Una-May O’Reilly – MoocDB: Taming MOOC Big Data while Fostering Collaboration in Online Education Research
Dr. Una-May O'Reilly is with the AnyScale Learning for All Group at CSAIL, MIT. While MOOCs exhaust mountains of behavioral data, it can come out the platform pipe in a disorganized, rapid, dynamic multi-stream manner. Conducting education research in the midst of this jumble is tedious, slow, resource consuming and unscalable. How can we routinely generate 1000's of analytic variables and 100's of visualizations from data corresponding to millions of events? How can technology leadership allow every education researcher to study the online course behavioral data captured from every course taught across the globe? What started as an effort to organize the 6.002x data from edX has now become "MOOCdb", a multi-institution effort converging on organization of MOOC data that supports multi-platform, open access, collaborative, online education research. I will present the MOOCdb project, and a number of projects that are spinning off as a result of MOOCdb which provide interactive visualizations and data analytics. These have allowed us to compare an MIT course on the Edx platform side by side with a Stanford course on the Coursera platform using the same variables and visualization. The MoocDB data model has allowed us to populate a 6.002x behavioral evidence database which we have used develop means for instructors to examine the large scale behavior of student problem solving.