adm1 codes (missing FIPS, partial HASC, need ISO)

[nvkelso]

See this thread in the forums: http://www.naturalearthdata.com/forums/topic.php?id=471&replies=3#post-1048

[nvkelso]

The 1.x series had FIPS codes. Those weren't updated for a long time and were out of date so they were dropped from 2.x series.

But then Geonames.mil GNS picked it back up and it's updated more often than ISO (every couple months versus 1ce a year) so time to add FIPS codes back in.

[bblack]

I'd also like to see the ISO codes. Looks like they exist for some countries' admin1 regions (e.g. US), but not others (e.g. AU).

[nvkelso]

3.0.0-pre is up on quattroshapes.org (and has been for a while, I need to merge it here, sorry).

Should have a good bit of what you need. http://static.quattroshapes.com/ne_adm1.zip

If you can add the rest that'd be huge!

On Tue, Sep 17, 2013 at 9:49 AM, Brian Black notifications@github.comwrote:

I'd also like to see the ISO codes. Looks like they exist for some countries' admin1 regions (e.g. US), but not others (e.g. AU).

— Reply to this email directly or view it on GitHubhttps://github.com/nvkelso/natural-earth-vector/issues/10#issuecomment-24603782 .

[karimbahgat]

Looks like the problem with missing codes is still an issue, which is undestandable given how huge a task it is. But I have a proposal. I need accurate codes for all provinces of the world and will be going through the statoids.com pages for every country as part of another project anyway, so I thought I might as well update the natural earth adm1 codes incl the admin "type" variable. Before I start though, I wanted to check if it is okay that I overwrite all Natural Earth adm1 codes to the codes listed on statoids.com, which in my view is likely to be very accurate given the amount of efforts the author of that site has put in? Specifically I will update HASC ("code_hasc"), ISO1 ("iso_31662"), and FIPS ("fips") -- are these the correct variables I should be updating. And what specifically are the "gns" versions of these variables ( #121 )?

I will probably make a new Github repo which I can link to, and that way track my changes, and make it easier/more manageable to let others contribute codes for individual countries if they want to speed it up.

Similar to #105, in order to better see the changes, I thought I could convert the original adm1 shapefile to geojson and then make my changes to the same geojson file. Then after you skim the changes if you decide to keep it you can just convert the geojson to shapefile?

One last thing, when I use Python to try to load and then save the dbf file using "utf8" I get encoding error, but not if I use "latin". So I wonder if the file encoding is "latin" instead of "utf8" as claimed, relevant for #89 .

[nvkelso]

Does this file have more of what you need? http://static.quattroshapes.com/ne_adm1.zip

Would be great if you could add more. Please retest your Python workflow against it (should certainly be UTF8 encoding).

Yeah, GeoJSON is much easier to do a diff with. Suggest each property is on it's own line to make this even easier. Just make a repo with the GeoJSON file before and after and I'll take care of merging it back.

_n

On Wed, Oct 7, 2015 at 3:01 AM, Karim Bahgat notifications@github.com wrote:

Looks like the problem with missing codes is still an issue, which is undestandable given how huge a task it is. But I have a proposal. I need accurate codes for all provinces of the world and will be going through the statoids.com pages for every country as part of another project anyway, so I thought I might as well update the natural earth adm1 codes. Before I start though, I wanted to check if it is okay that I overwrite all Natural Earth adm1 codes to the codes listed on statoids.com, which in my view is likely to be very accurate given the amount of efforts the author of that site has put in? Specifically I will update HASC ("code_hasc"), ISO1 ("iso_31662"), and FIPS ("fips") -- are these the correct variables I should be updating. And what specifically are the "gns" versions of these variables ( #121 https://github.com/nvkelso/natural-earth-vector/issues/121 )?

I will probably make a new Github repo which I can link to, and that way track my changes, and make it easier/more manageable to let others contribute codes for individual countries if they want to speed it up.

Similar to #105 https://github.com/nvkelso/natural-earth-vector/issues/105, in order to better see the changes, I thought I could convert the original adm1 shapefile to geojson and then make my changes to the same geojson file. Then after you skim the changes if you decide to keep it you can just convert the geojson to shapefile?

One last thing, when I use Python to try to load and then save the dbf file using "utf8" I get encoding error, but not if I use "latin". So I wonder if the file encoding is "latin" instead of "utf8" as claimed, relevant for #89 https://github.com/nvkelso/natural-earth-vector/issues/89 .

— Reply to this email directly or view it on GitHub https://github.com/nvkelso/natural-earth-vector/issues/10#issuecomment-146137252 .

[karimbahgat]

Hi,

Apologies for the belated response. No, from a quick scan through the quattroshapes version missing codes still seems pervasive for iso codes and occasional for hasc codes, I don't see any major differences in the codes. Since I will be going through all the statoid pages anyway I might just as well update everything.

But should I use the github version or the quattroshapes version? Which is the newer more updated one in general / did you ever get around to moving v3 over to github?

Thanks

[sventech]

Looks like this has been addressed here: https://github.com/nvkelso/natural-earth-vector/pull/128

[nvkelso]

Agree, closing. More improvements in PR https://github.com/nvkelso/natural-earth-vector/pull/199 are in the works.

[nvkelso]

~~