Visualizing Instructor Feedback for Video-Based Online Courses in Real-Time

Abstract

Modern online learning platforms allow instructors to post videos, texts, assignments, discussion forums, and other educational tools in order to enhance students’ learning and to widen the reach of the course content beyond the limits of a traditional classroom. Especially critical to evaluating the impact of these conventional supplements are feedback mechanisms for instructors that characterize students’ learning behaviors. This research project designs and implements a real-time video analytics toolkit for the Princeton 3ND website based on data obtained from a continuous browser logger of students’ course-related video watching. Specifically, performance-optimized Python and C algorithms intelligently parse the raw logger data to obtain students’ video completion percentage and pause locations. The pausing behavior is then correlated with grades on assignments related to the video content, demonstrating that a strong correlation exists between classroom performance and pausing behavior for certain students. The methods and results generated here present a framework for an extendable research topic in exploring effective learning behaviors related to online classroom learning