Henriques, Roberto André PereiraMira, Carolina Pratas Ferreira2024-11-132024-11-132024-10-31http://hdl.handle.net/10362/175134Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Business AnalyticsThere has been an advance in computing with the rise of large volumes of data. Education was one of the fields that has advanced to this point. As new technologies and information are continually integrated, datasets are now available from students' interactions with educational software and online learning platforms. Educational software like Moodle exemplifies e-learning solutions that combine traditional teaching methods with information technology resources. In this line of thought, this study investigates the potential use of Learning Analytics and Moodle log data to predict academic success in higher education, considering the first part and the whole course. The research uses a variety of machine learning algorithms, including Random Forest, Logistic Regression, Gradient Boost, Support Vector Machine, and Neural Networks, to uncover patterns in student behaviour and academic performance. The data used in the study is related to Moodle log data and sociodemographic information that comes from NOVA Information Management School's Master of Data Science and Advanced Analytics program, of three academic years (2021-2023). To achieve the objective a CRISP-DM methodology was implemented to serve as the base model for Machine Learning models. The Random Forest model had the highest prediction values; however, overfitting was detected in all models. Key findings show that student involvement, as measured by interactions with course materials and demographic parameters such as age and nationality, has a significant impact on academic success. The study underlines the importance of early and persistent student involvement with Moodle and offers instructors techniques to improve student engagement and performance. Future work will include the development of early warning systems and complete dashboards to offer instructors real-time knowledge.engLearning AnalyticsAcademic SuccessMachine LearningMoodleSDG 4 - Quality educationmaster thesis203776640