Clustering Crime-Prone Areas in Indonesia Using the K-Means Method

Chairani, T.Sofia (Universitas Negeri Medan)
Listia, Hijka (Universitas Negeri Medan)
Wardaniah, Sabina (Universitas Negeri Medan)
Wulandari, Siti (Universitas Negeri Medan)
Tampubolon, Putri Tasya Agustina (Universitas Negeri Medan)
Piliang, Arnita (Universitas Negeri Medan)

Bit (Fakultas Teknologi Informasi Universitas Budi Luhur)
Vol 21, No 2 (2024): SEPTEMBER 2024 (IN PRESS)

Crime in Indonesia is a significant problem that affects various aspects such as security, social, and economic. However, mitigation efforts are often hampered by a lack of structured information about crime-prone areas. This study aims to overcome this problem by grouping provinces in Indonesia based on crime patterns using the K-Means Clustering algorithm. Crime statistics data for 2014-2023 from the Central Statistics Agency (BPS) were analyzed by determining the optimal number of clusters that resulted in five clusters with evaluation using Python and the Scikit-learn library. The results showed a Silhouette Score of 0.593, which reflects the formation of a fairly good cluster. This clustering provides data-driven guidance for the government in developing more targeted security policies to reduce crime rates in Indonesia.