Analyzing Millions of Submissions to Help MOOC Instructors Understand Problem Solving

We present an analytics framework that helps analyze student behavior during problem solving in Massive open online courses. There are some important differences in how students solve problems online when compared to on campus education. However, massive amounts of information, at a very high granularity, capturing how students access the content and solve problems is currently stored. We present a structured way for assembling, aggregating, visualizing and statistical approaches to analyze this data. We are developing a comprehensive set of tools that will help and inform instructors. This paper presents our first few steps towards achieving that goal.