Improved place ontology

Places are currently classified with [type, class] (eg. 'admin', 'neighbourhood'), this ontology is not currently being used, other than for display.

Some more thought about a simple-yet-flexible place ontology could be very beneficial for discovering and filtering of places within the system.

In particular, it would be ideal if we could support a 'place rank' concept, so that discrete data sources could map their ontology to a common ranking dimension in a way that allows comparing two places:

• The lower rank represented the parent
• The higher rank represented a child
• Matching ranks represent sibling places at the same level in the hierarchy

I suspect this will be fairly difficult to do for OSM without first having the country-code, the semantics of their administrative_level tag integer values vary based on the country.

ref: https://wiki.openstreetmap.org/wiki/Tag:boundary%3Dadministrative
ref: https://github.com/whosonfirst/whosonfirst-placetypes

Start by reviewing how places currently store the [type, class] values and how that ontology is used for display; no implementation file or test is named. Compare the OpenStreetMap administrative_level reference with the Who's On First placetypes reference, then define how a shared parent, child, and sibling rank model would support discovery and filtering.