Abstract:

This presentation explores the application of new methods of gathering and visualising student data. Based on this analysis, we offer new insights into factors affecting student performance and resource allocation. In this paper we present diverse ways in which student engagement can be visualized from cohort of students (n=252) participating in first year undergraduate science modules.

The process involves data extraction from VLE (Moodle) which takes in the log files data that captures the student interaction and this is transformed into meaningful information. Accessible and open-source tools; Python and Jupyter Notebooks were used for the pre-processing, while Python and Power BI were used to create the visualizations. Assessing and presenting learner engagement and performance is made feasible by learning analytics and visualizations.

This process and the subsequent visualisations allow insights and trends into relationships between the type, times and frequencies of interactions. Highlights include cluster analysis using a multidimensional model and visualisations to discover at risk students. These were identified as early as after 20% of the course length. This prompted further investigation of the correlation of individual factors and student performance.
The goal of these visualizations is to offer instructors and course designers insights for assessing the use and efficacy of instructional materials and interventions.

Keywords:

Data, visualization, engagement, interactions.