【Abstract】This paper examines whether Twitter messages are related to stock market performance, which metric of Twitter has the strongest relationship, and whether one can use them to make better investment decisions in the financial market. We gathered four-month of Twitter, Google search, financial performance data of S&P 100 companies. We developed a Twitter-specific classifier to conduct sentiment analysis. Through an elaborated process of tweet feature reduction and spam removal, we were able to improve the accuracy of the sentiment analysis classifier for 6%. Our empirical findings suggest that (1) a higher tweet volume leads to an increase in trading volume and volatility; (2) A higher tweet bullishness leads to an increase in stock returns; (3) A higher tweet agreement among Twitter users leads to an decrease in trading volume; (4) Tweet quality strengthens the positive relationship between tweet bullishness and stock returns; (5) Tweet quality strengthens the negative relationship between tweet agreement and trading volume and stock returns. Furthermore, we simulated a set of trading strategies using tweet bullishness and tweet quality, our best result outperforms the market for more than 17% including transaction cost. To our knowledge, this is the first study that leverages Twitter information and device a practical and conceivable trading strategy for individual investors.