<![CDATA[The increasing adoption of information technology in higher education has generated large volumes of academic data that can be utilized to support evidence-based institutional decision-making. However, many Islamic higher education institutions still face challenges in identifying student graduation patterns, particularly in distinguishing students who graduate on time from those who experience academic delays. This study aims to analyze student graduation patterns using the K-Means clustering algorithm as an Educational Data Mining approach. The study utilized academic records of 634 undergraduate students from an Islamic higher education institution in Lampung, Indonesia, with 127 records (20% of the population) selected through representative sampling for clustering analysis. Data preprocessing included cleaning, normalization, and attribute selection before clustering was performed using the Orange Data Mining platform. The resulting clusters were evaluated through visual exploration using scatter plots to identify similarities and differences in graduation characteristics. The analysis revealed distinct student clusters representing different graduation profiles, including timely graduates and students with delayed completion. These findings provide valuable insights for academic management by enabling early identification of at-risk students and supporting data-driven academic interventions, strategic planning, and quality assurance initiatives. The proposed approach demonstrates that K-Means clustering is an effective method for discovering hidden patterns in academic data and can contribute to improving student success management in Islamic higher education institutions.]]>
