SilhouettePlot: An R-based app integrating silhouette scores into network plots for enhanced clarity and meaning

Network visualizations of author collaborations are widely used in academia, yet traditional graphs often provide limited insight into the cohesiveness of research teams. The Silhouette Score (SS), typically applied to evaluate clustering quality, has rarely been used to classify patterns of team collaboration. This study explores the feasibility of applying SS to author collaborations in SoftwareX articles published between 2017 and 2025. Affiliations of first and corresponding authors were analyzed, and collaboration patterns were categorized using silhouette scores into three groups: strong consistency (SS > mean SS), weak consistency ( 0 < SS ≤ mean SS), and divergence (SS < 0 ). An R-based application was developed to generate the SS plot. Results revealed 11 major clusters among the top 20 contributing countries, led by the United States, Germany, the United Kingdom, China, and France, with SS values of 0.75, 0.83, 0.64, 0.77, and 0.39, respectively, and an overall SS of 0.65. The United Kingdom was closest to Australia, while the United States collaborated more with Germany and also demonstrated a higher edge dominance coefficient (0.71) within its cluster. These findings highlight SS as an effective measure of cluster cohesiveness in research collaboration and suggest its potential application to an author’s publication profile across their career.