Evaluation of Usage Patterns Within Lecture Videos

Date

2018-05

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Lecture videos supplementing or replacing traditional classroom instruction have become common in academic settings and have been shown to contribute positively to students’ academic performance. Video usage patterns contain information that can be employed by instructors to improve course offerings. For instance, if a particular video, or a particular section of a video, is being viewed heavily by students, this indicates that the content of that video or segment is particularly difficult. Using this information, the instructor can intervene, if these patterns run counter to their expectations for the course. In this project, I developed a series of graphs which transform large quantities of video usage data into easily understood graphical formats. These graphs plotted usage by date, by content, and by user. I also developed graphs showing usage inside individual videos and a graph showing the order in which users viewed videos. Additionally, I generated a system for dividing users into clusters based on their weekly usage of lecture videos. This system uses an expectation maximization algorithm to find groups of users with similar behavior, without prior knowledge of the number of groups in the dataset. This project analyzed usage data from over 1600 students, participating in seventeen courses across three semesters and four fields of study. Most students were found to use small numbers of videos at specific points during the semester, likely coinciding with exams. Some users, most of whom were participants in certain classes, instead viewed videos regularly each week. Clustering users based on weekly activity found five distinct groups of users. Users grouped into a given cluster were shown to have distinct behaviors from users in other clusters, beyond the dates on which they viewed the videos; they focused on different videos, viewed the videos in different orders, and tended to be participants in different courses.

Description

Keywords

Clustering, Learning technologies, Learning, Lecture videos, Video analytics, Video usage patterns

Citation