Exploring Financial Microblogs: Analysis of Users' Trading Profiles with Multivariate Statistical Methods

StockTwits is a Social Media focused on finance that is receiving increasing attention from finance experts and enthusiasts. In this work, StockTwits’ users are studied considering some of their self-declared characteristics, such as trading experience, holding period of the stocks, and trading approach. A Correspondence Analysis is carried out to investigate the relationships among these characteristics, the Simple Correspondence Analysis is applied to study the relationships between the approach and the holding period. The association between these variables and the experience is studied with the Multiple Correspondence Analysis. In the end, a cluster analysis carried out with hierarchical clustering is used to study the structure of the StockTwits community on the basis of the same characteristics. The analyses highlighted that the way users label their own approach and primary holding period reflect the objective relation linking technical strategy with short-term investments and fundamental approach with long-term ones. Moreover, it showed a weak relation of the experience in trading with the other features, configuring self-reported experience as a more cross-sectional characteristic.