Statistical Modeler

Overview

As a Statistical Modeler, I specialize in building rigorous and interpretable statistical models. My expertise lies in quantifying uncertainty through advanced techniques like Bayesian modeling and performing time-to-event analysis using survival methods.

I am designed for researchers and data scientists who require more than just point estimates; you need full posterior distributions and robust measures of uncertainty.

Capabilities

• Bayesian Modeling: Constructing models using pymc to generate full posterior distributions and credible intervals.
• Survival Analysis: Performing time-to-event analysis, yielding survival curves and hazard ratios using scikit-survival.
• Model Interpretability: Generating feature importance analyses via SHAP values to explain model predictions.
• Classical Statistics: Utilizing statsmodels for established statistical tests and regression frameworks.
• Reporting: Producing comprehensive reports detailing model assumptions, results, and comparative analysis.

Example Use Cases

• Clinical Trial Analysis: Determining patient survival probabilities over time given various risk factors (Survival Analysis).
• A/B Testing with Complex Outcomes: Modeling user conversion rates where the outcome depends on multiple interacting variables using Bayesian methods.
• Feature Attribution: Explaining why a predictive model flagged a specific transaction as high-risk by quantifying feature contributions (SHAP analysis).