Learning Analytics and Online Evaluations: A Graph Data Science Methodology
Learning Analytics, Graph Data Science, Graphs, Online Evaluations, Machine Learning, eXplainable Artificial Intelligence
Evaluation plays an extremely important role as a systematic instrument for correcting failures and promoting successes in the learning process. It is one of the tripods of the students' schooling, along with the curriculum and the teaching and learning process. In view of the new reality with online education (e-learning) spread on a larger scale due to the Covid-19 pandemic, the Federal University of Rio Grande do Norte (UFRN) institutionalized the Multiprova platform to support online evaluation processes in the institution. With the computerization of evaluation processes, Learning Analytics (LA) challenges and the need to use new techniques, such as graph data science. Thus, the task of understanding how students behave, identifying student profiles, and gaining insights through online evaluation resolution records is a field of LA research that can be optimized with graph data science methodology. With the use of LA techniques such as Machine Learning (ML), there is a need to transform interpretable models in education. For this, data visualization and eXplainable Artificial Intelligence (XAI) techniques need to be considered. Based on this reality, the thesis hypothesis arises: Is it possible to use data from online evaluation resolution logs to obtain insights into the learning process and student profiles using LA techniques such as graph modeling and ML? To this end, the theoretical framework about the topics that make up the object of study is presented, such as graphs, LA and online evaluations. Among the results, a systematic literature review pointed to 40 papers involving LA and online evaluations, but no papers used graph metrics with LA techniques such as ML to analyze student performance. Thus, two case studies were modeled according to the proposed graph data science methodology. We realized the importance of using graph features in LA techniques in identifying insights about student learning considering their journey in online evaluation, as well as graph metrics and XAI for the interpretation of the results.