<![CDATA[Competition in streaming music services requires a better understanding of user behavior and churn tendencies. This study aims to segment Spotify users and analyze churn patterns based on demographic characteristics and service usage behavior using the K-Prototypes method on mixed-type data. The optimal number of Cluster was determined using the Elbow method, while the churn variable was used to evaluate the clustering results. The analysis shows that three user cluster were formed with distinct characteristics. The first cluster represents younger premium users with relatively high usage intensity, the second cluster represents student users with the highest churn proportion, and the third cluster represents free users with high ad exposure but the lowest churn proportion. These findings indicate that the K-Prototypes method is effective in grouping Spotify users and provides useful information for understanding user behavior and churn tendencies.]]>
Copyrights © 2026
