(366c) Data-Driven Analysis of Learning Behavior within a Student-Led Chemical Engineering Wiki

The ChemEng Wiki is a student-led initiative aimed at improving collaborative learning and fostering technical expertise within the chemical engineering cohort at Imperial College London. The primary functions of the Wiki are to enable students to share their interpretations of taught material with other students and both develop and disseminate new content effectively.

The Wiki has been operational for the past two academic years, following a pilot in early 2020. During this period we have had the opportunity to evaluate the impact of the platform using pseudo-anonymized usage data through Google Universal Analytics. Analysis of the temporal dataset obtained via applying machine learning techniques, such as dimensionality reduction, has uncovered substantial heterogeneity in the way the Wiki was used to facilitate learning throughout the cohort. This has allowed us to correlate differences in learning behavior. In particular we highlight clusters of students who access material across subject years (cross-reading) and who access material during semester-time versus non-term time (pre-reading). These clusters help to better understand how students access material, how they use to this build their engineering competence and how the format and framework of the Wiki can be improved.

Such analysis provides the basis of developing an integrated open-source data collection and analysis pipeline, which represents a novel way of approaching web usage analytics within higher education. This further demonstrates the untapped potential a data-driven approach holds in aiding the development of an authentic appreciation of student learning, and provides evidentiary support towards the use of wikis in higher education pedagogy.