Educational data mining (EDM) is a novel scientific area that focuses on developing and applying methods to analyze datasets generated within educational settings. This paper outlines the evolution, significance, and applications of EDM. With the increasing popularity of e-learning in web-based educational systems, EDM has expanded to include a variety of analytical methods and data sources. Some key methodologies addressed include classification, regression analysis, clustering techniques, association rule mining, and Natural Language Processing, among others. Additionally, this paper looks at how EDM can facilitate data-driven decision-making among other areas such as curriculum development and customization of learners’ experiences. It also touches on issues related to the challenges of the scientific field. Finally, some projections about EDM’s future trends are made, especially concerning its integration into AI technologies and development trends like augmented reality or virtual reality, which imply greater possibilities for changes than any other series witnessed before within this sphere.

Intro: In this survey the academic performance of primary and secondary school students in Greece, for three consecutive school years, was examined. The data concerned all Greek students of the last two grades of elementary school and the three grades of junior high school. Method: Unsupervised learning methods such as an X-means algorithm in combination with descriptive and inductive statistical methods were used, in order to examine students’ performance levels. The longitudinal stability of academic performance levels and the influence of demographic characteristics such as the region, gender and guardians’ profession were also examined. Results: The existence of four levels of academic performance and longitudinal stability of frequencies per performance level was confirmed. There was also statistically significant differentiation based on the profession of guardian, gender, and area of residence. Discussion: The results demonstrated specific challenges that the educational policy of the country has to address. The stability of the percentages of students in the four groups of academic performance that emerged over time, shows corresponding stability in the factors that affect academic performance. A gradual reduction in the performance of students in high School was found, as the level of difficulty of the courses increases from class to class. Some demographic characteristics of students are not independent of their performance. However, due to the compliance with the general regulation of personal data, there was no access to additional features that may be related to performance, such as nationality and exact place of residence.