The rise of ML in trading and everywhere else largely complements the data explosion that we covered in great detail. We illustrated in Chapter 2, Market and Fundamental Data how to access and work with these data sources, historically the mainstay of quantitative investment. In Chapter 3, Alternative Data for Finance, we laid out a framework with the criteria to assess the potential value of alternative datasets.

A key insight is that the state-of-the-art ML techniques like deep neural networks are successful because their predictive performance continues to improve with more data. On the flip side, model and data complexity need to match to balance the bias-variance trade-off. Managing data quality and integrating datasets are key steps in realizing the potential value.