Intelligent K-Median Clustering For Region Segmentation of Coffee Production and Consumption in Indonesia

A coffee drink is one of the popular beverages made from coffee beans. Indonesia is one of the countries that produces coffee fruits. Coffee beans are roasted and ground into coffee powder, then brewed with boiling water to make a coffee drink. The coffee beans and the instant coffee drinks have a prominent market. This paper implements Intelligent K-Median clustering to cluster districts in Indonesia according to the coffee production and consumption. The advantages of Intelligent K-Median is able to cluster a dataset without determining the number of clusters. It uses three datasets are coffee production, coffee bean consumption, and instant coffee consumption. The Coffee production from the Ministry of Agriculture of the Republic of Indonesia from 2015 to 2022. The coffee consumption datasets are provided by the Indonesian Central Statistics Agency from 2018 to 2024. Intelligent K-Median clusters the coffee production data into 3 clusters with an average Silhouette score of 0.437. Clustering consumption of coffee beans and instant coffee, each obtains 2 clusters with average Silhouette scores of 0.483 and 0.765, respectively. Finding clusters of districts in Indonesia is beneficial in identifying the districts with high potential for developing coffee production and market segmentation. The region segmentation based on the coffee bean production finds that the high coffee producers are from the districts in Suma- tra (e.g., Lahat, OKU Selatan, Empat Lawang, Lampung Barat, and Tanggamus) and Java (e.g., Temanggung, Malang, Jember, Banyuwangi). The places with high consumption of instant coffee are regions in West Java and Jakarta. Meanwhile, regions with high consumption of coffee beans include some parts of Sumatra, Java, Kalimantan, and Sulawesi.