Vis-Health: Exploratory Analysis and Visualization of Dengue Cases in Brazil

This paper presents a tool, Vis-Health, for analysis and visualization of public health data covariance with variables chosen by the user, in order to give her/him clues for a better understanding of disease occurrences. Vis-Health is a system that allows: (a) the government authorities store data of different nature and from different methods, (b) the user (common citizen or government) to choose an interest variable as dependent(output) and n other as dependent (or inputs), in order to study the covariance study between them and (c) visualize the result in a clear way. The covariance study is made by the use of Principal Component Analysis (PCA), resulting in a linear orthogonal regression. The method permits that many users, choosing different variables as dependent or independent, can use Vis-Health simultaneously, thanks to the optimized performance of the included algorithm based on PCA. Vis-Health is an open tool, and can help government to have more adequate actions, but any common citizen can use Vis-Health for making her/his particular exploration. As a illustration, a study of the numbers of cases of Dengue in Brazil is studied as a function of climatic variables.