NetPsy: An Open-Source Interactive R/ShinyWeb Application for Undirected Psychological Network Analysis

Psychological network analysis has gained considerable attention in the last years, yet the requirement of R programming still excludes many potential users. We present NetPsy, a free and open-source R/Shiny web application that wraps the full pipeline of undirected network analysis behind a graphical interface, from data import through automated reporting. Users can upload CSV or Excel files, run Goldbricker redundancy checks (which flag item pairs that may be too similar to distinguish empirically), and estimate networks for continuous data (Gaussian Graphical Models with EBICglasso or ggmModSelect), binary data (Ising models), or mixed data (mgm). The application computes standard centrality indices such as strength and expected influence, together with bridge centrality measures relevant to comorbidity research. Bootstrap routines provide case-dropping CS coefficients and nonparametric confidence intervals for edge weights. Group comparisons rely on the Network Comparison Test, reporting invariance statistics for network structure (M), global strength (S), and individual edges (E) with Holm correction. Visualization uses Fruchterman–Reingold layouts with adjustable parameters, and an optional module generates DOCX reports assisted by large language models. We illustrate these capabilities with simulated data and compare NetPsy against JASP Network and direct R scripting in terms of coverage, accessibility, and reproducibility.