Clustering Analysis of E-Learning Readiness in Java Island Indonesia with GIS Visualization

Abstract

E-learning readiness (e-readiness) is used as a tool to measure the success rate of using ICT in the academic process. Until now, a lot of research on e-readiness has been collected in various regions, especially Java Island, but the data has not been grouped and visualized, so it is difficult to know the level of readiness. The purpose of this study is to group e-readiness data using clustering analysis, then create a GIS-based map of the distribution of e-readiness clusters. To obtain optimal clustering results, researchers used the K-means and PCA combination as a cluster optimization method. The total dataset used is 27 locations' data with 2 parameters selected based on the level of e-learning readiness. Based on the results of the performance analysis using the selected internal clustering validation metrics, specifically the Davies Bouldin Index (DBI) and the average within centroid distance, each metric indicates the best cluster with values of 0.057 and 0.001, respectively. The most optimal cluster formed using the K-means and PCA methods, with a total of three clusters spread across various areas on the island of Java. As for the division of each cluster by its location point, namely, cluster 1 (amounting to 20 locations) for the ready level, cluster 2 (amounting to 4 locations) for the less ready level, and cluster 3 (amounting to 3 locations) for the very ready level.