- Crime is now a concern and concern for the community, especially those in the Pematangsiantar Region. Acts of crime that often occur are murder, theft, drugs, rape. With the rampant crime in Pematangsiantar City, it is necessary to group each region using the K-Means algorithm. The data source in this study is a collection of various documents of information Criminal Acts by Pematangsiantar Police Law. The data used in this study are data from 2019 consisting of 6 Districts. In this study, the K-Means algorithm is used to find out areas that have high crime rates and low crime rates areas.Keywords - Crime, K-Means, Clustering, Data Mining Abstrak - Kejahatan saat ini menjadi perhatian dan ke khawatiran bagi masyarakat terutama yang ada di Wilayah Pematangsiantar. Aksi kejahatan yang sering terjadi ialah pembunuhan, pencurian, narkoba, Pemerkosaan. Dengan maraknya kejahatan di Kota Pematangsiantar maka di perlukan pengelompokkan tiap daerah menggunakan algoritma K-Means. Sumber data pada penelitian ini merupakan kumpulan dari berbagai dokumen-dokumen keterangan Aksi Kriminalitas oleh Hukum Polres Pematangsiantar. Data yang digunakan pada penelitian ini adalah data dari tahun 2019 yang terdiri dari 6 Kecamatan. Dalam penelitian ini algoritma K-Means digunakan untuk mengetahui daerah yang memiliki tingkat kejahatan tinggi dan daerah tingkat kejahatan rendah.Kata Kunci -  Kejahatan, K-Means, Clustering, Data Mining

One of the problems related to population that still has to be faced by Simalungun is the problem of the imbalance in the distribution of the population. Incomplete division of the population brings problems to population density and population pressure in an area. This study uses data sources from the Central Statistics Agency (BPS) Simalungun. The data used in this study is data from 2015-2019 which consists of 32 Districts. Therefore, the researchers used the K-Means algorithm in clustering 32 sub-districts in Simalungun Regency. The data will be processed by clustering in 3 clusters, namely clusters with high population levels, clusters with moderate population levels and clusters with low population levels. The iteration process takes 5 times so that the results obtained are 4 sub-districts with high population level clusters (C1), 11 cluster sub-districts with moderate population level (C2) and 17 cluster sub-districts with low population level (C3)