Design and Evaluation of a Block-based Environment with a Data Science Context

As computing becomes pervasive across fields, introductory computing curricula needs new tools to motivate and educate the influx of learners with little prior background and divergent goals. We seek to improve curricula by enriching it with authentic, real-world contexts and powerful scaffolds that can guide learners to success using automated tools, thereby reducing the strain on limited human instructional resources. To address these issues, we have created the BlockPy programming environment, a web-based, open-access, open-source platform for introductory computing students (https://www.blockpy.com). BlockPy has an embedded data science context that allows learners to connect the educational content with real-world scenarios through meaningful problems. The environment is block-based and gives guiding feedback to learners as they complete problems, but also mediates transfer to more sophisticated programming environments by supporting bidirectional, seamless transitions between block and text programming. Although it can be used as a stand-alone application, the environment has first-class support for the latest Learning Tools Interoperability standards, so that instructors can embed the environment directly within their Learning Management System. In this paper, we describe interesting design issues that we encountered during the development of BlockPy, an evaluation of the environment from fine-grained logs, and our future plans for the environment.