Big data is allowing unprecedented views of information from large deposits of data that only a few years ago were too large to analyze and, in particular, visualize. Without the ability to visualize large amounts of data, important relationships may be missed that could have the potential of affecting public health policy and practice. Currently, we are analyzing a large repository of data called the Exposome which has over 20,000 variables. For this talk, we analyze a much smaller subset of 62 variables. Needless to say, the impact of the analysis has the potential for allowing identification of critical points of disease clusters that would benefit from strategically placed interventions to help at-risk populations. Our study considers the at-risk population of Cardiovascular Disease (CVD) at the detail of the county level. We look at CVD indicators across a geographical space (i.e. Interstate highway) to expose differing levels of factors through areas of small and large populations. We propose a novel visualization for the 62 variables of the 22 counties along interstate 20 represented as quintiles to find the diffusive relationships among the variables. Diffusion is defined as the effect of a variable across a geographic boundary. Intuition suggests that the variables should have a smooth diffusion from west to east. Smooth diffusion patterns within a variable might provide further opportunity to reduce the complexity of the model, by removing the variable. Sharp diffusion occurrences may alert of potential errors that need to be mitigated within the predictive model constructed from the data. Sharp diffusion occurrences may be counted as effective test cases for predictive models for hardening against noisy data. Our observations show 15 variables can be removed from the set since they have smooth diffusion patterns. The rest of the variables at least shows one occurrence of sharp diffusion.