DataLab: Introducing Software Engineering Thinking into Data Science Education at Scale

Data science education is a new area in computer science that has attracted increasing attention in recent years. However, currently, data science educators lack good tools and methodologies. In particular, they lack integrated tools through which their students can acquire hands-on software engineering experience. To address these problems, we designed and implemented DataLab, a web-based tool for data science education that integrates code, data and execution management into one system. The goal of DataLab is to provide a hands-on online lab environment to train students to have basic software engineering thinking and habits while maintaining a focus on the core data science contents. In this paper, we present the user-experience design and system-level implementation of DataLab. Further, we evaluate DataLab's performance through an in-classroom use case. Finally, using objective log-based learning behavior analysis and a subjective survey, we demonstrate DataLab's effectiveness.