Exploring Student Engagement in Online Programming Courses: A Two-Level K-means Analysis

This research examines student clustering in online Programming course using a two-level K-means clustering algorithm. The study identifies two distinct clusters, Cluster1 and Cluster0, based on student engagement and performance. Results reveal that Cluster1 demonstrates significantly higher levels of engagement and performance compared to Cluster0 across 10 out of the 11 analyzed features, as validated by t-tests. The accuracy of the clustering model is evaluated using the Silhouette score (0.65) and Confusion matrix (78%). These findings emphasize the efficacy of clustering algorithms in discerning varying levels of student engagement and performance in online Programming courses, thereby offering valuable insights for educators and practitioners.