Interesting how much of the Pacific Northwest is graded lower than a lot of the Midwest, even excluding the Great Plains. Wouldn't expect that.
They use a weighted average. (I used unweighted in my rankings mostly because I'm lazy, but also because I can't justify all of their weightings. Some subjectivity always creeps into such decisions.)
Their big weighting (15%) is from local educational attainment. The next three are public schools (10%), cost of living (10%), and housing (10%). The upper midwest (IA, MN, WI) and New England (ME, NH, VT, CT, RI, MA) generally have the best public schools on average, when ranked by achievement scores and the like, and those areas have generally high educational attainment, so the weighting accounts for much of that.
What strikes initially me about the map is that relatively densely populated areas are blue, and less densely populated areas are not. Consider the ten most populous cities in the US. All of them are in the blue-colored counties. I think that has a great deal to do with their algorithm as well. While they are not weighted as heavily, jobs (7.5%), diversity (7.5%), nightlife (5%), and activities (5%) all show up in the algorithm. There's also a totally subjective "composite overall score" (5%, reported by users) that shows up. Collectively those things would give the edge to cities and densely populated areas over rural areas and less densely populated areas.
Finally, there's a weather grade (5%) which gives an edge to sunny climates. This may also help explain another striking feature of the map: Red Alaska.