Demographic and behaviour clustering on chicken meat’s seller using K-means method

The supply chain for halal chicken meat in Indonesia is a unique and more complex than in other countries since there are so many actors who take a part in each chain. Study on behaviour for breeders, distributors, sellers and customers become essential to be carried out to obtain a comprehensive understanding of the critical point of halal chicken supply chain. The common study only discusses segmentation behaviour on the customer's perspective, while the critical point of halal lies more crucially on the sellers. Therefore, reviewing the sellers becomes important in the halal chicken meat supply chain. This paper presents a computational segmentation of halal chicken meat’s sellers using a clustering method. The clustering is based on demographic and behavioural attributes. The essential attributes of demographic cover gender, age, education, selling location, selling quantity, selling duration, and source of chicken while behaviour attributes are adapted from Theory of Planned Behaviour framework that involves knowledge, attitude, intention, and action. Data were collected through a quantitative questionnaire from 249 chicken meat’s sellers in Yogyakarta province in Indonesia. K-means cluster is used to group seller with the same interest in attributes and the number of clusters is optimized using the Elbow Method. From the calculation of Elbow Method, there is a sharp decline and shapes an angle between k=2 and k=3, so chicken meat sellers are grouped into three clusters. The composition shows that about 23.2% is included in the apathetic cluster, 45.7% belongs to religious, and 31.1% fits to rationalist. Cluster 1 and 3 have insignificant difference in halal behaviour but not with cluster 2. Based on analysis, the level of education is the most important demographic sub attributes in configuring halal behavioural pattern.