LatLon:Exists with point to point continues to create confusion in facet filter

[cncoleman]

Since the facet is used not only as a filter but also as a way to measure values, the LatLon:Exists presents problems when we have two potential coordinate points for each letter and we don't know which one is missing. The better solution—in the facet filter— is to use the Source facet and the Destination facet independently to explore blank values for each or for both.

This does not solve the problem of showing "unmappable records" in the timeline and timespan. For both the timeline and timespan we need the option "Group by" "empty source places" "empty target places" "empty source+target". It is definitely inelegant but at least it would show the comparative distribution of letters that cannot be mapped, which is the goal.