A Learning Analytics Framework to Analyze Corporal Postures in Students Presentations.

Area of Science:

• Educational Technology
• Human-Computer Interaction
• Data Science

Background:

• Effective communication is crucial for professional development and career advancement.
• Evaluating complex communication skills like corporal analysis in public speaking has been challenging.
• Emerging sensor technologies offer new possibilities for quantifying personal and contextual information in learning environments.

Purpose of the Study:

• To present a novel framework for analyzing and detecting patterns in students' corporal postures during oral presentations.
• To facilitate the quantitative assessment of non-verbal communication skills in an educational context.
• To enable automated feedback systems for improving students' presentation abilities.

Main Methods:

• A four-step framework: Data Collection, Statistical Analysis, Clustering, and Sequential Pattern Mining.
• Utilizing sensor technologies to capture student interactions and contextual information during presentations.
• Applying statistical analysis to understand data variations, clustering to group students by posture, and sequential pattern mining to identify posture sequences.

Main Results:

• The framework successfully segmented 222 Computer Engineering students into three distinct groups based on corporal postures.
• Statistical analysis revealed how student postures evolved throughout the academic year.
• Sequential pattern mining identified frequent postural sequences within different student groups.

Conclusions:

• The developed framework can guide the provision of automated feedback to students on their oral presentations.
• This approach offers valuable background data for comparing presentation skills across different undergraduate courses.
• Integrating technology for analyzing non-verbal cues enhances the evaluation of essential professional skills.