Mapping of Climatic and Environmental Factors for the Early Detection of Arbovirus Outbreaks
Project leader: Mauro César Cafundó de Morais
This project aims to develop an early warning and response system for arbovirus outbreaks. The approach is based on the integration of publicly available data from climatic sources—such as temperature, humidity, and precipitation—with health outcomes, using artificial intelligence tools.
The main objectives are:
Integrate multidimensional data: collection and organization of a geospatial and temporal database encompassing information on human cases, vectors, hosts, and environmental factors.
Identify transmission patterns: application of supervised and unsupervised learning techniques to detect spatial and temporal associations between environmental variables and case occurrence.
Validate predictive incidence models: training of machine learning models, such as Random Forests and Neural Networks, to generate dynamic risk maps and predict areas of higher transmission.
To address scientific challenges related to the quality and granularity of health data, a robust preprocessing phase will be implemented, using statistical methods to handle underreporting and missing data. In addition, the interpretability of predictions will be ensured through the use of Explainable AI (XAI) tools. The final outcome will serve as a public health decision-support tool, enabling the strategic allocation of resources and being replicable for the monitoring of other climate-sensitive diseases.