16
UAd Institute of Information Systems & Information Management Building Linked Data For Both Humans and Machines Wolfgang Halb, Yves Raimond and Michael Hausenblas LDOW2008, 2008-04-22, Beijing, China

LDOW08 at presentation: Building Linked Data for Both Humans and Machines

  • Upload
    whalb

  • View
    1.657

  • Download
    3

Embed Size (px)

DESCRIPTION

Presentation at the Linked Data on the Web (LDOW2008) workshop at the WWW2008 Conference

Citation preview

Page 1: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

UAd

Institute of Information Systems & Information Management

Building Linked Data For Both Humans and Machines

Wolfgang Halb, Yves Raimond and Michael Hausenblas

LDOW2008, 2008-04-22, Beijing, China

Page 2: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

2

UAd

Agenda rieseRDFizing and Interlinking the Eurostat data Set Effort

LinkingOpenData Eurostat Architecture Schema & Data Querying Interlinking Demo Inside

Page 3: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

3

UAd

riese

RDFizing and Interlinking the Eurostat data Set Efforthttp://riese.joanneum.at

Contributing to Linking Open Data project Offer Semantic Web version of Eurostat data ... for both humans and machines

Page 4: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

4

UAd

LinkingOpenData: Current State

Page 5: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

5

UAd

Eurostat Eurostat (http://ec.europa.eu/eurostat) publishes statistics in these themes:

General and regional statistics Economy and finance Population and social conditions Industry, trade and services Agriculture and fisheries External trade Transport Environment and energy Science and technology

about the European Union in detail and additional statistics for some major non-European countries

Page 6: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

6

UAd

Eurostat data dump provided as download (tables) updated twice a day additionally needed:

dictionary files to translate the data codes used table of contents for structure

Size of Eurostat data 5 GB data dump in approx. 4,000 files 350 million data values 80,000 different data codes

Page 7: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

7

UAd

riese: architecture

Page 8: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

8

UAd

riese: schema & data

riese:Item

xsd:String / xsd:Decimal

rdf:valueevent:Event

rdfs:subClassOf

riese:Dimension riese:dimension

xsd:String

dc:title

dimension:Geo

dimension:xxx

geonames:Feature

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

dimension:Flags

riese:flagrdfs:subClassOf

riese:Dataset

riese:dataset

xsd:String

dc:title

dimension:Time

rdfs:subClassOf

skos:Concept

rdf:type

skos:narrower /skos:broader

event:time

event:place

xsd:String

dc:title

geonames:parentFeature

Additional features for geo not detailed here

riese:datasetOf

Page 9: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

9

UAd

riese: schema & data Survey Modelling Statistics in RDF, Lee Feigenbaum

“The riese approach seems the best combination of flexibility and usability.” [1]

Generalised framework for statistics modelling to be released soon

Example data:<riese:Dataset

rdf:about="http://riese.joanneum.at/data/eb040"

dc:title="Inflation rate"

riese:data_end="2006"

riese:data_start="1980"

riese:last_update="08/01/2008“/>

[1] http://www.thefigtrees.net/lee/blog/2008/03/modeling_statistics_in_rdf_a_s.html

Page 10: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

10

UAd

riese: schema & data<riese:Item dc:title=“Inflation rate Austria 2006"

rdf:value=“1.7"

<riese:dimension rdf:resource="http://riese.joanneum.at/dimension/geo/at"/>

<riese:dimension rdf:resource="http://riese.joanneum.at/dimension/time/2006"/>

<riese:dataset rdf:resource="http://riese.joanneum.at/data/eb040"/>

</riese:Item>

Page 11: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

11

UAd

riese: schema & data XHTML + RDFa example:

<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN"

"http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd"><html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" xmlns:riese="http://riese.joanneum.at/schema/core#" ... ><head>...</head><body about="http://riese.joanneum.at/data/economy/"

instanceof="riese:Dataset"><span class="toc-entry"><a

href="http://riese.joanneum.at/data/bop/" rel="skos:narrower" class="dim">Balance of payments - International transactions</a></span>

Last update: <span property="dc:date" datatype="xsd:date">2008-01-09</span>

</body></html>

Page 12: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

12

UAd

riese: querying

Page 13: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

13

UAd

riese : interlinking Generic approach

Restrict source dataset to potential interlinking candidates Look up identifying feature from source in target dataset Restrict results by appropriate classifications or identifiers Create the interlink

User Contributed Interlinking

Page 14: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

14

UAd

riese: demo

Page 15: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

15

UAd

riese: inside Server:

Apache 2.2 SWI-Prolog p2r PHP 5 RDF/XML documents in the file system

(to be replaced by p2r data-pump)

Client XHTML+RDFa Javascript/Yahoo! User Interface Library [YUI]

Page 16: LDOW08 at  presentation: Building Linked Data for Both Humans and Machines

16

UAd

Let‘s discuss!