Creating an Active Learning Environment using Reproducible Data Science Tools

Abstract

After a decade of struggle to help students install and launch machine virtual machines in the cloud, the author migrated his computer science course to the Gigantum data science platform, which automates the delivery of complex software configurations. The goal was to make it easier for students to complete projects so that they could focus on programming rather than system administration. In the process, lectures were redesigned into an active learning experience in Jupyter notebooks in which students run and modify examples as they are presented and can reproduce exactly all work that they have done or has been demonstrated.