So, I've finished version 2.0 that breaks things down by race.
One issue I've ran into is that since I don't always have county level information for Asian and Hispanic groups, I just apply a statewide weight (i.e. a boost for Trump if there are more Vietnamese in a state, a detraction if there are more Hmong, as an example). for the remaining counties. This sometimes leads to "choppy" borders (where you can tell the boundary of a state). Accordingly, there might need to be some revision at work, though it might also be justified for some states.
Here are all the maps I've made for this project. I apologize for the current colors (I post these maps on the other websites I'm on, where everyone else is accustomed to a Red=Republican scheme). I'll change them to Atlas Colors once I finalize this.
Any suggestions would be appreciated!
I would suggest that you screen counties for no data, based on some threshold (1% or 2%) so that for example in rural areas of the Midwest and West you aren't trying to extract black or Asian voting behavior. There is not enough data in these areas to be statistically valid, or alternatively, enough persons of the group as to be relevant.