Missing fips codes in admin 1

[gka]

It seems that with the new 2.0.0 release FIPS codes were dropped in Admin 1 boundaries. Is that intentional? Actually these codes are pretty helpful b/c they’re still widely used in other databases.

[Nathaniel]

That standard has been retired and it’s quite out of date with much of the rest of the world beyond the USA. Do you still use it? If so, why and what countries? Is there a better local coding system to use there instead?

The FIPS codes were half moved to the code_local field for the USA since they are still current via the US Census codes. Looks like they shouldn’t have the US prefix. Now for instance California is US06 where the 06 is the fips code. For Maryland this is US24.

I notice a separate issue where MD and VA aren’t dividing right on the DelMarVa peninsula, oops!

It should be possible to add them back in, but they seemed out of date and a hassle to make sure they were all in fact coded right and then some would always be missing since many adm-1 splits have occurred since the standard was withdrawn.

[Nathaniel]

I should mention also the ISO_3166_2 field is where the new ISO codes should go. But they’re often missing now.

The HASC codes are mostly well maintained by someone else and are often in the data here in Natural Earth.

I have it planned to do more work on the adm-1 later this month and in January so please let me know what you’d like to see here.

[gka]

I need FIPS codes because they’re still used in MaxMind’s GeoIP database. This worked perfectly with Natural Earth.

Since the standard has been retired I agree that it makes sense to drop it. On the other hand, including legacy identifiers for backward compatibility seems to be worth the few kB.

Do you have a change-set of the regions that changed from 1.4.0 to 2.0.0 (e.g. New Zealand)? I need to go through them and do the mapping myself. I can probably use the FIPS codes from 1.4 and add the missing ones.

[Nathaniel]

Agree they should come back for legacy support reasons…

The 1.4 version is still available here:

https://github.com/nvkelso/natural-earth-vector/tree/master/archive

There are just a couple changes on the North Island in New Zealand (splits, adds). There will be a few more refinements there for version 2.1 but only affecting boundary alignments, not new features.

[gka]

Ok, that’s good to hear. Thanks for your support.

[Nathaniel]

Looks like FIPS 10-4 has been resurrected now as NGA Geopolitical Codes!?

http://earth-info.nga.mil/gns/html/gazetteers2.html

[WrinkledCheese]

I would also like to see FIPS10-4 brought back.

Technically US06 is the correct FIPS 10-4 code for California, but most datasets I’ve seen just drop off the prefix as it always matches the FIPS country code.

MaxMind’s GeoLite GeoIP CSVs use FIPS for all countries EXCEPT they use ISO 3166-1 for U.S. and Canada…for some reason.

For the interim, while I wait for FIPS to be brought back, I will use the old polygon data to make progress until the newer data sets have brough FIPS back.

The folks at MaxMind say it’s being looked into whether or not they’re going to add ISO 3166-2 codes to the GeoIP databases, but they don’t have a plan in place yet.

[WrinkledCheese]

Is there somewhere else I can easily get the 1.4 data?

https://github.com/blog/1302-goodbye-uploads

I ran a script and found that of all the codes, this is what seems to be the current state:

iso_a2 is consistent 3632 / 3632 present
adm1_code is consistent 3632 / 3632 present
mapcolor13 is reliable 3 / 3632 missing
code_hasc is inconsistent 249 / 3632 missing
iso_3166_2 is reliable 1 / 3632 missing
code_local is inconsistent 3581 / 3632 missing
hasc_maybe is inconsistent 3470 / 3632 missing
region_cod is inconsistent 3186 / 3632 missing
big_code is inconsistent 3428 / 3632 missing
gadm_level is inconsistent 111 / 3632 missing

What does adm1_code translate to?

When reading the above script output, keep in mind that the script doesn’t account for the validity of the data, just that there is data present. I know for a fact that ISO_3166_2 has a lot of dirty data and so does adm1_code, albeit to a lesser extent.

[WrinkledCheese]

Is it my understanding that Natural Earth Data v2.1 will have FIPS_10 codes brought back?

[Author]

Posts