Open kamicut opened 5 years ago
Comparing countries.sql
(visually) with countries.geojson
(jq -r '.features[] | .properties.ADM0_A3_US + " " + .properties.NAME_LONG' src/analytics/src/main/resources/countries.geojson
), it looks like the names are coming from the NAME_LONG
attribute.
In some cases, it looks like I manually edited things, e.g. Heard Island and McDonald Islands
(NAME_LONG
is Heard I. and McDonald Islands
; NAME_SORT
is the value I took, but that doesn't match the names in other cases).
In case it's relevant, the names of the regions in the State Department border geometries are here: https://gist.github.com/jpolchlo/54501716a7d105de3fa93361e86e1302
If we move to those boundaries, we'd have a hard time associating a short, 3-letter code with these regions, as they don't always correspond to ISO countries. We'd probably need to just put the name in the DB, or create a separate table with IDs of our choosing.
Right now osmesa-stat-server pulls country names from a different source than OSMEsa's
countries.geojson
. Ideally we should have the same names coming from a single source of truth.