VEMETER: A Tool for Evaluating Participation Levels In Virtual Class Sessions

While online education and virtual meetings increasingly form part of the modern learning environment, it is uniquely difficult to gauge participant engagement in virtual class sessions. Traditional approaches are usually based on either attendance counts or observation by a moderator and can seldom offer real-time, precise feedback as to the degree of individual contribution. This kind of approach is surely unsuitable for a large learning setting or even for remote learning, where active participation may be more difficult to gauge. The existing limitations, therefore, raise the need to introduce VEMeter, a tool designed to measure and quantify the level of participation in virtual class sessions. VEMeter analyzes transcripts and chat logs from virtual meetings using state-of-the-art text-processing techniques, including TF-IDF and Cosine Similarity. These techniques help in measuring the participant engagement against those of the presenter or facilitator. This enables a more objective and correct analysis of how participants in the session interact and contribute. To enhance its analysis, VEMeter cleans text by applying several advanced techniques: stopword removal, lemmatization, contraction expansion, and tokenization. Through the application of these pre-processing techniques, it will be easier to eliminate irrelevant content and noise inside the text for the generated engagement metrics to reflect meaningful participations and not trivial ones. Data privacy and security are at the very core of VEMeter. The tool anonymizes participant data through name encryption; thus, it guarantees protection of participants' personal information in an ethical analytics ecosystem. This already allows VEMeter to provide insights into engagement without compromising participant confidentiality. VEMeter enables rich analytics and visualizations through bar charts, scatter plots, and correlation heatmaps that explain the level of engagement and variation of participation based on the word count. Educators can track participation in virtual sessions in real time through these insights and make timely interventions to enhance engagement. This study contributes to engagement level assessment by introducing an NLP-driven participation analysis tool. Practically, it effectively provides educators with a data-driven methodology to measure student engagement in virtual classrooms. The following paper designs, implements, and evaluates a proof-of-concept VEMeter to enrich virtual education by supplying real-time insights into participant engagement.