Through evolving heatmaps of individual disease-risk landscapes across heterogeneity-mapped subpopulations to see chronic disease complexity in U.S society: order-1 effects of major feature's categories

{Categorical Exploratory Data Analysis (CEDA) is used to uncover evolving risk dynamics in chronic disease across heterogeneous subpopulations.} Consider the bivariate dynamics of (Stroke(STK), heart-disease(HD)) to better represent the whole chronic disease in U.S. society. Our data-driven study analyzed the Year 2015 Kaggle-BRFSS database. The evolving disease-risk dynamics across 5×5 subpopulation-map defined via (age, general-health(GenHL)) heterogeneity-axis are precisely computed and graphically displayed. Due to the entirely categorical nature of this database, the computational paradigm called Categorical Exploratory Data Analysis (CEDA) is particularly suitable. Each positive or negative chronic disease risk is expressed as a categorical conditional distribution of (STK,HD) given a covariate feature-category presented by a binary vector of absence-presence statuses across subjects within a subpopulation. Its finite sample confirmation via a minimum sum of Type-I and Type-II errors is evaluated as an overlapping area of the alternative and null entropy distributions. A subpopulation specific collective of all confirmed chronic disease risks constitutes a binary subject-vs-disease-risk bipartite network called heatmap. Five age-specific heatmap-series along the GenHL axis demonstrated evolving chronic disease risk dynamics with fundamental distinctions. The Age3 heatmap-series across the 5 GenHL stages assumed a pivotal role in societal chronic disease dynamics. As byproducts all predictive tasks are carried out based on the multiscale topological structures defined on the subject-axis of heatmaps.