This is a repository for a project that investigates the widely debated field, stock prediction, through the lens of data science. Specifically, natural language processing and machine learning is used to identify if the sentiment of Elon Musk’s tweets can be used to predict the movements in the Tesla stock price. Following this notion, the research of this paper includes text preprocessing, topic modelling, sentiment analysis, data standardization and the selection, deployment and evaluation of machine learning algorithms to predict stock movements.
Our findings show that investors seeking to trade Tesla stocks might be able to leverage predictions of the machine learning models to create a profitable investment strategy. However, viable predictions are only found in the paper when the sentiments of tweets related to Tesla are incorporated in the machine learning models. Our findings suggest that the best machine learning model for prediction of stock movements is the Support Vector Machine, providing an accuracy score of 61% when hyperparameters have been optimized through cross-validation, in line with other studies. In this regard, we note that our Support Vector Machine model was overfitting and that there is no guarantee that tweets can be used to predict the movement in the Tesla stock price in the future.