Missing some US Address data (Hawaii and NULL address_level)

[mlptownsend]

I noticed a few data gaps with the address set and I wanted to at least point them out.

SELECT 
	count(*) 
FROM
	read_parquet('s3://overturemaps-us-west-2/release/2025-04-23.0/theme=addresses/type=address/*', filename=true, hive_partitioning=1) 
WHERE
	country = 'US' 

There are no addresses with a state code for Hawaii. I also tried a bbox around Honolulu and the surrounding area and got nothing.

SELECT 
	*
FROM
	read_parquet('s3://overturemaps-us-west-2/release/2025-04-23.0/theme=addresses/type=address/*', filename=true, hive_partitioning=1)
WHERE 
	bbox.xmin > -158.31024169921878 AND bbox.ymin > 21.236902399759188 AND bbox.xmax < -157.62290954589847 AND bbox.ymax < 21.736539949711823 

Additionally, there are 216,558 records with country 'US' where there is no address_level. Range-wise, they run from minx,miny,maxx,maxy -106.288967840741, 41.0059183455904, -71.9555035789288, 45.0081344055916. Kind of sprinkled around here and there. I want to use address_level[1].value to avoid having to do a ST_Within on the state's geometry from division areas. That'll take forever to run.

SELECT 
	count(*) 
FROM
	read_parquet('s3://overturemaps-us-west-2/release/2025-04-23.0/theme=addresses/type=address/*', filename=true, hive_partitioning=1) 
WHERE
	country = 'US' 
AND 
	address_levels[1].value is null

And of the 110,819,056 US addresses, 33,177,156 have no secondary address_level / city designation.

SELECT 
	count(*) 
FROM
	read_parquet('s3://overturemaps-us-west-2/release/2025-04-23.0/theme=addresses/type=address/*', filename=true, hive_partitioning=1) 
WHERE
	country = 'US' 
AND 
	address_levels[2].value is null

[jeffdefacto]

Addressess are a work in progress still. Hawaii's counties do have open data but we haven't been able to get a clear usage license for any of it yet. Likewise, much of the US is a collage of various sources, some of which is more complete than others. We're always striving to get better data but it doesn't always exist out there yet.