There's actually no correlation between the differential we're talking about here and the results in 2014, for what it's worth. Where there is correlation is with the following map: Hispanic population.
If you redo the first map, taking out everyone but the non-Hispanic whites, the map gets a lot more purple (purpler?) -- meaning that those congressional districts (which is what we're talking about in this example) have more old whites than young whites. The notable exception is Utah.
Notice that the scale on that map is the same as the first one: the darkest parts are all 22-percentage-points-plus more people under 25 or over 55.
If we do the same map for Hispanics, it looks like this.
(Those light-colored areas in West Virginia and Kentucky didn't have data from 2013.)
It blows out the scale. Only a few urban areas and the lower tip of Florida have districts where there are more older Hispanic people than younger ones. If we adjust the scale to show more degrees of youth, we can see the distribution more easily.
These maps obscure a lot of information. In many of the areas above, there aren't very many Hispanics, for example. This looks at the entire population, not just those eligible to vote. There are also still far more white people in America, and, as we saw in 2014, those older white voters turn out to vote far more.
But a lot of Republican districts have an over-55 heavy white population and an under-25 heavy Hispanic one.
That's not a big political deal -- yet.