Maseeh Summer Undergraduate Research Experience

Funding Source

Undergraduate Research and Mentoring Program (URMP)

Document Type

Poster

Publication Date

2023

Abstract

Currently, the assessment of instructors in classrooms are time consuming and biased. Active learning has been determined to be the best teaching strategy in STEM based classes, but determining if active learning strategies are being used can be difficult. Previously, a tool called Decibel Analysis for Research in Teaching (DART) has been used to determine active learning in classrooms based on audio. Audio can be used to determine active learning in a more efficient way than DART presents. The “who speaks when” analysis used Pyannote.audio an open source pretrained speaker diarization pipeline. The pipeline segments the audio and labels speakers based on a clustering algorithm. Speaker diarization easily shows who is talking when and can be used to determine a more precise analysis of student and instructor interactions.

Share

COinS