Blended learning is becoming ubiquitous in higher education. It combines traditional face-to-face teaching and online learning that typically takes place in learning management systems which store all student interactions with the system in their log files. In addition to this data, education institutions store various electronic student datasets and survey responses. Learning analytics involves collection, analysis and mining of education data to better understand the learning process and obtain information that can inform decision-making with regard to enhancing student success. The purpose of the research is to improve the understanding of the blended learning process in higher education with the aim of identifying the factors that significantly influence student work in the online classroom, and determining activities that are most associated with student final performance. The empirical part presents two frequent learning analytics approaches carried out with open source software. The calculations were based on education data from various sources of the Faculty of Public Administration. Education data mining examined the predictive effectiveness of various models of predicting student final (non)performance based on their activity in the online classroom and predispositions (learning approach, previous performance).

Total universe/Complete enumerationTotalUniverseCompleteEnumerationTotalUniverseCompleteEnumeration

Computer-based observationObservation.ComputerBased

Self-administered questionnaire: Web-based (CAWI)SelfAdministeredQuestionnaire.CAWI