kmed: An R package for reproducible distance-based k-medoids clustering

k-medoids clustering provides a robust alternative to k-means when data contain outliers or non-Euclidean dissimilarities. We introduce kmed, an open-source R package providing a reproducible pipeline for various distance formulations, k-medoids variants, validation techniques, and visualizations. The package emphasizes flexibility and ease of use for researchers integrating k-medoids into reproducible research workflows. kmed implements established, peer-reviewed algorithms widely applied across statistics and machine learning. By providing a standardized R framework, kmed facilitates reuse and promotes reproducibility in clustering-based research, supporting both exploratory and confirmatory data analysis. • Provides a reproducible R framework for distance-based k-medoids clustering. • Supports numerical, categorical, and mixed-type distances in a unified workflow. • Integrates clustering, validation, and visualization in open-source software. • Enables transparent and reusable clustering analyses across scientific domains.