Publication-Grade Analysis: Multilevel Modeling & Complex Survey Data
The Challenge:
Handling massive, hierarchical datasets (DHS data) covering a 19-year period (2000–2019) to identify trends in public health. The data required complex cleaning and accounting for nested clusters (children within households).
The Statistical Solution:
As the Lead Statistician and First Author, I executed a Multilevel Logistic Regression (Mixed-Effects Model) to handle the data structure. I managed the entire pipeline:
✅ Data Cleaning: Merging large-scale datasets from multiple years.
✅ Advanced Modeling: Accounting for random and fixed effects.
✅ Visualization: Creating clear trend lines and odds-ratio tables.
The Result:
The manuscript was accepted and published in Frontiers in Nutrition (5.1 Impact Factor). This project demonstrates my ability to turn messy, complex data into rigorous, publication-ready insights.
Challenge: Analyzing 19 years of messy, hierarchical DHS survey data.
Solution: I executed a rigorous analysis and published it in a high-impact journal.