538.com has a remarkable tool for showing how differing patterns of participation and partisanship by different groups of people (blacks, Latinos, "Asians and others", "college-educated whites", and non-college-educated whites would vote in the 2016 Presidential election.
http://projects.fivethirtyeight.com/2016-swing-the-election/Here's 2012:
(I am not showing the districts of Maine and Nebraska because 538.com does not mention them).
The "Asian and other" category includes Asians and Pacific islanders, primarily; the second-largest group of them are what Canadians call First Peoples -- Native Americans and Alaska natives. These are heavily concentrated in non-swing states (AK, AZ, CA, HI, NM, OK, OR, TX, WA) generally nowhere near being close in 2012. Cut their participation to zero or raise it to 100%, and one gets roughly the same electoral results as in 2012. No map needed.
I start jiggling the amount for educated white people, and states start appearing in the Democratic column. At a 52R-48D, North Carolina goes D. At an even split, Georgia goes D. Around 55%, one starts to see Missouri, Indiana, Arizona, and South Carolina go D. At 57% among educated white voters, only two states slip away from the Republicans, but one of those is Texas.
2012 wins for the Democrat 332
at 50-50 among educated white voters (40% saturation) 363 (Obama got 365 in 2008)
at 55D-45R among educated white voters (30% saturation) 404
at 57D-43R among educated white voters (20% saturation) 448
I'm not going to show it on the map, but if well-educated white people voted like Asian-Americans, Hillary Clinton would end up with about 500 electoral votes and all states except Idaho, Utah, Wyoming, Oklahoma, Arkansas, Kentucky, and West Virginia.