Efforts to develop Pre-K-12 data science curricula have accelerated in recent years. Data science curriculum documents generally specify desired student learning outcomes, but they do not always specify what teachers need to know and do to support students in attaining the outcomes. This article addresses the issue by offering a process that can be used to set and refine goals for data science teacher education. The process draws upon research and theory about mathematical knowledge for teaching, statistical knowledge for teaching, and technological pedagogical statistical knowledge. These types of knowledge are defined and exemplified in the context of Pre-K-12 data science. Examples of teaching actions that require the different types of knowledge to support students’ learning are given to illustrate how meaningful goals for data science teacher education can be set. An agenda for future research and development is also proposed. The proposed agenda includes generating curriculum-specific data science teacher education goals, identifying and prioritizing teacher education strategies that have the greatest impact on students’ learning, and continuously refining and improving theory and practice in data science teacher education using empirical data.