HyperMOOC: Augmenting MOOC Videos with Concept-based Embedded Visualizations

Paper Title

HyperMOOC: Augmenting MOOC Videos with Concept-based Embedded Visualizations

Publication Info

• Topic area: Enhancing MOOC video learning through embedded visualizations and interactions.
• Keywords: MOOC, hypervideo, embedded visualizations, cognitive load, online learning, concept-based design, learner engagement, interactive videos, knowledge augmentation, education technology.

Background and Problem

• Problem / challenge: MOOC learners struggle to maintain knowledge context and understand complex concept relationships due to the lack of integrated, interactive support within video content. Existing solutions like navigation menus and hypervideo systems are either non-educational or lack curriculum-aware capabilities.
• Significance: Addressing these challenges can improve learning outcomes, reduce cognitive load, and enhance learner engagement in MOOCs, which are increasingly central to global education.
• Motivation and related work: Prior research has explored visual analysis of learning content, hypervideo, and video interaction methods, but these approaches often treat content analysis levels in isolation or fail to embed visualizations directly into videos. This paper builds on these foundations to address the gap in integrated, concept-based augmentation for MOOC videos.

Solution

• Proposed approach: HyperMOOC, a tool that augments MOOC videos with concept-based embedded visualizations and hyperlink-based interactions to enhance learning engagement and understanding.
• Novelty:
  1. Development of a two-level, four-dimensional concept-based design space for MOOC video augmentation.
  2. Implementation of HyperMOOC, featuring multi-glyph designs for concept demonstration and multi-stage interactions (Play, Focused, Paused).
  3. Evaluation of HyperMOOC through a user study with 36 learners and expert interviews, demonstrating its effectiveness in improving learning outcomes and usability.
• Procedure and key techniques:
  • Extract data from MOOC videos using deep learning models and a bottom-up pipeline.
  • Embed visualizations at three levels (element, event, conclusion) with multi-glyph designs and radial visualizations.
  • Support multi-stage learner interactions (Play, Focused, Paused) with mouse-hover and hyperlink-based navigation.
  • Evaluate usability, cognitive load, and learning outcomes through user studies and expert feedback.

Results

• Concrete findings:
  • FULL mode of HyperMOOC significantly improved learning outcomes compared to RAW and AUG modes (e.g., Task 1: p=0.010, d=1.34; Task 2: p=0.013, d=1.30).
  • SUS usability scores and NASA-TLX cognitive load rankings favored FULL mode, with 63.9% of participants reporting reduced cognitive load.
• Advantage over baselines:
  • FULL mode outperformed RAW and AUG modes in test scores, usability, and cognitive load reduction.
  • Embedded visualizations and hyperlink-based interactions enabled deeper course understanding and engagement.
• Experiments / evaluation:
  • User study with 36 participants across three academic majors (Mathematics, Education, Art Design) using three learning modes (RAW, AUG, FULL).
  • Post-study questionnaires, SUS, NASA-TLX, and interviews assessed usability, cognitive load, and engagement.
  • Expert interviews provided additional insights into real-world applicability and design improvements.
• Limitations and future work:
  • Limited sample size and focus on immediate learning outcomes; future studies should explore long-term retention and larger, diverse cohorts.
  • Challenges in adapting visualizations for non-slide-based or abstract content.
  • Need for adaptive visualization strategies to accommodate diverse learner backgrounds and cognitive loads.

Summary

This paper introduces HyperMOOC, a tool that enhances MOOC videos with concept-based embedded visualizations and hyperlink-based interactions. By addressing challenges in maintaining knowledge context and reducing cognitive load, HyperMOOC improves learning outcomes and engagement. A user study with 36 participants demonstrated significant benefits of the FULL mode over traditional video learning, while expert feedback highlighted areas for future improvement, including adaptive visualizations and expanded customization. The findings provide actionable insights for designing interactive, concept-driven learning systems in MOOCs and beyond.