For this map, I found the percentage of each state’s total population voting for each candidate in all presidential elections from 2000 to 2016. The candidate who had the highest percentage (in any election year) is the “winner.”
I realize that it would be more ideal to use the actual eligible voting population, rather than the total population, but I have been unable to find that data further back than 2000, and I am currently working on a second version which will incorporate more past election years. So, for the sake of uniformity, I have gone with total population. I do not believe that this greatly affects the results, at least not in the 2000-2016 time frame.
This map is a little tricky to explain succinctly, so bear with me. What is being shown is the racial/ethnic group in each county that is most disproportionate to the makeup of the United States.
For example, my home of Travis County, Texas, is 50% White, 34% Hispanic, 8% Black, 6% Asian-Pacific Islander and 0% Native American. The United States as a whole is 62% White, 17% Hispanic, 12% Black, 5% Asian-Pacific Islander and 1% Native American. While Travis County is less White, Black and Native American than the US as a whole, it has a higher share of Hispanic and Asian-Pacific Islander residents.
The share of the population which is Hispanic is 2.8x that of the US, and the Asian-Pacific Islander population is 1.2x. Therefore, Travis County is shown on this map as Hispanic, even though Whites constitute the largest racial/ethnic group within the county.
Source for total number of eligible voters: United States Census Bureau American Community Survey
Source for county-level election results (except Alaska): Tony McGovern
Source for estimated Alaska borough-level election results: RRH Elections