Social Media Behavior Analysis of Marine Engineering Learners using Natural Language Processing and Time Series Analysis

Abstract

Social media undoubtedly has a profound impact on how students communicate and perceive the world. The Philippines is a maritime country and maritime seafarers can also affect the economy of the Philippines. In the Philippines education system, Maritime baccalaureate programs were developed such as marine engineering and marine transportation. Maritime programs are constantly evaluated to the global standard set by the international community. In this study, we value the importance of investigating maritime students' academic performance. This research explores the relationship between marine engineering students' academic standing and social media behavior by employing Bi-term topic modeling on social media texts collected from students of varying academic standings. For marine engineering students with high academic standing, the content of discussions often relates to academic and existential themes. In contrast, marine engineering students with lower academic standing tend to engage more with entertainment and politically oriented material. A significant finding from the correlation analysis is the negative association between the frequency of social media usage and academic performance in mathematics, suggesting that a high frequency of social media usage may adversely affect academic success. Additionally, K-means clustering was used to group Facebook posts, identifying patterns that align with key societal and educational events. These results highlight the complex influence of social media within educational settings, underscoring both its positive features and potential risks. Time series analysis was also performed to check for patterns in the number of students' Facebook posts.