There are 243,557 distinct combinations of interpreted_countrycode and v_stateprovince in where v_stateprovince is not null in gazetteer.locations_distinct_with_scores. Given that there should be on the order of 3700 first order subdivisions this represents a huge potential for normalization and better matching.
What would be useful here is a lookup table containing interpreted_countrycode plus v_stateprovince and interpreted_stateprovince.
There are 243,557 distinct combinations of interpreted_countrycode and v_stateprovince in where v_stateprovince is not null in gazetteer.locations_distinct_with_scores. Given that there should be on the order of 3700 first order subdivisions this represents a huge potential for normalization and better matching.
What would be useful here is a lookup table containing interpreted_countrycode plus v_stateprovince and interpreted_stateprovince.