Riza Rumayanti Dewi
Universitas Horizon Indonesia

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

K-Means Clustering of Karawang Regency Sub-Districts Based on Population Educational Attainment to Support Educational Equity Riza Rumayanti Dewi; Anwar Hilman; Deden Moh Alfiansyah; Arif Budimansyah Purba; Jajang Mulyana
RISTEC: Research in Information Systems and Technology Vol. 7 No. 1 (2026): RISTEC: Research in Information Systems and Technology
Publisher : Institut Pendidikan Indonesia Garut

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31980/0tpvbx54

Abstract

Educational attainment is unevenly distributed across the 30 sub-districts of Karawang Regency, yet regional open data remain in raw form and offer no objective guidance for equity-oriented policy. This study segments the sub-districts by the educational attainment of their residents using K-means clustering. Secondary open data from the Karawang Regency One Data portal for 2024, covering 2,612,065 residents in ten education categories, were pivoted, converted into percentages of each sub-district's population, and standardized with z-scores. The number of clusters was examined with the elbow method and the silhouette score for k = 2 to 9, and the final model used k-means++ initialization with ten restarts. Three clusters were obtained, with a silhouette score of 0.386: an advanced cluster of three sub-districts (38.51% of residents with senior high school education or above), a moderate cluster of five sub-districts (35.03%), and a lagging cluster of 22 sub-districts (17.23%) that is home to 56.1% of the population and is dominated by primary school graduates. The silhouette score was higher for k = 2 (0.467) and k = 4 (0.421), and k = 3 was selected for its policy interpretability; the advanced cluster is weakly separated (mean silhouette 0.165). The Tirtajaya value (54.71% not yet in school) could not be confirmed against a downloadable provider record; no correction or sensitivity result is claimed. The segmentation offers a screening basis for equity planning, particularly for the Lagging cluster, but represents one year and one dimension of attainment.