Predicting Tesla: Stock Market Forecasting Using Facebook's Prophet

The stock market can potentially deliver great returns for investors who invest in high-growth companies at the right time. The aim of this paper is to evaluates the use of Facebook's open-source forecasting tool, Prophet, to predict future stock price movement based on historical performance. This paper includes the fitting of a 12-month forecast model based on the past 5 years' stock price performance of the leading electrical vehicle manufacturer Tesla (TSLA) as the model case study. The model's performance is furthermore evaluated and the hyperparameters tuned to improve accuracy. Key findings include the usefulness of the Prophet's model to forecast future performance as well as the possibility to increase model accuracy further by means of extensive hyperparameter tuning. For example. our results show that forecasting 21 days ahead using Prophet has been improved significantly from an initial 21.2% mean absolute percentage error to 12.7% by hyperparameter tuning. The paper can bring value to investors, stock market enthusiast, and other stakeholders eager in using predictive analytics to determine whether a particular stock is worthy of an investment.