pascalantwi/Thesis-Code

Spatio-Temporal Generalized Linear Modeling of Climatic Influences on Malaria Incidence in Ghana # Spatio-Temporal Modeling of Malaria Incidence in Ghana This repository contains the code and analysis for my Master's thesis: **"Spatio-Temporal Generalized Linear Modeling of Climatic Influences on Malaria Incidence in Ghana"** (Strathmore University, 2025). ## Overview Malaria remains a major public health challenge in Ghana. This study quantifies the lagged effects of climatic variables on monthly malaria cases (2020–2024) across 16 regions using a combination of Generalized Linear Models (GLMs) and Bayesian spatio-temporal models. ## Key Features - **Data**: Monthly confirmed malaria cases (Ghana Health Service) and climatic data (rainfall, temperature, humidity, dew point, solar energy, UV index, etc.) from Visual Crossing. - **Methods**: - Overdispersion testing → Negative Binomial regression. - Optimal lag selection (0–3 months) via correlation analysis. - Bayesian spatio-temporal modeling with INLA (CAR spatial prior, RW1 temporal prior). - Hotspot detection using Local Indicators of Spatial Association (LISA). - **Main Findings**: - Rainfall positive at lags 2–3 months; temperature immediate positive. - UV index at lag 1 strongly positive; solar energy protective. - Dew point inversely associated at lags 0–1. - Persistent high-risk clusters: Upper West, Upper East, Bono, Ahafo, Western North. - Ashanti as a low-high anomaly (low-risk surrounded by high-risk neighbors).