Overview
Develop and validate country-level hierarchical Bayesian models for Maternal, Newborn, Child, and Adolescent Health (MNCAH) indicators for UNICEF’s reporting and monitoring. This role will advance estimation processes, including methodology, toolset, estimates, and validation approaches.
Tasks Summary
- Review existing modelling approaches and recommend temporal structures, short-term deviations, autocorrelation structures, and hierarchical random effects.
- Compare time-only, covariate-driven, and multi-source models, assessing suitability for different indicators and data contexts.
- Implement systematic covariate selection strategies.
- Extend the modelling framework for data-rich and data-sparse indicators.
- Explore random-walk and intervention-sensitive models.
- Develop and apply rigorous validation strategies.
- Recommend suitable model specifications for each MNCAH indicator.
- Identify a framework to evaluate model results.
- Convert existing JAGS models to brms/cmdstanr equivalents on Databricks.
- Refactor existing code into a modular, reusable, and transparent codebase.
- Develop a visualization toolkit with reusable plotting utilities.
- Set up a structured GitHub repository with documented scripts and reproducible workflows.
- Document all methodological decisions, assumptions, covariate treatment, and validation processes.
- Produce country-level profiles combining model outputs with annotated data sources and explanatory notes.
- Prepare a final technical report summarizing methods, validation results, limitations, and recommendations.
Experience Requirements
- At least 4 years of experience in statistical modelling
- Demonstrated expertise with Bayesian statistical methods and modelling.
- Knowledge and understanding of key issues and modelling challenges for maternal, neonatal, and child health data.
Qualification Requirements
• Masters in biostatistics, statistics, public health, or related discipline
