Still under construction
Massive Open Online Courses (MOOCs) are expanding their presence not just for online learners, but also for traditional students in universities. Georgia Institute of Technology runs an EdX (an online MOOC provider) course on CS1301x (Introduction to Python), which is an online course open to non-institutional online learners and traditional students on-campus. Using tools such as MongoDB, we aggregate and analyze the data provided by EdX to determine metrics to define and predict student success. We take a set theory approach and organize students into four different classes based on intersecting features. The classes are: Overachieving students, Median students, Underachieving students and Disinterested students. These classes are determined by engagement during video lectures, performance in assignments & tests, and presence of prior coding experience. Using this system, we will propose changes to the course, with the goal of making it more accessible to underachieving students and more challenging to the overachieving students. We also propose automated solutions for addressing specific challenges such as figuring out which assignment questions have faulty answers in the autograders. We present results achieved by the system on a dataset consisting of the on-campus population over a time period of three semesters. Once we formalize these results, we plan on making the system modular so that it can be applied to other online courses and scaled up to larger datasets.