What happens when tens of thousands of archival photos are shared with open licenses, then mashed up with geolocation data and current photos? Or when app developers can freely utilize information and images from millions of books? On this panel, we'll explore the fundamental elements of Linked Open Data and discover how rapidly growing access to metadata within the world's libraries, archives and museums is opening exciting new possibilities for understanding our past, and may help in predicting our future. Our panelists will look into the technological underpinnings of Linked Open Data, demonstrate use cases and applications, and consider the possibilities of such data for scholarly research, preservation, commercial interests, and the future of cultural heritage data.
3. to expose ‘The London Art World 1660-1735’
dataset - several years of history of art
research trawling primary and secondary
sources of information about into art sales,
people, places and artworks all contained in
spreadsheets
5. ok, so it’s on the web, it has some
links, it’s open, right? can I go
now? ... not so fast, I’m not done
yet
6. how does a machine know
and these are links to info about
that this is about an art people and places?
sale?
and how can someone get
at this info and do like enrich it with
interesting things with it? information from
elsewhere?
7. linked open data
... describing real-world things and
the relationships between them in a
machine-readable way
8. in walks RDF:
Resource (identifying resources on the
web) Description (and describing them)
Framework (with a model based on
triples and graphs)
9. PREDICATE
SUBJECT (aka relationship) OBJECT
<someArtist> <occupied> <somePlace>
<someArtist> <painted> <somePainting>
<somePainting> <soldIn> <someSale>
<someSale> <happenedIn> <somePlace>
all of these
<someCatalogue> <describes> <someSale> will be uris
<someSaleItem> <soldFor> <somePrice>
RDF - <someBuyer> <purchased> <someSaleItem>
all this is not
about the rdf you
are looking
triples for
10. an ontology standardized representation
is a
of knowledge as a set of concepts within a
domain, and the relationships between those
concepts. It can be used to reason about the
entities within that domain, and may be used
to describe the domain (wikipedia)
or, put another way “a standard way of describing stuff
for a given domain” (me)
11. we should either use terms from
existing ontologies or create and
publish our terms using standard
approaches
12. we created an event-driven ontology based
on DUL (DOLCE Ultra Lite) and LODE
(Linked Open Events)
why?
because we wanted to create rich and
specific data but ensure our data could still
be understood in a generic and low barrier
way
15. we linked our people to viaf and some of
our places to geonames ...
<ourPerson> <sameas> <viafPerson>
<ourPlace> <sameas> <geonamesPlace>
... a data consumer can start following this
network of links
18. spreadsheet cleanup with scripting, a database
and some Google refine action*
* google refine is very useful for dealing with
messy spreadsheets + has an rdf plugin
19. a turtle* document for each of our 38,000
primary ‘entities’
sale
person
place
artwork
source
stored in dlib.york.ac.uk and indexed in
sindice.com** semantic search engine
* a format for creating rdf data
** try a search for sale domain:dlib.york.ac.uk
20. foaf:primaryTopic <http://dlib.york.ac.uk/id/place/34867>;
rdf:type foaf:Document, dctype:Text .
<http://dlib.york.ac.uk/data/place/34867/turtle>
void:inDataset <http://dlib.york.ac.uk/data/void.ttl#OpenART>;
rdf:type foaf:Document, dctype:Text .
<http://dlib.york.ac.uk/data/place/34867/rdf>
void:inDataset <http://dlib.york.ac.uk/data/void.ttl#OpenART>;
rdf:type foaf:Document, dctype:Text .
<http://dlib.york.ac.uk/id/place/34867>
mapping:hasResearchID "3.0548"^^<xsd:string>;
SUBJECT rdfs:label "The Green Doors in the Little Piazza, Covent Garden; sale venue";
vocupper:hasPlaceName "The Green Doors in the Little Piazza, Covent Garden";
vocupper:hasBuildingName "The Green Doors";
vocupper:hasStreetName "Little Piazza";
vocupper:hasCity "London";
vocupper:hasCounty "Greater London";
vocupper:hasCountry "England";
vochoa:hasContributorOfSource "Richard Stephens";
OBJECT
oactxt:venueOfSale <http://dlib.york.ac.uk/id/sale/3494
8>;
PREDICATE oactxt:venueOfSale <http://dlib.york.ac.uk/id/sale/34949>;
oactxt:venueOfSale <http://dlib.york.ac.uk/id/sale/34950>;
oactxt:venueOfSale <http://dlib.york.ac.uk/id/sale/34951>;
oactxt:venueOfSale <http://dlib.york.ac.uk/id/sale/34952>;
vocupper:liesWithin [
owl:sameas <http://www.geonames.org/6269131/>;
rdf:type model:Place, vocupper:Country,
LINKED
owl:NamedIndividual
];
rdf:type model:Place, owl:NamedIndividual .
21. DISCLAIMER
ours was one approach
it is very experimental and is
imperfect in various ways
it showed that we could do linked
data with an existing system
we want to do more
22. linked open data is leap of
faith -
you have to expose data
before people can consume
data
23. aim high -
if we all put out high quality
rich data we can do high
quality AND low barrier
things with it
24. there are 77981 results for
‘York’ in geonames
we had a little project in York ...
25. we had a little project in
http://www.geonames.org/2633352/
...
26. credits
Richard Stephens: data creator
Tate: data partners
Martin Dow: ontology dev
Stephen Bayliss: ontology dev
Paul Young: data transform
LOCAH project: inspiration
Jon Voss: lodlam guru
University of York: institutional support
JISC: funding
@julieallinson
julie.allinson@york.ac.uk
http://tinyurl.com/dlib-openart
#LODLAM #sxsw