Step-by-step TMLE implementation. Covers influence functions, fluctuation steps, and variance estimation with worked examples.
Causal Inference & TMLE
Doubly robust methods, AIPW, and TMLE in Stata using the
eltmle command with real-data examples.Derivación matemática de TMLE: funciones de influencia, teoría semiparamétrica, y estimadores doblemente robustos.
Variance estimation using the delta method for epidemiological estimands. R notebook with derivations and worked examples.
Stata simulation comparing
eltmle vs competitors under dual misspecification.Open-source Stata module for TMLE estimation of the average treatment effect with SuperLearner integration.
Survival Analysis
Kaplan-Meier estimator, log-rank tests, and Cox proportional hazards models. ULB workshop tutorial.
Standardization of net survival in a cohort analysis. For Stata users working with population-based cancer data.
Modeling net survival using STRS and Poisson regression in a cohort analysis under the relative survival framework.
Royston-Parmar flexible parametric survival models for net survival estimation. Extended tutorial with Stata code.
Machine Learning
K-fold cross-validation, LOOCV, and bootstrap resampling for model selection and evaluation. Interactive R notebook.
Stata module to compute cross-validated area under the ROC curve with bootstrap confidence intervals.
Interactive Shiny Apps
Calculate expected delivery date with gestational age uncertainty intervals.
Visualize comorbidity patterns among cancer patients using interactive network analysis.
Explore parametric survival distributions (Weibull, log-normal, Gompertz) and their hazard functions interactively.
Interactive demonstration of collider bias and selection bias effects in causal inference and epidemiology.
Stata Packages
Targeted Maximum Likelihood Estimation for Stata. Implements TMLE for the average treatment effect with SuperLearner ensemble learning.
Cross-validated area under the ROC curve for binary outcome predictors. Bootstrap confidence intervals and k-fold CV.
Tabulation of matched pairs in 1:1 matched case-control studies by exposure status. Simple, fast, and well-tested.