This week, in the lead-up to the AUT Postgraduate Research Symposium on Friday, we are putting the spotlight on some of the talented and hard-working researchers who will present their work at the Symposium. Keep an eye out for more stories about our AUT research students throughout the week!
Ni Yang is a financial analyst for the real world.
While studying for his Masters in Finance and Investment at the University of Nottingham, he noticed that people were relying less on traditional news media like newspapers and television, and more on fragments of news received via social media. With his finance training, he realised that this practice of absorbing news-by-browsing would affect the way that people make investment decisions. This drove him to pursue doctoral study at AUT exploring the relationship between social media and stock prices.
As he puts it, “finance and investment decisions are more than just numbers, formulas and money, but also about overcoming the flawed human nature.” If our human nature leads us to check Twitter multiple times per day, what effect will that have on financial decision-making… and ultimately, on stock prices?
To answer this question, Ni worked with a massive sample of 2.2 million tweets. He used Natural Language Processing to scan and categorise each tweet as ‘bullish’ or ‘bearish’ (i.e. describing increasing or decreasing stock values). He then used a modified vector autoregression model to determine how tweets were affecting stock returns minute-by-minute.
His results showed that the more tweet activity there is around a given stock, and the stronger the sentiment in those tweets, the higher the price impact of trades. Those short 280-character missives can lead to real changes in the stocks are priced; to put it in the language of metaphor, these big bulls and bears can be moved by a little tweety bird.
Not only do Ni’s findings contribute to the nascent literature on social media in financial markets, they can also have implications for anyone involved in the stock market. “I hope [my research] has a social impact in the real world. Practically, it would be fantastic to help people have a dialectical attitude towards information and sentiment from social media while making investment decisions.”
You can hear more about Ni’s research at the AUT Postgraduate Research Symposium this Friday. His talk will take place at 2:25pm as part of the ‘New insights from big data & modelling’ session. Register here.