<![CDATA[The use of alumni data to support curriculum evaluation continues to face challenges due to the limitations of conventional data collection methods, such as manual surveys, which often result in low response rates. Meanwhile, LinkedIn provides relatively comprehensive and up-to-date alumni career information; however, its potential for supporting tracer studies remains underutilized. This study aims to analyze the career patterns of higher education alumni using LinkedIn data collected through web scraping and analyzed with the K-Means clustering algorithm within the Knowledge Discovery in Databases (KDD) framework. The proposed approach applies the KDD process to generate a data-driven mapping of alumni career patterns as a complement to conventional tracer studies. The dataset consisted of 133 alumni profiles, which were processed through the stages of data selection, preprocessing, transformation, clustering, and evaluation. The results indicate that the majority of alumni are employed in the technology sector and occupy mid-level or specialist positions. The K-Means algorithm identified three distinct career clusters, representing career tendencies in business process and operations, systems and technology development, and data utilization and software quality assurance. These findings reveal the distribution of alumni competencies across business, data, and technology domains. However, the clustering quality was relatively low, as indicated by a Silhouette Score of 0.0321 and a Davies-Bouldin Index of 3.0487, suggesting limited separation among the identified clusters. Therefore, the clustering results should be interpreted as an initial mapping of alumni career patterns rather than definitive classifications. Overall, this study demonstrates the potential of professional social media data as a valuable resource for supporting data-driven alumni career analysis and complementing traditional tracer study practices.]]>
Copyrights © 2026
