Klasterisasi Pola Nilai Impor Migas Bulanan Berdasarkan Pelabuhan Bongkar di Indonesia Tahun 2024 Menggunakan Algoritma K-Means

Abstract

Indonesia’s oil and gas (O&G) distribution relies on imports through various unloading ports with different monthly patterns. This study aims to cluster O&G ports in 2024 based on monthly import values using the K-Means algorithm. The method follows Knowledge Discovery in Databases (KDD) stages: data selection, preprocessing, transformation, clustering, evaluation, and visualization. Analysis was conducted in Google Colab using Python with Scikit-learn, Pandas, and Matplotlib. Results show three main clusters: ports with high, medium, and low import volumes. Evaluation using Elbow Method and Silhouette Score confirmed that three clusters offer optimal separation. PCA visualization clearly shows cluster distribution. These findings support more efficient energy logistics planning and port infrastructure development based on data-driven insights.