Opinion Maps for tracking and visualizing feedback on digital video content

摘要

An increasingly large amount of digital video content continues to become available on the Internet. This content ranges from short entertainment clips by amateurs lasting a few minutes to full-fledged, significantly longer, serious course lectures and presentations by subject matter experts. As these video bits of content continue to get pushed out one-way, to the learner, the content authors have no quick or easy way of getting an objective, machine-processable opinion on arbitrary subsections of their content. In this paper, we explore the design and implementation of an innovative software visualization system that uses ‘Opinion Maps’ and a WYSIWYG (When You See Is When You Give) approach to bring a much needed real-time focus to the feedback aspect of the video mode of instruction. In addition to using routine graphical charting to visualize server-side tabular data, an Opinion Map employs a combination of (i) a Tukey Box⁺ plot — a simple but powerful extension to the Tukey Box plot — to present N-point ordinal opinion scores, collected from learners potentially for every single second of the video content, (ii) a per-second opinion score histogram, and (iii) overall ‘best’ and ‘worst’ opinion score visualizations. The system is nonintrusive; it requires no special content authoring processes and bolts on to prerecorded digital video content.