Skripsi
ANALISIS KLASTERING KELOMPOK PEKERJAAN MENGGUNAKAN METODE K-MEANS PADA DATA SIKA PT PERTAMINA ZONA 4 TAHUN 2023-2025
The oil and gas industry involves high-risk activities that require accurate utilization of operational data to support occupational safety. However, at PT Pertamina Hulu Rokan Zona 4 Field Limau, the Work Permit Safety System (SIKA) is still used mainly as an administrative document and has not been leveraged to analytically identify work permit patterns. This study aims to explore clustering patterns of work permits based on location, job type, and permit duration using 4,453 SIKA entries from 2023–2025. The data were cleaned, encoded using one-hot encoding, and normalized with Min-Max scaling before being analyzed with the K-Means algorithm. Evaluation using the Silhouette, Davies–Bouldin Index, Calinski–Harabasz, and Inertia Elbow metrics indicates that two clusters provide the optimal configuration, supported by a Silhouette Score of 0.9306 and a DBI of 0.4437. The results reveal two main patterns: the first cluster represents routine jobs with short permit durations and a predominance of cold work, while the second cluster reflects technical jobs with longer permit durations, all conducted in Field areas. These findings provide an objective basis for understanding work permit patterns and offer an initial reference for planning safety control measures in operational environments.
No other version available