Exploratory Regression Analysis helped narrow down the most significant variables. Explanatory Regression also helped identify the best models (variable combinations) to run against the dependent variables. Geographic Weighted Regression (GWR) was only conducted after identifying best models through OLS. very high correlations for 2016 Republican votes between College degree & Blue Collar were achieved. We anticipate moderately high correlations for 2024 Presidential Democratic votes based on vaccination rates alone. However, further analysis will be needed to identify which counties have the highest correlations. SVI index doesn’t seem to be as highly correlated at predicting for 2024 Presidential Democratic votes.

The data analytic approach for this work was designed with the intention to solve the defined problem. The thorough data analysis that was performed in this project was conducted with extensive datasets, appropriately fit to better understand county-level communities across the United States. I believe that the results are conclusive and effective in solving the proposed problem & answering the research questions. The results from this project have improved my understanding of unique & evolving community characteristics in the US at the county-level. Some trends that have been identified from this analysis regarding government trust and compliance are a bit alarming. It seems there are “resistance areas” where vaccination rates are low (less than 20%). Perhaps this analysis can be helpful to health & human services stakeholders in addressing this problem. Stakeholders such as political actors, municipal organizations, non-profit and philanthropic businesses, and commercial & private organizations should be able to benefit from this analysis work as well.
It is with optimism that future data scientists and practitioners find inspiration from this work to further expand this research.

Here is the ArcGIS link to all Maps for this project
ArcGIS Online Maps Project files

Here is the link to the Map generated after executing Geographic Weighted Regression (GWR) on the 2016 Presidential Election Data. The dependent variable we are trying to predict is the Percent Republican Vote and the control variable is the percent with College degree for the US Counties

AecGIS Battle Ground States 2016