Syndat

Syndat computes quality metrics for synthetic tabular data and visualizes how closely it reproduces the real data. It implements three base metrics: Jensen-Shannon distance for feature distributions, normalized correlation difference for inter-feature relationships, and discriminator AUC from a cross-validated Random Forest classifier, and aggregates them into normalized 0-100 distribution, correlation and discrimination scores. It also provides stratified visualizations (raincloud, GOF, trajectory plots) for clinical subgroups.