This paper presents an in-depth analysis of student behavior and score prediction in the ASSISTments online learning platform. We address four research questions related to the impact of tutoring materials, skill mastery, feature extraction, and graph representation learning. To investigate the impact of tutoring materials, we analyze the influence of students requesting hints and explanations on their performance in end-of-unit assignments. Our findings provide insights into the role of guidance in learning and inform the development of superior tutoring strategies. Additionally, we explore the correlation between mastery/non-mastery of specific skills during in-unit problems and performance in corresponding end-of-unit assignments, shedding light on the efficacy of standard-aligned curricula. In terms of feature extraction, we extract relevant features from extensive student activity data and determine their importance in predicting assignment grades. Furthermore, we employ graph representation learning techniques to model the complex relationships between different entities in the dataset. This yields a more nuanced modeling of factors influencing student performance and facilitates the development of more accurate predictive models. Overall, our study contributes to the practical application of data mining techniques in online learning contexts, with implications for personalized learning, interventions, and support mechanisms. The code is publicly available in https://github. com/DSAatUSU/EDMCup2023. [For the complete proceedings, see ED675485.]