Author |
Message |
01/04/2007 10:08:14
|
marc
Joined: 08/12/2005 07:39:47
Messages: 4416
Offline
|
For Sweden we are using the following administrative divisions :
ADM1 = counties (21)
ADM2 = municipalities (290)
not yet included in geonames :
ADM3 = parishes
Thanks to Grimar for the file:
Description |
Tab-separated txt-file encoded in UTF-8 |
Download
|
Filesize |
133 Kbytes
|
Downloaded: |
8567 time(s) |
|
|
|
|
04/09/2007 13:49:39
|
PeterBengtson
Joined: 04/09/2007 13:33:58
Messages: 1
Offline
|
Looking at the data for Sweden, it seems that even if there are administrative subdivisions on the ADM1 and ADM2 levels, the locations themselves only are classified according to ADM1 - the location geoname records all have ADM2 empty.
Thus, there are counties and municipalities in an administrative sense, but the actual locations are ordered only according to county.
This makes it very difficult to construct a menu system, as the geonames themselves do not correspond to the administrative subdivisions. For the logic to be completely generalised on a global level (not just for Sweden), we have to look forward, down one level, to see if there are any locations "inside" the administrative division (which is expensive), or predetermine how many levels are possible for each country (which is fugly and nondynamic).
My questions are: (1) will this be rectified for Sweden in particular, and (2) is this the case for other countries as well, in general?
Also, (3) in Germany and also in France, there are many instances of administrative subdivisions containing only one single geoname with the same name as the subdivision. Apart from requiring extra checks to compensate for it, is it correct and intended?
|
|
|
12/09/2007 14:10:41
|
marc
Joined: 08/12/2005 07:39:47
Messages: 4416
Offline
|
Hi Peter
PeterBengtson wrote:
My questions are: (1) will this be rectified for Sweden in particular, and (2) is this the case for other countries as well, in general?
We need the information to be able to add it to the places. As soon as we have got the information we will add it. You can help us find the required information, for example in the form of shape information.
This is the case for most countries.
PeterBengtson wrote:
Also, (3) in Germany and also in France, there are many instances of administrative subdivisions containing only one single geoname with the same name as the subdivision. Apart from requiring extra checks to compensate for it, is it correct and intended?
Toponyms that serve as both as administrative divison and populated places have received two entries to be able to identify them distinctly.
Regards
Marc
|
|
|
|
22/11/2007 22:02:20
|
fairhair
Joined: 22/11/2007 18:34:53
Messages: 4
Offline
|
When you move or create names in parts of Sweden where the ADM1 division has been changed, the names get an obsolete ADM1 association. For instance the ADM1 entity "Västra Götalands län" is a merger of the former "Göteborgs och Bohus län", "Älvsborgs län" and "Skaraborgs län". When I move a name in the former "Göteborgs och Bohus län", the ADM1 association is automatically changed from "Västra Götalands län" to the obsolete "Göteborgs och Bohus län" so I have to change it back manually. Also, also when I create names I have to change ADM1 manually. Its possible the same problem exists in "Skåne län" which is a merger of "Kristianstads län" and "Malmöhus län".
|
|
|
25/11/2007 12:10:58
|
marc
Joined: 08/12/2005 07:39:47
Messages: 4416
Offline
|
Thanks.
I have changed the fips codes for the admin division polygons 01,04,17 to 28 and 11,13 to 27.
Cheers
Marc
|
|
|
|
26/11/2007 12:42:08
|
fairhair
Joined: 22/11/2007 18:34:53
Messages: 4
Offline
|
Unfortunately, the thing is a bit more complicated. At the creation of Västra Götalands län, two ADM2 units were transferred from the former Skaraborgs län to Jönköpings län. Also, at 2007-01-01 one ADM2 unit was transferred from Västmanlands län to Uppsala län so these two polygons have changed. This leads us to the issue of polygons for ADM2 units. Such polygons are freely available as shape files for all the Nordic countries. The resolution is rather low however. This map, which is roughly 100x100 sqkm, will give you an idea:
There is another drawback too. There are no borders at sea or in the big lakes and only larger islands are represented. Thus, objects at sea such as small islands, shoals, lighthouses etc. fall outside the polygons.
|
|
|
02/01/2008 21:06:42
|
Admin
Joined: 19/11/2005 16:33:24
Messages: 46
Offline
|
Hi Fairhair
Do you know which ADM2 municipality was moved?
Cheers
Marc
|
|
|
02/01/2008 21:32:17
|
fairhair
Joined: 22/11/2007 18:34:53
Messages: 4
Offline
|
The two muncipalities moved from Skaraborgs län to Jönköpings län were:
- Habo kommun, Geoname Id 2709492
- Mullsjö kommun, Geoname Id 2690899
The municipality moved from Västmanlands län to Uppsala län was:
- Heby kommun, Geoname Id 2707055
|
|
|
12/01/2010 22:32:13
|
Ingmar Berthelsen
Joined: 12/01/2010 22:06:32
Messages: 1
Offline
|
I suggest that you in some way add the word "län" to the ADM1 division, just to avoid confusion. Some of the ADM1 names are also names for ADM2! (eg Stockholm)
And some ADM1 names are identical to the older division in "landskap" eng "province", (eg Södermanland)
In fact the norwegian service yr.no is using your database, and I'd say that they cannot add "län" by themselves.
You might consider using "county" instead of "län".
Ingmar Berthelsen
|
|
|
09/10/2011 19:58:57
|
blajo
Joined: 07/10/2011 09:49:33
Messages: 4
Offline
|
I'm intressed in getting parishes into geonames. Is it a shape-file with the borders of parishes and municipalities that is needed? In that case I will make a effort to get that information.
Is it also correct to assume that when you have the shape-files all of the places will automatically be placed in the correct parish and municipality.
|
|
|
09/10/2011 22:52:18
|
marc
Joined: 08/12/2005 07:39:47
Messages: 4416
Offline
|
For the parishes and municipalities the centroid lat/lng would be sufficient, but as you correctly say with the boundary shapes we can automatically assign the other features to the right parish and municipality.
Attributes we are interested in: official admin codes (often by the national statistical offices, names in local languages, population.
Best Regards
Marc
|
|
|
|
11/10/2011 06:04:10
|
blajo
Joined: 07/10/2011 09:49:33
Messages: 4
Offline
|
Attached are a csv-file with population data for municipalities of Sweden as of Q1 2011. With which you could do a batch update of the 290 municipalities population.
Description |
Pop Sweden Municipalities Q1 2011 |
Download
|
Filesize |
16 Kbytes
|
Downloaded: |
7556 time(s) |
|
|
|
27/10/2011 17:58:35
|
marc
Joined: 08/12/2005 07:39:47
Messages: 4416
Offline
|
Thanks a lot. The data has been included.
Best Regards
Marc
|
|
|
|
09/12/2015 22:34:05
|
Kitayama
Joined: 03/05/2014 15:22:16
Messages: 3
Location: Sweden
Offline
|
The current structure in Geonames for Sweden is:
ADM1 = counties (21)
ADM2 = municipalities (290)
ADM3 = parishes "församling" (1786)
ADM4 = parishes, old structure (used until 1952) "socken" (2430)
Jan 1, 2016, there will be a new administrative division in Sweden, called "Distrikt" (District). There will be 2.523 districts. The districts are based on the parishes as they were at the end of 1999 - which is almost the same as the old "socken", with some changes made in 1952-1999.
The parishes will no longer be an administrative division used by the Government. However, it will still be used by the Church of Sweden.
The number of parishes has shrunk since the Church took over in 2000,
there are now only 1.378 parishes (so we have about 400 too many in Geonames).
The question now is:
Which administrative divisions shall we maintain in Geonames for Sweden?
Current ADM3 - only used by the Church of Sweden from Jan 1, 2016.
Current ADM4 - not used since 1952
The new entity "Distrikt".
In most cases there will be a 1:1-relationship between ADM4 (Socken) and the new Distrikt. In many cases there's also a 1:1-relationship between
ADM3 (Församling) and ADM4/Distrikt.
The purpose of the new districts is to keep a track back to the old socken system for research purposes. There will be published population data on district level, but they will not be used in the Local Administration.
I have two different suggestions:
1) Remove all data below ADM2-level. That's the lowest level used by Local Administration and the lowest level used by the society. The quality of the current data on ADM3 / ADM4 level is quite poor.
ADM3 exists only for 11.7% of the records.
ADM4 exists only for 9.8% of the records.
ADM2-level is easy to maintain.
2) Somehow merge current ADM3 and ADM4 into ADM3 and rename them to districts.
Before the merge, someone must go through the current structures for ADM3 and ADM4 and match them with the new district stucture.
Additional information:
A district can, just like a parish/socken, belong to more than one ADM2.
The number of districts and their borders are fixed and will not be changed in the future, as the purpose is only to serve researchers.
The districts will get a new numeric code.
|
|
|
14/06/2016 22:54:28
|
jenifferhomes
Joined: 16/05/2014 09:35:17
Messages: 1
Offline
|
This makes it very difficult to construct a menu system, as the geonames themselves do not correspond to the administrative subdivisions. For the logic to be completely generalised on a global level (not just for Sweden), we have to look forward, down one level,
|
|
|
|