Segmentasi Pelanggan PDAM Berdasarkan Pemakaian Air Bulanan Menggunakan Time Series K-Means Berbasis Dynamic Time Warping

Inayatulloh, Faza (2026) Segmentasi Pelanggan PDAM Berdasarkan Pemakaian Air Bulanan Menggunakan Time Series K-Means Berbasis Dynamic Time Warping. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221094-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Pemakaian air pelanggan PDAM yang dicatat secara bulanan membentuk data runtun waktu sehingga analisis tidak hanya memperhatikan besar konsumsi, tetapi juga pola perubahannya dari waktu ke waktu. Penelitian ini bertujuan untuk menganalisis karakteristik pemakaian air pelanggan rumah tangga kode tarif 1.3 di Zona 2 PDAM Surya Sembada Kota Surabaya serta melakukan segmentasi pelanggan berdasarkan pola pemakaian air bulanan menggunakan Time Series K-Means berbasis Dynamic Time Warping. Data yang digunakan merupakan data sekunder pemakaian air bulanan pelanggan periode Januari 2023 hingga September 2025 yang diperoleh dari PDAM Surya Sembada Kota Surabaya. Setelah dilakukan pre-processing, diperoleh 16.114 pelanggan aktif yang memiliki data lengkap selama 33 periode pengamatan dan tersebar pada 28 subzona pelayanan. Tahapan analisis meliputi eksplorasi data, pre-processing, penentuan jumlah cluster menggunakan Elbow Method, serta segmentasi pelanggan menggunakan Time Series K-Means dengan jarak Dynamic Time Warping. Hasil eksplorasi menunjukkan bahwa rata-rata pemakaian air pelanggan berfluktuasi pada kisaran 24 hingga 29 m³ per bulan dan jumlah pelanggan antar subzona tidak tersebar merata. Hasil Elbow Method menunjukkan jumlah cluster yang dipilih sebanyak empat cluster. Cluster 0 merupakan kelompok terbesar dengan 12.940 pelanggan atau 80,30% dan merepresentasikan pola pemakaian umum mayoritas pelanggan. Cluster 2 terdiri atas 3.169 pelanggan atau 19,67% dengan pola pemakaian yang relatif lebih tinggi dibandingkan mayoritas pelanggan. Sementara itu, Cluster 1 dan Cluster 3 masing-masing terdiri atas 2 dan 3 pelanggan dengan pola pemakaian yang berbeda dari mayoritas pelanggan. ================================================================================================================================
Monthly water consumption data provide information on changes in customer consumption behavior over time. Therefore, the analysis of water consumption needs to consider both the amount of water used and the pattern of consumption changes across observation periods. This study aims to analyze the characteristics of water consumption among household customers under tariff code 1.3 in Zone 2 of PDAM Surya Sembada Kota Surabaya and to segment customers based on their monthly water consumption patterns using Time Series K-Means based on Dynamic Time Warping. This study used secondary data consisting of monthly customer water consumption from January 2023 to September 2025 obtained from PDAM Surya Sembada Kota Surabaya. After the pre-processing stage, 16,114 active customers with complete data across 33 observation periods were obtained, distributed across 28 service subzones. The analysis stages included data exploration, pre-processing, determination of the optimal number of clusters using the Elbow Method, and customer segmentation using Time Series K-Means with Dynamic Time Warping distance. The exploratory analysis showed that the average customer water consumption fluctuated between 24 and 29 m³ per month, while the number of customers was unevenly distributed across subzones. The Elbow Method indicated that the optimal number of clusters was four. Cluster 0 was the largest cluster, consisting of 12,940 customers or 80.30%, and represented the general consumption pattern of the majority of customers. Cluster 2 consisted of 3,169 customers or 19.67%, with a relatively higher water consumption pattern than the majority of customers. Meanwhile, Cluster 1 and Cluster 3 consisted of 2 and 3 customers, respectively, with consumption patterns that differed from the majority.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Dynamic Time Warping, Elbow Method, Pemakaian Air, Segmentasi Pelanggan, Time Series K-Means, Customer Segmentation, Dynamic Time Warping, Elbow Method, Time Series K-Means, Water Consumption
Subjects:	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:	Faza Inayatulloh
Date Deposited:	30 Jul 2026 01:19
Last Modified:	30 Jul 2026 01:19
URI:	http://repository.its.ac.id/id/eprint/138930

Actions (login required)

View Item