Intro to Linked Data: Data Modeling

An
Introduction
to
Linked Data

Part 2 of 5 David Hyland-Wood
Data Modeling University of Mary Washington

Monday, May 16, 2011

Data modeling
without context

Modeling without Context


• Model the Real World


• Look for objects of interest


• Put aside the immediate needs of any
application


application
• Investigate how others are already
modeling similar or related data


application
• Investigate how others are already
modeling similar or related data
• These techniques allow for re-use and
merging


• Look for duplication (normalise)


• Think about time and events, as well as
things


things
• Consider how the data will change over
time


things
• Consider how the data will change over
time
• Separate real-world (non-information
resources) and online (information
resources)


Objects of Interest

Apollo 11 Command and crew
Buzz Aldrin
Service Module


Objects of Interest

Buzz Buzz Aldrin
Aldrin
Service Module


Objects of Interest
20 January 1930

Person

born

a

Buzz Buzz Aldrin
Aldrin
Service Module


Objects of Interest
20 January 1930

Person

born

a

Buzz Aldrin
Service Module

name

Buzz Aldrin


Apollo 11 Command and launchsite
Cape Canaveral
Service Module

name name
Apollo 11 Command and Service Module Cape Canaveral, United States

launched
16 July 1969

mass
28801.0

a
Spacecraft


URIs:
Naming Things
on the Web

Naming Things on the Web
• Use http:// URIs
There are other URIs, but http URIs can be used to get more
information quickly and easily

• Use clean, stable URIs
Abstract away implementation details
Ensure they are reliable

• There are two types of resource, both
require URIs
Information Resources — Web pages, images, CSS ﬁles etc.
Non-Information Resources — People, places, events, concepts etc.


Distinguishing Resources
http://example.com/1969-059A ← Thing

http://example.com/1969-059A.rdf ← RDF data

http://example.com/1969-059A.html ← HTML page

http://example.com/1969-059A ← Thing

http://example.com/1969-059A/rdf ← RDF data

http://example.com/1969-059A/html ← HTML page

http://example.com/resource/1969-059A ← Thing

http://example.com/data/1969-059A ← RDF data

http://example.com/page/1969-059A ← HTML page


Use URIs

http://.../topics/physics
URIs provide names for resources
You can say things about resources

"Physics"
Literal values are displayable labels
You can't say more things about "Physics"


Use Your Own Domain

http://yours.com/spacecraft/1959-059A
I own the domain, so I can control the meaning of the URI
I can also commit to its permanence and provide data at this
address

http://theirs.com/spacecraft/1959-059A
I don't own this domain, so I can't put any data there
The domain owner may choose to use this URI to mean
something different to me


Use Neutral URIs

http://.../spacecraft/1959-059A
The URI contains meaningful, natural or system-neutral keys.
I can route these URIs to any implementation, meaning they
can live forever
Bonus: I'm not advertising my technology choices

http://.../index.aspx?id=3
Nobody wants to know that you're using ASP.Net, really
How will you keep URIs the same when you change or
upgrade your technology?


Use http:// URIs

everyone know what to do with http:// URIs
http:// URIs provide a quick, easy and scalable lookup
mechanism

info:space/1969-059A
http://info-uri.info
info URIs require specialised knowledge
info URIs provide no lookup mechanism


Use Natural Keys

Anyone working with the original data can 'guess' your URIs
You can visually debug, and remember, your own URIs

http://.../809c6e
Nobody can guess this


Use Containers (sparingly)

Containers help keep natural keys separate
Containers provide a natural place to put lists

http://.../1969-059A
What URI do I put a list of all the spacecraft at now?


Use Dates (sparingly)

http://.../statistics/2010/agencies
Dates provide a way to show data from a period in time
Most useful for things like statistics, samples or readings

http://.../statistics/agencies
Where am I going to put next year's statistics?


Distinguishing Resources
http://example.com/1969-059A.html#self ← Thing

http://example.com/1969-059A.html ← HTML page


Fragment Identiﬁers #
• Useful, with caution
• Fragment identiﬁers are not sent to the
server
• Limits server-side decision making
• Limits granularity of the response
• Sometimes used with schema when
terms link to human-readable
documentation on the Web.


Turtle:
Terse RDF Triple
Language

Turtle
• Terse Triple Language (TTL)
http://www.w3.org/2007/02/turtle/primer/

• A human-readable notation for RDF
• Provides useful syntax for quickly
hand-writing RDF
• Forms the basis of patterns in
SPARQL


http://... a http://...
/1969-059A /Spacecraft

<http://.../1969-059A> a <http://.../Spacecraft> .


http://...
/1969-059A

<http://example.com/spacecraft/1969-059A>


http://...
/1969-059A

a space:Spacecraft ;

a http://...
/Spacecraft


http://...
/1969-059A

name
Apollo 11 Command and Service Module

space:name "Apollo 11 Command and Service Module" ;

a http://...
/Spacecraft


http://...
/1969-059A

name

launched
16 July 1969

space:launched "1969-07-16"^^xsd:date ;

a http://...
/Spacecraft


http://...
/1969-059A

name

launched
16 July 1969

mass
28801.0 a space:Spacecraft ;
space:mass 28801.0 .
a http://...
/Spacecraft


http://...
/1969-059A

name

launched @prefix space: <http://example.com/space-schema/> .
16 July 1969
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

mass
a http://...
/Spacecraft


Example Turtle
@prefix space: <http://example.com/space-schema/> .



http://... launchsite http://...
/1969-059A /cape+cana...

name name

16 July 1969

mass
a http://...
/Spacecraft


/1969-059A /cape+cana...

name name

16 July 1969

mass
;
a http://... space:launchsite <http://.../cape+cana...> .
/Spacecraft


/1969-059A /cape+cana...

name name

16 July 1969

mass
;
a http://... space:launchsite <http://.../cape+cana...> .
/Spacecraft

<http://.../cape+cana...>
space:name "Cape Canaveral, United States"@en-us .


Example Turtle

space:mass 28801.0 ;
space:launchsite <http://.../cape+cana...> .

<http://.../cape+cana...>
space:name "Cape Canaveral, United States"@en-us .


http://... a http://...
/1969-059A /Spacecraft

<?xml version="1.0" encoding="utf-8"?>
<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">

<rdf:Description rdf:about="http://.../1969-059A">
<rdf:type rdf:resource="http://.../Spacecraft"/>
</rdf:Description>

</rdf:RDF>


/1969-059A /cape+cana...

name name

launched
16 July 1969

mass
28801.0

a http://...
/Spacecraft


Example RDF/XML
<?xml version="1.0"?>
<rdf:RDF
xmlns:space="http://example.com/space-schema/"
xmlns:xsd="http://www.w3.org/2001/XMLSchema#"

<rdf:Description rdf:about="http://example.com/spacecraft/1969-059A">
<rdf:type
rdf:resource="http://example.com/space-schema/Spacecraft">
<space:name>
</space:name>
<space:launched rdf:datatype="http://www.w3.org/2001/XMLSchema#date">
1969-07-16
</space:launched>
<space:mass rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">
28801
</space:mass>
<space:launchsite
rdf:resource="http://.../cape+cana..." />
</rdf:Description>

<rdf:Description rdf:about="http://.../cape+cana...">
<space:name xml:lang="en-us">
Cape Canaveral, United States
</space:name>
</rdf:Description>

</rdf:RDF>


Example RDF/XML
<rdf:RDF

<rdf:Description rdf:about="http://example.com/spacecraft/1969-059A"> subject .
<rdf:type property .
rdf:resource="http://example.com/space-schema/Spacecraft"> value .
<space:name> property .
Apollo 11 Command and Service Module value .
</space:name>
<space:launched rdf:datatype="http://www.w3.org/2001/XMLSchema#date"> property .
1969-07-16 value .
</space:launched>
<space:mass rdf:datatype="http://www.w3.org/2001/XMLSchema#integer"> property .
28801 value .
</space:mass>
<space:launchsite property .
rdf:resource="http://.../cape+cana..." /> value .
</rdf:Description>

<rdf:Description rdf:about="http://.../cape+cana..."> subject .
<space:name xml:lang="en-us"> property .
Cape Canaveral, United States value .
</space:name>
</rdf:Description>

</rdf:RDF>


Example RDF/XML
<rdf:RDF

<space:Spacecraft rdf:about="http://example.com/spacecraft/1969-059A">
<rdf:type
rdf:resource="http://example.com/space-schema/Spacecraft">
<space:name>
</space:name>
1969-07-16
</space:launched>
28801
</space:mass>
<space:launchsite
rdf:resource="http://.../cape+cana..." />
</space:Spacecraft>

</space:name>
</rdf:Description>

</rdf:RDF>


Example RDF/XML
<rdf:RDF

<space:Spacecraft rdf:about="http://example.com/spacecraft/1969-059A">

<space:name>
</space:name>
1969-07-16
</space:launched>
28801
</space:mass>
<space:launchsite>
</space:name>
</rdf:Description>
</space:launchsite>
</space:Spacecraft>

</rdf:RDF>


Other RDF Formats

Writing RDF
• Turtle
Easily human readable, to become a W3C Recommendation
(standard)

• RDF/XML
Machine to Machine Transfer, W3C Recommendation (standard)

• N-Triples
One statement per line, subset of Turtle, grep compatible

• RDF/JSON
Useful for AJAX Applications and anything with else JSON support

• RDFa
Writing RDF within HTML pages


Example N-Triples
<http://example.com/spacecraft/1969-059A> <http://
www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://
example.com/space-schema/Spacecraft> .
example.com/space-schema/name> "Apollo 11 Command and
Service Module" .
example.com/space-schema/launched>
"1969-07-16"^^<http://www.w3.org/2001/XMLSchema#date> .
example.com/space-schema/mass> "28801"^^<http://
www.w3.org/2001/XMLSchema#integer> .
example.com/space-schema/launchsite> <http://.../cape
+cana...> .
<http://.../cape+cana...> <http://example.com/space-
schema/name> "Cape Canaveral, United States"@en-us .


Example N-Triples
<http://example.com/spacecraft/1969-059A> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.com/space-schema/Spacecraft> .
<http://example.com/spacecraft/1969-059A> <http://example.com/space-schema/name> "Apollo 11 Command and Service Module" .
<http://example.com/spacecraft/1969-059A> <http://example.com/space-schema/launched> "1969-07-16"^^<http://www.w3.org/2001/XMLSchema#date> .
<http://example.com/spacecraft/1969-059A> <http://example.com/space-schema/mass> "28801"^^<http://www.w3.org/2001/XMLSchema#integer> .
<http://example.com/spacecraft/1969-059A> <http://example.com/space-schema/launchsite> <http://.../cape+cana...> .
<http://.../cape+cana...> <http://example.com/space-schema/name> "Cape Canaveral, United States"@en-us .


Example RDF/JSON
{
"http://example.com/spacecraft/1969-059A" : {
"http://example.com/space-schema/launched" : [ {
"value" : "1969-07-16",
"datatype" : "http://www.w3.org/2001/XMLSchema#date",
"type" : "literal"
} ],
"http://example.com/space-schema/launchsite" : [ {
"value" : "http://.../cape+cana...",
"type" : "uri"
} ],
"http://example.com/space-schema/mass" : [ {
"value" : "28801",
"datatype" : "http://www.w3.org/2001/XMLSchema#integer",
"type" : "literal"
} ],
"http://example.com/space-schema/name" : [ {
"value" : "Apollo 11 Command and Service Module",
"type" : "literal"
} ],
"http://www.w3.org/1999/02/22-rdf-syntax-ns#type" : [ {
"value" : "http://example.com/space-schema/Spacecraft",
"type" : "uri"
} ]
},
"http://.../cape+cana..." : {
"http://example.com/space-schema/name" : [ {
"value" : "Cape Canaveral, United States",
"lang" : "en-us",
"type" : "literal"
} ]
}
}


RDF Schema

Who has written a
schema before?


Step 1.
Forget what you
think you know
about schema


Differences
• In a relational database:
• the schema is external to the data
• the schema comes ﬁrst
• you must know the schema to query
• In RDF:
• the schema is in the data
• you may query at any time


The Open World
• RDF Schema follow the Open World
assumption
• This model assumes that you may not
have all of the information — that more
things may be said elsewhere
• This means that missing isn't
necessarily broken
• RDF Schema are not used for
validation


If RDF Schema
aren't for
validation, then
what are they for?


RDF Validation
• The Web of Data is about modeling for
re-use
• Reusability is provided by modeling
outside of any one application's
context
• Validation is speciﬁc to an application's
context


RDF Schema
• RDF Schema are descriptive
• They add additional information to the
graph, through inference
• They add labels to properties and
classes that make dynamic interfaces
possible
• They provide human-readable
descriptions to support schema re-use
• RDF Schema are written as RDF

Important
• Properties are ﬁrst class citizens
they're not hung off classes

• Classes are not OO classes
they do not have any behavior (methods)
or properties...


launchsite
http://... http://...
/1969-059A /cape+cana...

name name

launched
16 July 1969

mass
28801.0

a http://...
/Spacecraft


Schema Re-use
• RDF & RDFS
rdfs:label — labels for UI use

• Dublin Core
Bibliographic Metadata — title, creator

• FOAF
Friend of a Friend — people, organisations, creations, topics

• SIOC
Online communities — users, accounts, forums, blogs, microblogs

• Geo
Geographical locations — latitude, longitude

• SKOS
Existing taxonomies, thesauri and classiﬁcation schemes


Schema Re-use
Property Schema
name (on Spacecraft)

launched

mass

launchsite

name (on Launchsite)


Schema Re-use
Property Schema
name (on Spacecraft) foaf:name

launched space:launched

mass space:mass

launchsite space:launchsite

name (on Launchsite) foaf:name


RDF Schema
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .


space:launchsite a rdf:Property ;
rdfs:label "Launch Site"@en ;
rdfs:description "The site from which something was launched,
examples would be the launch site of a spacecraft, rocket or
ship"@en ;
rdfs:isDefinedBy space: .

space:mass a rdf:Property ;
rdfs:label "Mass"@en ;
rdfs:description "The mass of an object, measured in Kg" ;


RDF Schema


rdfs:description "The site from which something was launched,
examples would be the launch site of a spacecraft, rocket or ship"@en ;


space:Spacecraft a rdfs:Class ;
rdfs:label "Spacecraft"@en ;
rdfs:description "A Spacecraft is any man-made object intentionally
launched into space" ;


Typing with Ranges and Domains

• Properties can add class information to
the subject and value either side of them
• A domain adds type information to the
subject
• A range adds type information to the
value
• Domains and ranges add type
information from the schema to the
instance data


A Launch Site is
any place where
something was
launched


RDF Schema


rdfs:description "The site from which something was launched, examples
would be the launch site of a spacecraft, rocket or ship"@en ;
rdfs:range space:Launchsite ;


rdfs:description "A Spacecraft is any man-made object intentionally
launched into space" ;


/1969-059A /cape+cana...

name

a http://...
/Spacecraft


/1969-059A /cape+cana...

Each
property name

is a

resource

http://...
/launchsite

a http://...
/Spacecraft


/1969-059A /cape+cana...

Each
property name

is a

resource

http://... http://...
/launchsite /Launchsite

range

a http://...
/Spacecraft


/1969-059A /cape+cana...

Each
property name

is a

resource

http://... a http://...

range

a http://...
/Spacecraft


Property and Class Hierarchies

• Properties and classes can be more
specialised versions of others
• This is done through rdfs:subPropertyOf
and rdfs:subClassOf


RDF Schema
@prefix lode: <http://linkedevents.org/ontology/> .


rdfs:description "The site from which something was launched, examples would
be the launch site of a spacecraft, rocket or ship"@en ;
rdfs:range space:Launchsite ;
rdfs:subPropertyOf lode:atPlace ;


rdfs:description "A Spacecraft is any man-made object intentionally launched
into space" ;


/1969-059A /cape+cana...

name

http://... a http://...

range

a http://...
/Spacecraft


/1969-059A /cape+cana...

name

http://... a http://...

range

a http://... subPropertyOf http://...
/Spacecraft /atPlace


atPlace

/1969-059A /cape+cana...

Each
property name

is a

resource

http://... a http://...

range

a http://... subPropertyOf http://...
/Spacecraft /atPlace


Naming Convention
• Properties are generally in camel-case
name; launchDate; launchVehicle; crewMember

• RDF Class names are usually in title
case
Person; Spacecraft; Discipline

• Properties use the simplest name they
can
launch; time; destination

• Inverse properties often use 'of'
name & nameOf; crewMember & crewMemberOf


Thanks to

• Rob Styles, Talis Group Ltd for his CC-BY
licensed “Two Day Introduction to the Web of
Data” slides and spacecraft example data.


This work is Copyright © 2010 David Hyland-Wood
Based partially on CC-BY licensed content Copyright © 2010 Talis Systems Limited by Rob Styles.
It is licensed under the Creative Commons Attribution 3.0 Unported License.
Full details at: http://creativecommons.org/licenses/by/3.0/

You are free:

to Share — to copy, distribute and transmit the work

to Remix — to adapt the work

Under the following conditions:
Attribution. You must attribute the work in the manner speciﬁed by the author or licensor (but
not in any way that suggests that they endorse you or your use of the work).

• For any reuse or distribution, you must make clear to others the license terms of this work.
• Any of the above conditions can be waived if you get permission from the copyright holder.
• Nothing in this license impairs or restricts the author's moral rights.
• Some Content in the work may be licensed under different terms, this is noted separately.


Intro to Linked Data: Data Modeling

More Related Content

Similar to Intro to Linked Data: Data Modeling

More from David Wood

Recently uploaded

Intro to Linked Data: Data Modeling