Abstract :
Users, especially students, go on a journey with each new programming project they create, and until recently assessments could not access that journey, having to depend entirely on the final project artifact. New technologies allow the development environment itself to be instrumented to collect all of the intermediary steps of that journey. Learning analytics, powerd by data mining and machine learning, can be used to extract patterns of the numerous code snapshots collected. This abstract describes new work that applied these ideas in a new way, in a blocks-based, mobile-targeted, live-development environment, MIT App Inventor.