Research
Projects
Current Projects
Goal: Identify and characterize evolving HIV hotspots across multiple African countries
Methods: Geostatistical interpolation, Bayesian smoothing, LLM-assisted text mining of health survey reports
Partners: University of Cincinnati, Harvard, local ministries of health
Output: Targeted maps for national program use; publications in Lancet HIV, BMJ Global Health
Goal: Develop adaptive dashboards powered by machine learning to inform health agencies in real time
Methods: LLM-assisted EDA, unsupervised clustering of health behavior datasets, temporal anomaly detection
Impact: Pilot in U.S. urban settings for syndromic surveillance and behavioral intervention planning
Goal: Analyze joint burden of HIV and non-communicable diseases using population-level geospatial data
Methods: Hotspot mapping, spatial autocorrelation, multilevel modeling
Use Case: South Africa, Zambia
Past Projects (Selected)
Identified critical spatial disparities in HIV testing using DHS data. Supported targeted resource allocation for provincial health planners.
Developed one of the first mechanistic models for disease interaction at a systems level. Helped shape understanding of interdependent disease burdens.
Created fine-scale geospatial analysis of ART access, diagnosis rates, and viral suppression. Published results directly influencing subnational HIV programs in SSA.
Proposed and validated new epidemiological indicators to prioritize intervention zones. Widely cited in strategy documents and used in dashboarding projects.