W3C home > Mailing lists > Public > public-lod@w3.org > May 2009

Re: How to query for Country Specific Data

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Sun, 24 May 2009 13:10:35 -0400
Message-ID: <4A197F8B.5010100@openlinksw.com>
To: richard.hancock@3kbo.com
CC: public-lod@w3.org
richard.hancock@3kbo.com wrote:
> Hi all,
>
> I am wondering what are the best queries to run against the various LOD
> datasets in order to get an idea of what Linked Data is currently
> available for a specific country, e.g. New Zealand.
>
> My plan was to start with Dbpedia and move out from there but so far my
> Dbpedia queries so far haven't yielded the results I was hoping for.
>
> My initial reasoning was that I could use the Dbpedia country URI ( i.e. 
> <http://dbpedia.org/resource/New_Zealand> ) as the beginning point for my
> search and get a list of properties relating things to New Zealand using
>
> SELECT DISTINCT ?p
> FROM NAMED <http://dbpedia.org>
> WHERE {
> ?s ?p <http://dbpedia.org/resource/New_Zealand> .
> }
>
> and a list of New Zealand properties using
>
> SELECT DISTINCT ?p
> FROM NAMED <http://dbpedia.org>
> WHERE {
> <http://dbpedia.org/resource/New_Zealand> ?p ?o  .
> }
>
> The first thing I looked for were properties that would let me find towns
> and cities in New Zealand.
>
> When I couldn't find obvious properties relating New Zealand towns and
> cities to New Zealand I tried taking the URI of a specific city,
> Wellington to see what properties linked it to New Zealand.
>
> The query below shows that no property linked Wellington to New Zealand.
>
> SELECT *
> FROM NAMED <http://dbpedia.org>
> WHERE {
>  <http://dbpedia.org/resource/Wellington> ?p
> <http://dbpedia.org/resource/New_Zealand> .
> }
>
> gives no results.
>
> Going the other way showed that Wellington was the capital of New Zealand
> but I was also expecting to see another property indicating that
> Wellington was a city of New Zealand. I had hoped to find the second
> property so that I could use it to find all other cities and towns within
> New Zealand.
>
> SELECT *
> FROM NAMED <http://dbpedia.org>
> WHERE {
>  <http://dbpedia.org/resource/New_Zealand> ?p
> <http://dbpedia.org/resource/Wellington>.
> }
>
> indicates that New_Zealand has a capital via
> <http://dbpedia.org/property/capital> of Wellington
>
> The best indication that Wellington is located in New Zealand was found
> using the location property.
>
> SELECT *
> FROM NAMED <http://dbpedia.org>
> WHERE {
> <http://dbpedia.org/resource/Wellington>
> <http://dbpedia.org/property/location> ?o .
> }
>
> gives the literal "Wellington, New Zealand"@en.
>
> A scenario similar to the one between Wellington and New Zealand exists
> between Berlin and Germany but here the literal value of the location
> property proves less useful
>
> SELECT ?location
> FROM NAMED <http://dbpedia.org>
> WHERE {
> <http://dbpedia.org/resource/Berlin>
> <http://dbpedia.org/property/location> ?location .
> }
>
> gives the literal "Berlin"@en
>
> At this point I thought it was time to ask for help.
>
> For the current Linked Data sets what is the best approach to take when
> seeking to find what information has already been published for a specific
> country ?
>
> Depending on what you are looking for are some datasets better starting
> points than others? E.g. for finding the towns and cities is geonames
> (http://sws.geonames.org/) ,CIA Factbook or some other dataset, a better
> starting point?
>
> For Dbpedia are there plans to migrate properties like location
> <http://dbpedia.org/property/location> which currently use literals to use
> URIs instead?
>
> Cheers,
>
> Richard Hancock
>
>
>
>
>
>
>
>
>
>
>
>   
Richard,

One place you can try is: http://lod.openlinksw.com (SPARQL via: 
http://lod.openlinksw.com/sparql). If you want to get to specific data 
sets also look at: http://lod.openlinksw.com/void/Dataset .

Hope it helps you.

-- 


Regards,

Kingsley Idehen	      Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO 
OpenLink Software     Web: http://www.openlinksw.com
Received on Sunday, 24 May 2009 17:11:13 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:21 UTC