Unveiling the Patterns of Employability in Higher Education: A DBSCAN Cluster Analysis of College Programs

This paper applies DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering to examine patterns in employability across 26 college programs over three years, 2021 to 2023. This study uses secondary data sources on graduate numbers and employability rates to look for possible understructures and groupings among academic programs based on employment outcomes. To reduce the dimensionality, PCA was used while performing DBSCAN based on the values of ε=0.73 and MinPts=6, which were derived from k-distance graph analysis. The obtained clusters are then analyzed about measures such as the Silhouette Coefficient of 0.3990, Calinski-Harabasz Index of 12.8237, and Davies-Bouldin Index of 1.0487. The analysis made clear that the divisions of three groups differentiated: large numbers of programs with stable employability rates, relatively small numbers with consistent growth, and outliers with more volatile patterns. This research draws together findings on graduate employability that describe the dynamics across fields of study, which inform targeted interventions and strategic decision-making in higher education. The study contributes to understanding the complex relationship between academic programs and employment outcomes and UN Sustainable Development Goals 4, Quality Education, and 8, Decent Work and Economic Growth. This proposed research develops an evidence-based approach to supporting increasing students' employability and facilitating better connections between higher education and labor markets.