Data Science in Financial Markets: Characterization and Analysis of Stocktwits

Abstract

Online social networks provide a bunch of useful information that can help to solve different problems. In this context, we present a data characterization and analysis of Stocktwits, a financial online social network, in order to get insights and views that can be applied to financial markets and algorithmic trading (e-commerce). Furthermore, we consider feelings information in messages to create a social indicator, which can be used with a prediction model to support decisions as a strategy for operating in stock markets. Our characterization reveals users behavior and content patterns in the network. Also, our social indicator shows to be useful in the strategy, since it diminished the number of triggers or operations in the market and improved the assertiveness of the model.