uSTAT — Browser-based Statistical Analysis Platform
uSTAT is a free, browser-based statistical analysis platform aimed at clinicians, biostatisticians, and medical researchers. It spans hypothesis testing, regression (linear, logistic, Firth, ordinal, Poisson, negative binomial, GEE, mixed models, restricted cubic splines), and survival analysis including Cox with time-varying covariates, Kaplan-Meier, Fine-Gray competing risks, RMST, landmark analysis, shared frailty, recurrent events, and multi-state models with dynamic prediction. Causal-inference tools cover propensity score matching with Austin 2011 diagnostics, IPTW, instrumental variables, mediation, difference-in- differences, regression discontinuity, and sensitivity analysis for unmeasured confounding. It also provides diagnostic accuracy (ROC with DeLong comparison, calibration, decision-curve analysis, NRI/IDI), method agreement, meta-analysis, missing-data handling with MICE and MNAR sensitivity, factor analysis, machine learning, time series, Bayesian tests, Table 1, and power calculations. Reported statistics are cross-validated against SciPy, statsmodels, and lifelines, and every analysis emits publication-ready methods text plus the equivalent R code. Data are held in memory on the server with a short TTL and no persistent storage by default.