GeoNames Forum

This is an interesting find. I have added it to the geonames test cases.
I post the text here as it will soon disappear in the orginal feed.

Title : "Bush nombra al polémico general Michael Hayden como nuevo director de la CIA"
Text : "El presidente de EEUU, George W. Bush, ha elegido al general de la Fuerza Aérea Michael Hayden como segundo responsable de la inteligencia estadounidense, después de John Negroponte. Hayden ucederá a Porter Goss al frente de la CIA. Leer. Escuchar",

I have had a look at how the geonames search engine is dealing with this text.
The words 'CIA', 'EEUU' and 'estadounidense' indicate that the text is about the United States. The word 'Hayden', however, is a placename in the US which makes the search engine think the text is about the place 'Hayden' in the US.

I will play a little bit with the relevance calculation for placenames in geonames, as the place 'Hayden' has only a population of 11,000 and has received too high a relevance.

http://www.geonames.org/search.html?q=hayden&country=US

Changing the relevance Algorithm will certainly improve it but I am not sure whether this will already solve the problem. An other possible improvement would be to use a list of firstnames and have the search algorithm make use of it. 'Michael' happens to be the second most popular firstname in English (after Jacob). The next word after 'Michael' is thus likely to be a family name and not a place name.

As you can see it is a pretty complex task to find an algorithm working for all different kind of texts. If you spot other problems let me know.

http://www.ssa.gov/OACT/babynames/

Hi Joerg

The full text search accepts now one or more parameters fcode.

http://ws.geonames.org/search?q=berlin&country=DE&maxRows=100&fcode=PPLC

Wildcards are also possible : fcode=PPL? will return all fcodes beginning with PPL.

Best Regards,

Marc

Hi xavi

I don't have an ansi file ready, but you can easily generate one for you.

SELECT countrycode,admincode1, ansiname FROM geonames WHERE feature_code='ADM1'

Hope this helps

Marc

Hi xavi

And with
SELECT * FROM geonames_places WHERE feature_code='ADM2' ;
you should get all provinces and with 'ADM3' all 8000+ municipalities ...
As you can see there are only 43 provinces and 1 municipalities in the geonames database. Please feel free to add the missing provinces. (I don't think it makes sense to manually add the 8000 missing municipalities

The file I was referrering to is now available in the download section : http://download.geonames.org/export/dump/admin1Codes.txt
It contains the admin1codes and the name for each code.

For Spain the fips to iso code mapping is 1-1 for most regions and provinces with exception of Ceuta and Melilla. I have update the page with the fips codes for the provinces, and fixed a lot of wrong iso codes on this page :
http://www.geonames.org/ES/administrative-division-spain.html

Marc

Now I understand. I was first thinking you wanted to say placenames were missing in these states.

The four states Massachusetts, Pennsylvania, Kentucky and Virginia are really missing in the gnis dataset.
I have added them and also updated the feature code of the four capitals :

http://www.geonames.org/recent-changes.html

Hi Thomas

How did you check the administratives regions? All states are in the geoname database, so I wonder what makes you think some states are missing.

Marc

You are right 'GB' it the official code. But 'UK' it definitely more used.

The English wikipedia for example contains 33.648 times 'UK' but only 625 times 'GB'.

I was just about to change the code to 'GB' when some one suggest to change LI to FL as FL is more common.
http://forum.geonames.org/gforum/posts/list/60.page

Marc

Hi Nenad

Funny you raise the question just now. In an other thread some one suggested to change the country code for the UK from 'UK' to 'GB' :
http://forum.geonames.org/gforum/posts/list/58.page

I was just about to change it as the main ISO code really is 'GB'. What do you think about UK/GB?
Should we always go with the main ISO code (GB,LI) or with the most common code (UK,FL) ?

Marc

You are right the file really should be part of the download. If you give me your email I can send you the file I am using, though it is not ready to be officially realeased.

Regards,

Marc

Hi xavi

Both UK and GB are valid iso codes for the United Kingdom. I just don't know what is more common. The top level domain for example is UK.
Do you think GB is more common?

Marc

Hi xavi

It is planned to move from fips to isocodes. The challenge with this is that there is no 1<->1 relationship between fips and iso codes, it has to be sorted out country by country. We therefore don't have something like allCountries.zip with iso codes.

Any help with this is welcome ...

Marc

Hi Sameer

The service findNearbyPlaceName takes now an argument "style" with the possible values SHORT,MEDIUM,LONG,FULL (default = MEDIUM).

http://www.geonames.org/export/#findNearbyPlaceName

Regards,

Marc

Hi Dirk

It was a quick fix indeed. Thanks for reporting it.

Marc

Thomas wrote:

I don't know if you have check only the french database but it seems that others countries are concerned by the same problem.

I have restored admin1 code for all countries.
It is just the admin2 update running now that is for france.

Good luck with your project.

Marc

The process updating the population has accidentally removed the admin1 code. I have now restored the admin1 code where it has disapeared.

Thanks for finding this bug and letting us know.

I am now running an admin2 update on the French dataset. In the dump of tomorrow you will find the restored admin1 code and new admin2 codes for many records. The admin2 code here is the iso code for the French Départments.
http://www.insee.fr/fr/nom_def_met/nomenclatures/cog/dbf/2005/txt/depts2005.txt

Let me know if the admin2 code is useful, I will then run the update for other european countries as well.

Posting errors like this one here in the forum or in an email is fine. I will fix them. Thanks a lot.

Marc

Edit :
It is fixed now : http://www.geonames.org/search.html?q=Castilla+La+Mancha&country=

Hallo Eddie

Da habe ich aber Glück gehabt, dass Du deinen Beitrag noch gespeichert hattest. Den Letzten Satz über die Länderdaten habe ich beim erstenmal gar nicht gesehen.

Genau, habe ich gesehen! Da fehlt dann ggf. noch die "Zwischenebene" mit überregionalen Einteilungen.

Richtig, so ist es gedacht. Oder sogar mehrere Spalten wie beim Ländercode, falls mehrere Standards unterschieden werden sollen. (Was Du unten ja auch schreibst.)

Prinzipiell ist das Ziel schon, obige Datei countryInfo.txt ganz im normalen DB-Dump aufgehen zu lassen, oder?

Sicher wird jeder Regionen-Eintrag eine geonameId erhalten. Wenn man sich auf einen Standard einigen kann wird die Haupttabelle einfach um eine Spalte erweitert, wo dann der RegionenCode steht. Fall aber mehrere sich widersprechende Standards untersützt werden sollen muss diese Info zwangsläufig separat gespeichert sein.
Die Datei countryInfo.txt wird sicher länger (oder immer) bestehen da wahrscheinlich noch mehr Länderinfos dazukommen werden wie Telefonvorwahl(en), Währung(en), Landesprachen, Nachbarn, Postleitzahlformat etc. Gut wäre wenn man das CIA Worl Factsbook in einem besseren Format als HTML erhalten könnte.

Das wuerde mich ohnehin interessieren, wie ist das mit den Namen? In der name-Spalte steht immer der Name in Landessprache, oder?

Es steht eher der international gängige Name in der Namesspalte. Bei den Alternativnamen gibt es noch kein Sortierkriterium. Das muss ich noch einbauen. Auf der Datenbank bereits vorhanden aber noch nirgends angezeigt ist die Spalte 'isOfficialName'.

Marc

Anonymous wrote:

<script type="text/javascript" src="jsr_class.js"> </script>
<script type="text/javascript">
// Define the callback function
function getGeo(jsonData) {
alert('...');
}
// The web service call
var req = 'http://ws.geonames.org/findNearbyPostalCodesJSON?lat=45&lng=-123';

You have to add a parameter callback=getGeo to tell our werbservice the name of the function where you are processing the result.
http://ws.geonames.org/findNearbyPostalCodesJSON?lat=45&lng=-123&callback=getGeo

The code was broken. Thanks for letting us know about the problem.

Please try again, it is fixed now and should work.

Marc

Who is 'they'? Is it a company starting with 'Y'? If yes, I don't think 'they' offer reverse geocoding.
But this is no problem as you can use our service for this

Marc