Complex-survey epidemiology in Python (NHANES-style): design weights, strata and PSU clusters, Taylor-linearized variance, survey-weighted logistic regression with design-corrected sandwich standard errors (like R survey::svyglm), weighted Table 1, and odds-ratio tables. Validated by recovering known true effects on a synthetic multistage cohort.

Open-source R tools for complex survey analysis, public health indicators, and spatial small-area estimation using Peruvian population surveys.

Simulates complex survey designs and applies Raking (IPF) calibration to reduce bias in demographic and market research studies.