40
Linking your resources to the Data Web Thomas Baker Dublin Core Metadata Initiative Linked Open Data @ AIMS 4 December2012

Linking your resources to the Data Web

Embed Size (px)

DESCRIPTION

Presentation delivered by Tom Baker at Linked Open Data @ AIMS Webinars. http://aims.fao.org/linked-open-data-webinars-at-aims http://aims.fao.org/linked-open-data-webinars-at-aims/tom-baker Abstract ------------- Linking data into the Semantic Web means more than just making data available on a Web server. It means using Web addresses (URIs) in data as names for things; tagging resources using those URIs - for example, URIs for agricultural topics from AGROVOC; and using URIs to point to related resources. This talk walks through a simple example to show how linking works in practice, illustrating RDF technology with animated graphics. It concludes with a recipe for linking your data: Decide what bits of your data are most important, such as Subject, Author, and Publisher. Use URIs in your data, whenever possible, such as Subject terms from AGROVOC. Then publish your data in RDF on the Web where others can link to it. Simple solutions can be enough to yield good results.

Citation preview

Page 1: Linking your resources to the Data Web

Linking your resourcesto the Data Web

Thomas BakerDublin Core Metadata Initiative

Linked Open Data @ AIMS4 December2012

Page 2: Linking your resources to the Data Web

2

Linking Open Datacloud diagramby Richard Cyganiak and Anja Jentzschhttp://lod-cloud.net/

Data Web – the Cloud…

Page 3: Linking your resources to the Data Web

Data Web – zooming in…

3

AGRIS record

AGROVOC concept

DBPediaconcept

VIVO Expert

BBC Web page

BBC News

Wikipediapage

Page 4: Linking your resources to the Data Web

How to link data into the “Semantic” Web

1. Publish your data on the Web.

2. Use Web addresses (URIs) as names for things.

3. Tag your resources with URIs (Agrovoc!).

4. Pointing to related resources with URIs.

See: http://aims.fao.org/sites/default/files/CIARD-2011-06-10.pdf http://www.w3.org/DesignIssues/LinkedData.html

Page 5: Linking your resources to the Data Web

Describe things in a standard, reusable way

Simple “statements” using bits of data to describe things.

5

ResourcerelatedTo

ResourceA ResourceB

ResourcedescribedBy

ResourceA Text data

Page 6: Linking your resources to the Data Web

RDF – the grammar for data statements

These are called “triples”.

6

ResourcetranslatedBy

http://bookA http://personB

ResourcehasName

http://personB “Alan Oh”

Page 7: Linking your resources to the Data Web

EXAMPLE: “LINKING” TWO RESOURCES

7

Page 8: Linking your resources to the Data Web

Resource A: Article indexed in AGRIS database

8

Topic of the articleIs “acrididae”, i.e.“grasshoppers”.

One author is named“Lu Hui”.

Page 9: Linking your resources to the Data Web

Statements about Resource A as “pseudo-triples”

Subject Predicate Object

Resource A has the Title 典型草原三种蝗虫种群死亡率的研究

Resource A has an Author Lu Hui

Resource A has a Topic Acrididae (grasshoppers)

9

Page 10: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 has the Title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author Lu Hui

agris:CN2009002389 has a Topic Acrididae (grasshoppers)

10

Resource A is identified by the URI http://agris.fao.org/resource/CN2009002389.Let’s abbreviate this as agris:CN2009002389.

Page 11: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 has the Title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author agris-author:luhui

agris:CN2009002389 has a Topic Acrididae (grasshoppers)

11

The author is identified by the URI http://agris.fao.org/author/luhui.Let’s abbreviate this as agris-author:luhui.

Page 12: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 has the Title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author agris-author:luhui

agris-author:luhui has the Name Lu Hui

agris:CN2009002389 has a Topic Acrididae (grasshoppers)

12

The author agris-author:luhui has the name Lu Hui.

Page 13: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 has the Title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author agris-author:luhui

agris-author:luhui has the Name Lu Hui

agris:CN2009002389 has a Topic agrovoc:c_4416

13

The topic is identified by the URI http://aims.fao.org/aos/agrovoc/c_4416.Let’s abbreviate this as agrovoc:c_4416.

Page 14: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 has the Title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author agris-author:luhui

agris-author:luhui has the Name Lu Hui

agris:CN2009002389 has a Topic agrovoc:c_4416

agrovoc:c_4416 has preferred label Acrididae (en)

agrovoc:c_4416 has preferred label 蝗科 (zh)

14

The description of http://aims.fao.org/aos/agrovoc/c_4416 in the AGROVOCConcept Scheme tells us how this concept is labeled in English and in Chinese.

Page 15: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 has an Author agris-author:luhui

agris-author:luhui has the Name Lu Hui

agris:CN2009002389 has a Topic agrovoc:c_4416

agrovoc:c_4416 has preferred label Acrididae (en)

agrovoc:c_4416 has preferred label 蝗科 (zh)

15

The Dublin Core element “Title” is identified by the URI http://purl.org/dc/terms/title.Let’s abbreviate this as dct:title.

Page 16: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416

agrovoc:c_4416 has preferred label Acrididae (en)

agrovoc:c_4416 has preferred label 蝗科 (zh)

16

Ditto for author (i.e., creator) and topic (i.e. subject).Name is defined in the FOAF vocabulary.

Page 17: Linking your resources to the Data Web

Triplifying the description…

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416

agrovoc:c_4416 skos:prefLabel Acrididae (en)

agrovoc:c_4416 skos:prefLabel 蝗科 (zh)

17

A property for the preferred label of a concept has been definedIn the Simple Knowledge Organization System (SKOS) vocabulary.

Page 18: Linking your resources to the Data Web

The triples seen as a graph

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416

18

agris:CN…389dct:title 典型草…

Lu Hui

agris-author:luhui

dct:subject

agrovoc:c_4416

foaf:name

dct:creator

Page 19: Linking your resources to the Data Web

Resource B: AGRIS article in German

19

Page 20: Linking your resources to the Data Web

Triples describing Resource B seen as a graph

Subject Predicate Object

agris:CH2001000179 dct:title Heuschrecken brauchen ökologische Ausgleichsflächen

agris:CH2001000179 dct:creator agris-author:peterb

agris-author:luhui foaf:name Peter, B.

agris:CH2001000179 dct:subject agrovoc:c_4416

20

agris:CH…179dct:title Heuschrecken

Peter, B.

agris-author:peterb

dct:subject

agrovoc:c_4416

foaf:name

dct:creator

Page 21: Linking your resources to the Data Web

Merging triples for Resource A…

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416

21

Page 22: Linking your resources to the Data Web

…with triples for Resource B…

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416

agris:CH2001000179 dct:title Heuschrecken brauchen ökologische Ausgleichsflächen

agris:CH2001000179 dct:creator agris-author:peterb

agris-author:luhui foaf:name Peter, B.

agris:CH2001000179 dct:subject agrovoc:c_4416

22

Page 23: Linking your resources to the Data Web

The computer detects matching URIs...

Subject Predicate Object

agris:CN2009002389 dct:title 典型草原三种蝗虫种群死亡率的研究

agris:CN2009002389 dct:creator agris-author:luhui

agris-author:luhui foaf:name Lu Hui

agris:CN2009002389 dct:subject agrovoc:c_4416agris:CH2001000179 dct:title Heuschrecken brauchen

ökologische Ausgleichsflächen

agris:CH2001000179 dct:creator agris-author:peterb

agris-author:luhui foaf:name Peter, B.

agris:CH2001000179 dct:subject agrovoc:c_4416

23

Page 24: Linking your resources to the Data Web

The computer detects matching URIs...

24

agris:CN…389dct:title

典型草…

Lu Hui

agris-author:luhui

dct:subjectagrovoc:c_4416

foaf:name

dct:creator

agris:CH…179dct:title Heuschrecken

Peter, B.

agris-author:peterb

dct:subject

agrovoc:c_4416foaf:name

dct:creator

Page 25: Linking your resources to the Data Web

…and merges (“links”) the graphs into one.

25

agris:CN…389dct:title

典型草…

Lu Hui

agris-author:luhui

dct:subject

foaf:name

dct:creator

agris:CH…179dct:title Heuschrecken

Peter, B.

agris-author:peterb

dct:subject

agrovoc:c_4416foaf:name

dct:creator

Resources in differentlanguages, linked by reference to a commonKOS Concept.

An “infinitely expandabledescription”…!

Page 26: Linking your resources to the Data Web

CASTING A WIDER NET

26

Page 27: Linking your resources to the Data Web

http://www.bbc.co.uk/nature/life/Orthoptera

27

Page 28: Linking your resources to the Data Web

http://www.bbc.co.uk/nature/life/Orthoptera.rdf

28

Page 29: Linking your resources to the Data Web

The BBC Grasshopper page links to a news article…

29

Page 30: Linking your resources to the Data Web

http://news.bbc.co.uk/earth/hi/earth_news/newsid_8172000/8172168.stm

30

Page 31: Linking your resources to the Data Web

…and to a DBPedia category…

31

Page 32: Linking your resources to the Data Web

http://dbpedia.org/page/Orthoptera

32

Page 33: Linking your resources to the Data Web

…using information extracted from Wikipedia.

33

Page 34: Linking your resources to the Data Web

34

What else links to the DBPedia concept…?

Page 35: Linking your resources to the Data Web

35

…the description of an expert!

Page 36: Linking your resources to the Data Web

What else links to the DBPedia concept?

36

“exactly matches”

“broader than”

Page 37: Linking your resources to the Data Web

“Six degrees of (linked-data) separation”…

37

AGRIS record

AGROVOC concept

DBPediaconcept

VIVO Expert

BBC Web page

BBC News

Wikipediapage

Page 38: Linking your resources to the Data Web

From one record, many links

38

AGRIS record

AGROVOC

AGRIS Authors Dublin Core

FOAF

SKOS

Page 39: Linking your resources to the Data Web

Recipe for Linking your Data

1. Decide what bits of your data are most important (eg, Subject, Author, Publisher…).

2. Whenever possible, use URIs (eg, Agrovoc terms) to name those things.

3. Publish your data in RDF on the Web, where others can link to it.

4. Keep it simple. Simple solutions can be enough to yield good results.