Penerapan ST-DBSCAN pada Clustering Spasial-Temporal Keluhan Pelanggan Perumda Air Minum Surya Sembada Surabaya

Aisha, Dinda Fitri Nur (2026) Penerapan ST-DBSCAN pada Clustering Spasial-Temporal Keluhan Pelanggan Perumda Air Minum Surya Sembada Surabaya. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221097-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Peningkatan jumlah pelanggan dan luas wilayah pelayanan Perumda Air Minum Surya Sembada Surabaya menyebabkan pengelolaan keluhan pelanggan semakin kompleks. Analisis agregat belum mampu mengidentifikasi konsentrasi keluhan yang berdekatan dalam ruang dan waktu. Penelitian ini bertujuan menerapkan Spatio-Temporal DBSCAN (ST-DBSCAN) untuk mengidentifikasi klaster kepadatan keluhan, menganalisis karakteristik hotspot, dan menyusun rekomendasi prioritas penanganan. Data yang digunakan berupa pengaduan pelanggan periode Januari–Oktober 2025 yang diagregasikan menjadi 1.432 observasi subzona-bulan. Koordinat longitude dan latitude ditransformasikan ke sistem UTM Zona 49S sehingga diperoleh koordinat Easting dan Northing dalam meter, sedangkan waktu diubah menjadi indeks bulan. Penentuan parameter dilakukan melalui evaluasi kombinasi Eps spatial, Eps temporal, dan MinPts berdasarkan jumlah klaster, proporsi noise, Silhouette Coefficient, Davies-Bouldin Index, distribusi anggota, serta keterinterpretasian hasil. Kombinasi Eps_spatial = 1.750 meter, Eps_temporal = 2 bulan, dan MinPts = 20 dipilih sebagai model final. Model tersebut menghasilkan empat klaster dan 286 observasi noise atau 19,97%, dengan nilai Silhouette Coefficient sebesar 0,2444 dan DBI sebesar 0,9310. Klaster 1 menjadi klaster dominan dengan 801 observasi dan cakupan 81 subzona. Jenis pengaduan dominan meliputi Meter PV/Jarum/Lebih dari 5 Tahun, Situasi Keadaan Meter, dan GMB–Meterisasi. Hasil penelitian menunjukkan bahwa setiap klaster memiliki karakteristik spasial, temporal, dan jenis pengaduan yang berbeda. Oleh karena itu, prioritas pelayanan perlu disesuaikan berdasarkan wilayah, periode peningkatan keluhan, dan jenis pengaduan dominan melalui inspeksi meter, evaluasi meterisasi, verifikasi tarif, pemeliharaan jaringan, serta komunikasi gangguan secara proaktif.
===================================================================================================================================
The growing number of customers and the expanding service area of Perumda Air Minum Surya Sembada Surabaya have increased the complexity of customer complaint management. Aggregate analysis has not been sufficient to identify concentrations of complaints occurring close together in space and time. This study aims to apply Spatio-Temporal DBSCAN (ST-DBSCAN) to identify complaint-density clusters, analyze the characteristics of the resulting hotspots, and formulate service-handling priorities. The data consist of customer complaints recorded from January to October 2025 and aggregated into 1,432 subzone-month observations. Longitude and latitude were transformed into Easting and Northing coordinates in meters using UTM Zone 49S, while time was converted into a monthly index. Parameter selection was conducted by evaluating combinations of Eps_spatial, Eps_temporal, and MinPts based on the number of clusters, noise proportion, Silhouette Coefficient, Davies-Bouldin Index, member distribution, and interpretability. The final model used Eps_spatial = 1,750 meters, Eps_temporal = 2 months, and MinPts = 20. It produced four clusters and 286 noise observations, equivalent to 19.97%, with a Silhouette Coefficient of 0.2444 and a DBI of 0.9310. Cluster 1 was the dominant cluster, containing 801 observations across 81 subzones. The dominant complaint types included PV Meter/Needle/Over Five Years, Meter Condition, and GMB–Meterization. The results show that each cluster has distinct spatial, temporal, and complaint-type characteristics. Therefore, service priorities should be adjusted according to location, periods of increasing complaints, and dominant complaint types through meter inspection, meterization evaluation, tariff verification, network maintenance, and proactive communication of service disruptions.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Clustering Spasial-Temporal, Hotspot Pengaduan, Keluhan Pelanggan, Perumda Air Minum Surya Sembada, ST-DBSCAN, Customer Complaints, Perumda Air Minum Surya Sembada, Spatio-Temporal Clustering, Spatio-Temporal Hotspot, ST-DBSCAN
Subjects:	Q Science Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Dinda Fitri Nur Aisha
Date Deposited:	31 Jul 2026 02:39
Last Modified:	31 Jul 2026 02:39
URI:	http://repository.its.ac.id/id/eprint/138936

Actions (login required)

View Item