Dataset for Feature Importance Analysis in the Construction of the Composite Dengue Mortality Risk Index in Mexico
Description
This dataset contains the integrated clinical, environmental, and territorial variables used in the feature importance analysis for the construction of the Composite Dengue Mortality Risk Index in Mexico. The epidemiological data were obtained from the Open Data Portal of the General Directorate of Epidemiology (DGE) of the Mexican Ministry of Health. Environmental variables were derived from satellite-based remote sensing products accessed through Google Earth Engine, and sociodemographic indicators were obtained from the National Institute of Statistics and Geography (INEGI). The dataset integrates clinical variables (demographics, comorbidities, and hospitalization indicators), environmental predictors (temperature, precipitation, humidity, vegetation index, terrain and hydrological indicators), and territorial variables (population size, density, territorial area, and human exposure index). These variables were used as predictors in a machine learning feature importance analysis to identify the main determinants associated with dengue-related mortality risk across the 32 federal entities of Mexico. The dataset does not contain personal identifiers and was prepared exclusively for research and reproducibility purposes.
Files
Steps to reproduce
The dataset was constructed by integrating epidemiological, environmental, and sociodemographic variables associated with dengue cases in Mexico. 1. Epidemiological data were obtained from the Open Data Portal of the General Directorate of Epidemiology (DGE) of the Mexican Ministry of Health. 2. Environmental variables (temperature, precipitation, relative humidity, and NDVI) were derived from satellite-based remote sensing products using Google Earth Engine. 3. Additional terrain and hydrological indicators were computed using digital elevation models and spatial datasets. 4. Sociodemographic indicators were obtained from the National Institute of Statistics and Geography (INEGI). 5. All variables were cleaned, harmonized, and integrated into a single analytical dataset. The resulting dataset was used as input for a decision-tree-based Feature Importance algorithm implemented in Python to identify the predictors most strongly associated with dengue-related mortality risk. The dataset provided in this repository corresponds to the final version used in the Feature Importance analysis described in the associated research article.