Aspect Based Sentiment Analysis of Distance Learning Using Natural Language Processing Techniques

Abstract

Understanding public opinion on key issues is essential given the exponential growth of online discourse. This study investigates aspect-based sentiment analysis, concentrating on tweets pertaining to online learning. Measuring attitudes is becoming more crucial as the global education landscape shifts rapidly in order to provide more flexible and effective online learning environments.

The research is rigorous, beginning with a comprehensive statistical analysis of a substantial set of 202,645 entries from Twitter. Through exploratory data analysis, trends in the tweet distribution across different geolocations are uncovered, highlighting temporal patterns and global interaction dynamics. Two feature engineering techniques—type pre proxy and vectorization—allow for a thorough examination of textual material.

The main component of the project is aspect-based sentiment analysis, which determines sentiments for a number of factors like instructor participation, platform usability, and course content. Using the TextBlob library, this study provides a deep understanding of how users express their viewpoints on many aspects of remote learning by revealing attitudes at the category and aspect levels.

Visualizations like word clouds, geographic plots, and hourly sentiment distributions provide rich insights into the dialog. The project's innovative technique includes sentiment scoring for certain aspect categories, such as positive and negative sentiment scores for "course content" and "platform usability." These scores, which are based on the frequency and sentiment polarity of phrases, offer a dynamic picture of the variables that most influence moods.

Dialogue can be deeply understood by visualizations such as word clouds, geographic plots, and hourly sentiment distributions. Sentiment scoring for certain aspect categories—such as positive and negative sentiment scores for "course content" and "platform usability"—is one of the project's novel techniques. Based on the frequency and sentiment polarity of phrases, these scores provide a dynamic picture of the factors that most affect moods.

In summary, our study advances the rapidly evolving field of sentiment analysis and provides a comprehensive framework for examining attitudes in the intricate realm of online conversation, especially with regard to remote learning.