Description
Software engineering education today is a technologically advanced and rapidly evolving discipline. Being a discipline where students not only design but also build new technology, it is important that they receive a hands on learning experience in the form of

Software engineering education today is a technologically advanced and rapidly evolving discipline. Being a discipline where students not only design but also build new technology, it is important that they receive a hands on learning experience in the form of project based courses. To maximize the learning benefit, students must conduct project-based learning activities in a consistent rhythm, or cadence. Project-based courses that are augmented with a system of frequent, formative feedback helps students constantly evaluate their progress and leads them away from a deadline driven approach to learning.

One aspect of this research is focused on evaluating the use of a tool that tracks student activity as a means of providing frequent, formative feedback. This thesis measures the impact of the tool on student compliance to the learning process. A personalized dashboard with quasi real time visual reports and notifications are provided to undergraduate and graduate software engineering students. The impact of these visual reports on compliance is measured using the log traces of dashboard activity and a survey instrument given multiple times during the course.

A second aspect of this research is the application of learning analytics to understand patterns of student compliance. This research employs unsupervised machine learning algorithms to identify unique patterns of student behavior observed in the context of a project-based course. Analyzing and labeling these unique patterns of behavior can help instructors understand typical student characteristics. Further, understanding these behavioral patterns can assist an instructor in making timely, targeted interventions. In this research, datasets comprising of student’s daily activity and graded scores from an under graduate software engineering course is utilized for the purpose of identifying unique patterns of student behavior.
Reuse Permissions
  • Downloads
    pdf (1.5 MB)

    Details

    Title
    • Continuous assessment in agile learning using visualizations and clustering of activity data to analyze student behavior
    Contributors
    Date Created
    2016
    Resource Type
  • Text
  • Collections this item is in
    Note
    • Partial requirement for: M.S., Arizona State University, 2016
      Note type
      thesis
    • Includes bibliographical references (pages 74-77)
      Note type
      bibliography
    • Field of study: Engineering

    Citation and reuse

    Statement of Responsibility

    by Suhas Xavier

    Machine-readable links