add metadata to existing datasets about geosnap HOT 9 CLOSED

thats a good point. The shape data come directly from census, but the table that osnap refers to as cbsa is the county crosswalk, probably from here

from geosnap.

rather than storing the CBSA crosswalk, it might be better to just read it directly when necessary (like we do with the inflation table)

the benefits would be

1. the documentation is implicit since the URL defines the original source (though we'd want to document explicitly as well)
2. we could provide methods for users to call MSAs as defined in other years

2019 cbsa-county definitions

from geosnap.

And an older source for 2003 cbsa definitions

from geosnap.

workflow:

1. store "appropriate version in osnap repo
2. when user calls a dataset, osnap checks the canonical dataset by downloading from the URL
3. osnap checks to see whether stored version matches canonical version
4. if versions don't match, raise a warning, etc.

from geosnap.

if we're checking the dataset integrity with each use, is there any real use in caching it?

from geosnap.

https://github.com/quiltdata/open-data-registry

from geosnap.

we should use this source. It might be best to always grab from the web like we do for the inflation adjustment table, though this would be used more and more prone to failure if the census api goes down. That would also mean you always need an internet connection to make MSA queries

from geosnap.

should be resolved by #215 but give it a look

from geosnap.

https://spatialucr.github.io/geosnap/data.html

from geosnap.