NLP-Based System Real-Time Analysis of Students Textual Feedback to Enhance Online Class Session Insight

This paper presents a real-time class feedback analysis system designed to process and analyze textual feedback provided by students during classroom sessions. Leveraging natural language processing techniques, the system preprocesses feedback data, extracting sentiment information crucial for understanding student perceptions. Through text preprocessing steps such as tokenization, lowercasing, stop word removal, lemmatization, emoji sentiment handling, contraction expansion, and symbol elimination, the system ensures data quality. Subsequently, sentiment analysis using the Text Blob library provides polarity scores, aiding in categorizing feedback as negative, neutral, or positive. The project emphasizes visualizing sentiment analysis results via graphs or plots, offering educators actionable insights into classroom sentiment trends, potential concerns, and areas of satisfaction, thereby enhancing teaching effectiveness and student engagement. In addition to sentiment analysis, this paper incorporates an analysis of interaction and interactivity within the session. By dividing the duration of each session into twenty-minute intervals and counting the total number of feedback messages from each student in these intervals, we calculate the average interactivity of individual students and the session overall. Standard deviation measures the variability in interaction counts, offering insights into how consistently students participate. We further visualize the interaction statistics results, presenting graphs and plots that illustrate student engagement levels over time. This approach offers educators a deeper understanding of student engagement levels throughout the session, enabling them to tailor instructional strategies to optimize student participation and learning outcomes.