In recent years, the paradigms of data-driven science have become essential components of physical sciences, particularly in geophysical disciplines such as climatology. The field of hydrology is one of these disciplines where machine learning and data-driven models have attracted significant attention. This offers significant potential for data scientists' contributions to hydrologic research. As in every interdisciplinary research effort, an initial mutual understanding of the domain is key to successful work later on. In this work, we focus on the element of hydrologic rainfall--runoff models and their application to forecast floods and predict streamflow, the volume of water flowing in a river. This guide aims to help interested data scientists gain an understanding of the problem, the hydrologic concepts involved, and the details that come up along the way. We have captured lessons that we have learned while "coming up to speed" on streamflow prediction and hope that our experiences will be useful to the community.