Our dataset provides estimates of prevalence (and associated confidence intervals) for various health measures in communities (i.e., census tracts) across all states (based on modeling). As such, we plan to explore how state, preventative health behaviours (e.g., doctor's visits), mental health, and smoking predict model-estimated prevalence of obesity, cancer, and stroke. We will use scatterplots, correlation matrices, bar graphs, and multiple linear regression to explore the relative contribution of each of these predictors to the outcomes of interest.