| Author |
Message |
|
|
Hi Guido
Geocoding webpages has a lot of new aspects compared to geocoding of natural language text. Location information is often encoded in tabular address formats. You could first look at how addresses can be identified on a webpage. (Microformats) A next problem is the visual proximity of address elements. For natural language geocoding you have one long string of text, whereas on a webpage you have ofen many elements between address element. The geocoder will have to detect this information.
We are not using any third party library for natural language geocoding. In a first step we identify the part of speech of all elements, then we search matching placenames and in the third step we find the best places from the potential places depending on proximity of other places in the same text and some gramatical rules.
Regards
Marc
|
 |
|
|
Hi Alexey
The only restaurant database I know of is the chefmoz project :
http://chefmoz.org/
It is part of the dmoz project.
Last time I have checked they did not have any lat/lng information.
Regards
Marc
|
 |
|
|
Hi
Administrative divisions for Poland are not uptodate. I think we can delete (or change to ADMD) all administrative divisions that do not have 'Voivodeship ' in their names.
See also this other thread about the topic :
http://forum.geonames.org/gforum/posts/list/604.page
Marc
|
 |
|
|
Hi Fabrizio
I think we can delete (or change to ADMD) all administrative divisions that do not have 'Voivodeship ' in their names.
The same issue is also discussed in this thread:
http://forum.geonames.org/gforum/posts/list/602.page
Regards
Marc
|
 |
|
|
The ascii names are included in the alternatenames column but have been forgotten to be included in the alternatenames.txt dump
Marc
|
 |
|
|
Hi
Second order administrative divisions for Australia have not yet been fully integrated into GeoNames. If you just care about one or two divisions you can add them with the google maps interface. In order for them to show off on the drop down list and work properly we will need a code for the administrative division. Normally we use the official code used by local authorities.
I have some official data files for Vic, SA and NSW, but I did not yet have the time to integrate them in GeoNames. The main work is tolook at the featurecodes in these datasets and map them to the GeoNames codes before we can insert the data. We could need help with this. Would you like to help us with Australian data?
Regards
Marc
|
 |
|
|
You are right the column asciiname is not exported in the dump file. I completely forgot about this. I think we cannot just add the column like this without forewarning since a couple of GeoNames users have daily import scripts to import the daily dump.
Regards
Marc
|
 |
|
|
Hi
'Duesserdorf' is in the file alternatenames.txt in the column asciiname for all alternatenames 'Düsseldorf'.
Cheers
Marc
|
 |
|
|
Hi
Yes, this was a bug. Thanks for reporting it.
Cheers
Marc
|
 |
|
|
I have added the missing 'e'.
To edit alternate names you have to open the info window of the main name and click on 'alternate names' : http://www.geonames.org/manual.html#alternateNames
Cheers
Marc
|
 |
|
|
Hi Xav
I have fixed the admin1 code for Essonne. Bye the way, you can also fix errors like this one. Go to the info window on top of googlemaps for this toponym, click on 'edit', click on the admin level you want to edit, choose the right admin division, click 'save'.
'Brittany' is the English name for 'Bretagne'. In the adminCodes file we include the English name. If you want the french name you have to use the file alternateNames.txt and get the name with the language code 'fr'.
Cheers
Marc
|
 |
|
|
Hi xav
Chefchaouen belongs to Tanger-Tétouan (fips 57). I have fixed this.
Aousserd and Oued-Ed-Dahab are in Western Sahara. Here we have a political problem. Some people say Western Sahara is part of Morocco while others (Polisario) say it is an independent country. What shall GeoNames do?
Western Sahara has its own ISO country code EH. So I think, we should make these two regions part of the country Wester Sahara and not Morocco.
Any thoughts about this?
Regards
Marc
|
 |
|
|
Hi Fred
Only the parts of composite names were previously indexed in the search index. As of now the parts as well as the full composite names are indexed and fuzzy search on composite names are now supported.
Regards
Marc
|
 |
|
|
Hi Alexey
It looks like an encoding problem. The GeoNames dump files are encoded in UTF8. You should use the same encoding.
There is also a file alternatenames.txt in the dump directory. It gives you the name of the place with the ISO language code. For Limoges it will be Лимож with the languge code 'ru'.
Cheers
Marc
|
 |
|
|
Hallo Michi
Ich glaube du könntest recht haben. Da stehen nirgends irgendwelche Einschränkungen für diese Daten. Würdest du helfen die Zuordnung fips-iso in Ordnung zu bringen?
Wir haben noch nicht auf iso gewechselt, da wir so viel wie möglich auf einen Schlag wechseln wollen. Wenn wir jede Woche ein Land wechseln würden gäbe das ein ziemliches Durcheinander und einige Anwendungen, die auf GeoNames beruhen würden Probleme kriegen.
Gruss
Marc
|
 |
|
|
thanks missokhay.
missokhay wrote:
can you tell me about the jData parameter how the function getLocation got it and what is the callback?
the jData is the result GeoNames is returning. When you add the 'callback' parameter to the web service call GeoNames will return the result in a way that the function you pass as 'callback' will automatically be called and you can process the data in your own function. Cool, isn't it?
Cheers
Marc
|
 |
|
|
Denkst du dass, alle x-hundertausend PLZs alphabetisch sortiert werden sollen? Mir scheint das nicht sehr sinnvoll. Wenn es sehr viele sind, ist es doch sinnvoller dies nach Wichtigkeit oder Relevanz zu sortieren. Alphabetische Sortierung kann dann zum Zuge kommen wenn nur noch wenige übrig bleiben.
Gruss
Marc
|
 |
|
|
Es ist leider richtig, dass sich bei der Namensgebung einige Inkonsistenzen eingeschlichen haben. Im Nachhinein ist es unglücklicherweise nicht mehr ganz leicht dies zu ändern.
Für einige Länder haben wir bereits begonnen, die PLZ in den GeoNames Datensatz zu integrieren. Dies erfolgt als 'alternate name' mit der Sprache 'post'.
Marc
|
 |
|
|
Hallo Michael
Wenn es nur um die 200 CHF ginge hätte ich die Daten schon lange gekauft. Leider würden es die Lizenzbestimmungen kaum erlauben die Daten mirnichtsdirnichts bei GeoNames zum Download anzubieten. Das heisst wir müssen die Daten etwas mühsamer zusammentragen. Häufig beruht der ISO code auf den Codes die im jeweiligen Land verwendet werden, häufig vom Amt für Statistik (wie ZH für Zürich). Im Zuge dieser Arbeiten an den administrativen Landes Unterteilungen aktualisieren wir jeweils auch die ISO codes. Eine weitere Quelle ist natürlich auch Wikipedia.
Gruss
Marc
|
 |
|
|
Hallo Michael
Das ist ein cooles Tool. Hoffentlich verwenden es auch viele Leute und ich hoffe es mag auch mithelfen die unsägliche Templateschwemme bei Wikipedia ein bisschen einzuschränken.
Bei mir scheint es für Zürich zu funktionieren, nur mit dem Umlaut gibt es noch ein Problem. Verwendest du UTF8?
Gruss
Marc
|
 |
|
|