
With 2.4 billion active users on Facebook and 1 billion on Instagram, social media is the largest source of online data on consumer behaviour. The power of social media data has profound benefits to both businesses and investors. Twitter, in particular, has become a key player in predicting stock market trends. How does it do this?
Each time we interact on social media and express our emotions and opinions, we create emotional data. We consume this information from each other, and in turn, our decisions are influenced by the sentiments and opinions of others.
Now take Twitter, a social beehive of 330 million active users that generates 500 million tweets each day. The high volume of real-time emotional data creates the perfect grounds for sentiment analysis. Data analysis firms like Dataminr and Social Market Analytics have shown how social media can be leveraged to obtain the earliest news even before Wall Street.
In 2013, Dataminr picked up the news leak of Blackberry’s collapsed buyout 180 seconds before Wall Street. Dataminr’s investment clients got a head start on shorting the stock. Last August, Social Market Analytics covered 400,000 Twitter accounts and was able to alert its clients that positive chatter was building on Apple just before billionaire investor Carl Icahn tweeted that he bought a large volume of Apple shares. Clearly, investors with access to social sentiment technology have an extra advantage.
So how do data firms scan and analyze vast, unorganized information?
StockPulse is a data company currently providing social media insights to Nasdaq’s US Market Surveillance team. StockPulse starts by scoring Twitter users based on reputation. Verified users like Warren Buffet and Elon Musk have a higher impact and therefore a higher score. Among these credible users, StockPulse then monitors their number of tweets sent, likes and favourites, mentions, and retweets. The same statistics are also monitored on all followers of the credible users. This compiles all network of accounts that have the most relevance to a particular industry or stock.
StockPulse also follows key words on social media such as “Merger”, as well as variations of the keyword like “merger & acquisition” and “m&a”. Other keywords include “IPOs”, “Going Private”, “Joint Venture”, and “Board Member Resignation”. StockPulse collects key words in English, Chinese and German, then creates an alert system for any rumours or events to the surveillance team.
Social Market Analytics creates algorithms that calculate Twitter criteria such as “averages, change, volume, volatility, dispersion of tweets and risk” to generate S-Scores, a quantifiable value indicating whether people’s feelings towards a company or stock are positive or negative. According to Social Market Analytics, 10.0% of the Twitter feeds they examine reveal investment opportunities, which is a significant number. Data provider Markit states that “positive social media sentiment stocks have shown cumulative returns of 76.0% compared to -14.0% from negative sentiment stocks.
The following are examples of tweets that heavily impacted the market:
Tesla shares soared 11.0% following Elon Musk’s tweet on taking Tesla private.
Kylie Jenner expressed negative sentiment towards Snapchat and its stock lost $1.3 billion.
Donald Trump’s tweet against Toyota caused a $1.2 billion drop.
The value of Twitter’s social media data to the investment community is undeniable. Twitter recognized this and made $47.5 million from licensing its data in 2012. In 2013, Bloomberg implemented its own Twitter feed tracker, monitoring tweets from Wall Street analysts, and alerting subscribers of positive or negative social media chatter on any chosen company.
Featured image by Robin Worrall.