Presentation given at OpenBelgium 2015.

1. Open Data Mashups:
linking fragments into mosaics
Philippe Duchesne, HighLatitudes
phd@highlatitud.es
@pduchesne
Open Tools & Standards workshop
OpenBelgium 2015
Namur, February 23rd, 2015

2. Linked Data
2
Data is live and
interoperable.
Almost there…

3. ds
Background
Observation: when collating
hypermedia resources, data
curators face issues such as
► not owning referred data
sources
► need resources identified by
URIs
► need to organize
annotate
store
exchange
the result of their curation as
a resource of its own Remote resources

4. Linked Data
4

5. Linked Data
5

6. Linked Data
6

7. Linked Data
7
is
has
contains
near
about

8. Linked Data
8
is
has
contains
near
about

9. Linked Data
9
Different use cases,
different link sets
►Urban planning
►Traffic
Monitoring
►Tourism
►…
… a catalog of contexts of use

10. Representation
– RDF, JSON-LD
Link vocabularies : Ontologies
– DublinCore, FOAF, DCAT, Schema.org,…
– Numerous domain-specific ontologies
– see http://lov.okfn.org
Linked Data
Standards
10

11. <rdf:RDF xmlns:foaf="http://xmlns.com/foaf/0.1/" xmlns:owl="http://www.w3.org/2002/07/owl#"
xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dcat="http://www.w3.org/ns/dcat#" xmlns:dct="http://purl.org/dc/terms/">
<dcat:Dataset rdf:about="http://portal.openbelgium.be/dataset/wms-des-sites-natura-2000-en-vigueur">
<owl:sameAs rdf:resource="urn:uuid:2a1d4058-50cd-42f2-947b-71ea9d0173ce"></owl:sameAs>
<dct:description>Le jeu de données géographiques NATURA2000 vise à localiser les sites Natura 2000 wallons et
à l'intérieur de ceux-ci les unités de gestion. A chaque unité de gestion correspond un ensemble de mesures de
gestion</dct:description>
<foaf:homepage rdf:resource="http://portal.openbelgium.be/dataset/wms-des-sites-natura-2000-en-
vigueur"></foaf:homepage>
<rdfs:label>wms-des-sites-natura-2000-en-vigueur</rdfs:label>
<dct:identifier>wms-des-sites-natura-2000-en-vigueur</dct:identifier>
<dct:title>WMS des sites Natura 2000 en vigueur</dct:title>
<dcat:distribution>
<dcat:Distribution>
<dcat:accessURL
rdf:resource="http://geoservices.wallonie.be/services/FAUNE_FLORE/NATURA_2000/WMSServer"/>
<dct:format>WMS</dct:format>
<dct:title>Natura 2000</dct:title>
</dcat:Distribution>
</dcat:distribution>
<dct:relation>
<rdf:Description>
<rdfs:label>responsible-party</rdfs:label>
<rdf:value>[{"name": "SPW-DGO3-DEMNA-DCD", "roles": ["pointOfContact"]}]</rdf:value>
</rdf:Description>
</dct:relation> […]
</dcat:Dataset>
</rdf:RDF>
Linked Data
Examples
11
http://portal.openbelgium.be/dataset/wms-des-sites-natura-2000-en-vigueur.rdf

12. 12
http://lod-cloud.net
Linked Data
Examples

13. ds
Background
Observation: when collating
hypermedia resources, data
curators face issues such as
► not owning referred data
sources
► need only parts of resources
identified by URIs
► need to organize
annotate
store
exchange
the result of their curation as
a resource of its own Remote resources

14. http://server.com/myResource
URI fragments
14
#fragment

15. ► URI spec leaves the fragment part
unspecified
► Several standardization efforts exist
regarding specific mime types
– MediaFragments (W3C Recommandation)
• Focused on temporal (#t) and pixel space fragments
(#xywh)
• Dedicated to audio/video media
– text/csv fragments (IETF Draft)
• Defines #col,row,cell
– text/plain fragments (IETF RFC 5147)
• Defines #line
Formalizing
fragments
15

16. Other mainstream usages
► #t for YouTube videos
► #page for PDF resources
► #<id> implicit id fragment used in HTML,
RDF, …
But no harmonized definition of fragments
for all media types
Formalizing
fragments
16

17. Identified dimensions that are subject to
fragments
► Textual #line
► Temporal #t
► Pixel space #xywh
► Tabular #col,row
► Geospatial space #bbox
► Tree path #xpath
Formalizing
fragments
17

18. Id Text Time Pixel Spatial Tabular Tree
html ✔ ✔
text/plain ✔
pdf ✔ ✔
video/* ✔ ✔
audio/* ✔
image/* ✔
geotiff ✔ ✔
kml ✔ ✔ ✔ ✔
csv ✔
json ✔
Formalizing
fragments
18

19. ► Objective : build mosaics documents from
fragments
► Define mosaic model on top of existing
vocabularies
► Open Annotations
http://www.openannotation.org
► Seamless integration in regular browser
flow
Mosaics
19

20. Demo
20
http://demo.mosaics.highlatitud.es/#/samples/demo

21. ► Work Context storage and sharing
– Disaster management
► Embedding original data
– Open Science
► Fact checking
– Political transparency
– Data Journalism
It’s all about stitching together heterogeneous
data fragments into meaningful and
contextualized patchworks
Use cases
21

22. • Seamless integration of authoring/viewing
avoids breaking train of thought  favors
engagement and better captures domain
knowledge
• Data mosaics act as exchangeable “mind
map” containers, capturing contextual
metadata in a domain-specific perspective,
for specific use cases
Conclusions
22

23. phd@highlatitud.es
@pduchesne
Thank you
23