A machine learning class project where we predicted prostate cancer survival of patients in the compartor arms of several large clinical trials. Methods used include random survival forests, elastic net Cox models, and gradient boosted survival modals.
Spline-fitting using sparsity-inducing penalties (fusion and lasso), and comparing these methods to penalized splines fit via mixed linear models.
A summer internship project where I ran extensive simulation studies to determine under what conditions standard right-censored survival models were inadequate and mixture cure models might be preferable. Application area was reliability testing of military systems.
A Python tool for parsing and visualization of Facebook messenger conversations.
An R Shiny app that uses dynamically generated Latex and Markdown code to create nicely formatted PDF recipe documents.