Modeling the Effects of Meteorological and Geospatial Patterns on Malaria Cases in Uganda

Malaria is one of the leading causes of public health issues in Uganda, with outbreaks becoming increasingly prominent during rainy seasons. Stagnant water and high humidity create optimal breeding and survival conditions for mosquitoes to thrive, leading to spikes in transmission cases. This study investigates how short and long term weather fluctuations and geospatial environmental features, specifically dewpoint, humidity, precipitation, and elevation correlate with weekly malaria case counts across the regions of Uganda from 2020 to 2025. Using modeling tools such as Linear Regression, Random Forest, HistGradientBoost, and XGBoost, this study evaluates whether the inclusion of temporal-environmental features improve the accuracy of models trained on historical cases. Initial results show a minute, but positive trend with the inclusion of geospatial embeddings, suggesting that environmental features can be an early predictor for outbreaks in high risk regions. With an emphasis on lagged case data and environmental predictors, these models demonstrate the potential of short-term forecasting for cases with the goal of supporting an early outbreak warning system, ultimately highlighting the value of combining epidemiological, meteorological, and geospatial datasets for anticipating disease transmission patterns. Future development in this project can provide clinicians and public health practitioners tools to rapidly develop interventions strategies in high-risk regions before outbreaks occur.