The shifting classroom: impact of heightened seasonal heat in education through sentiment and topic modeling

Abstract

This research applies text mining techniques to examine sentiments and themes among Filipino students adjusting to full in-person classes after pandemic-driven flexible learning, focusing on their experiences during April to June 2023–a period usually marked by vacations due to intense heat. By applying the natural language toolkit (NLTK) for sentiment analysis and Scikit-learn for topic modeling, the study gathered data from Filipino students on their in-person class experiences during this unique calendar shift. Post data cleaning, NLTK was used for sentiment analysis and latent Dirichlet allocation for topic modeling. The findings indicate that the high temperatures adversely affected students, as evidenced by frequent references to terms such as “room,” “focus,” and “hard.” The study identified a mix of positive and negative sentiments and highlighted key issues like academic challenges and the learning environment’s impact. This study also offered insights into students’ coping strategies during extreme heat. These results stressed the importance of considering environmental factors in educational planning and provide actionable insights for institutions to enhance the in-person learning experience, particularly in challenging weather conditions. Moreover, this study demonstrates the effectiveness of sentiment analysis and topic modeling in understanding and unraveling student experiences in specific contexts.