Understanding Characteristics of Thai Students through PISA 2018 Data Using Data Visualization Techniques

Abstract

Education is the cornerstone of a nation's development, economy, and competitiveness. A clear and deep understanding of the current educational situation is needed. The Programme for International Student Assessment (PISA) has been widely used as an indicator of a nation's school system. This study is conducted using a topological data analysis technique called UMAP to obtain a deeper understanding of Thai students based on PISA 2018 data. For each subject---mathematics, reading, and science---students's performances are preprocessed into high-dimensional vectors. Our proposed method reduces the high-dimensional vectors into two dimensions. The results show deeper insights beyond mere average scores, which are unattainable through other techniques. For example, although the Thai students' average scores are below the OECD averages, their performance across various skills was not uniformly poor. Thai students excelled in the interpretation skill in both mathematics and science compared to others. In reading, Thai students performed well in the cognitive process of locating information but struggled with the process of gaining understanding. Students with higher scores were found to read more books, have higher expectations of completing a bachelor's degree, and experience less bullying. These insights are valuable for policymakers and educators aiming to improve the Thai education system.