For discussion: Reshaping Election dataset

[acbart]

Instead of a list of location's votes, it might be better to do a list of locations/candidates.

So every row would have the location and the particular candidate. We'd also include candidate specific information (possibly just "affiliation": {"R", "D"}). This would allow students to do more interesting things with the visualizer and possibly python.

[acbart]

(this isn't a solid decision, want to bring it up for discussion).

[RyanWhitcomb-VT]

Leave structure as it is, would be better if it had the month of the primary and population of each county.