ld@em - what linked data does, what linked data needs

20
Linking Data at Euromoney 1 25/02/14 Fabio Colasanti

Upload: fcolasanti

Post on 06-May-2015

403 views

Category:

Technology


2 download

DESCRIPTION

Business drivers, opportunities and design patterns behind a real life implementation of Linked Data in a global publishing enterprise.

TRANSCRIPT

Page 1: LD@EM - what Linked Data does, what Linked Data needs

Linking Data at Euromoney

1

25/02/14Fabio Colasanti

Page 2: LD@EM - what Linked Data does, what Linked Data needs

A five course menu

• The problem space• Drivers and opportunities• Challenges• Patterns• Next steps

Page 3: LD@EM - what Linked Data does, what Linked Data needs

The enterprise

• Part of the Daily Mail group• Global publishing & data business• Conferences and seminars• 32 countries• 3.5K employees• 13 Business Units• 100+ publications • 100+ websites

Page 4: LD@EM - what Linked Data does, what Linked Data needs

Coverage

• Macroeconomics (BCA, …)• Capital Markets (EW, DI, GCM, …) • Metals price assessment (MB, AMM, SF, …)• Hedge funds investment (HFI, II, IISearches…)• Legal (IFLR, IP, ITR, ...)• Insurance (Insider, Reatcions, …) • Specialist (Air Finance, IndMin, …)• …

Page 5: LD@EM - what Linked Data does, what Linked Data needs

Project Delphi: drivers and vision

• Exploit high overlap of concerns to transform bounded information and segregated data into shared knowledge

• Ability to generate new transversal products offering by leveraging the knowledge graph of existing content

Page 6: LD@EM - what Linked Data does, what Linked Data needs

6

Euromoney’s Linked Data cloud

Page 7: LD@EM - what Linked Data does, what Linked Data needs

Not all fun and games

• Retain ownership • Independent change management• Editorial style• Ubiquitous language• Reference data• Information governance • Data governance

Page 8: LD@EM - what Linked Data does, what Linked Data needs

Multi-layered bounded contexts

Page 9: LD@EM - what Linked Data does, what Linked Data needs

Multi-layered bounded contexts

Page 10: LD@EM - what Linked Data does, what Linked Data needs

Multi-layered bounded contexts: example

Page 11: LD@EM - what Linked Data does, what Linked Data needs

Data and Information management

Partitioned by ownership – Ontologies– Non-Information Resources– Reference data

Risk matrixGovernance process

Page 12: LD@EM - what Linked Data does, what Linked Data needs

Labels management

@prefix eml: <http://data.emii.com/ontologies/location/> .@prefix emcore: <http://data.emii.com/ontologies/core/> .@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

<http://data.emii.com/locations/usa> rdf:type eml:Location ;emcore:canonicalLabel "United States"@en ;emcore:gcmCanonicalLabel "USA"@en ;emcore:bcaCanonicalLabel "U.S."@en ; emcore:alias ”US"@en , "U.S.A."@en , "United States"@en , "United States of America"@en .

Page 13: LD@EM - what Linked Data does, what Linked Data needs

Labels management

Page 14: LD@EM - what Linked Data does, what Linked Data needs

Labels management

@prefix emcore: <http://data.emii.com/ontologies/core/> .@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#> .@prefix dc: <http://purl.org/dc/elements/1.1/> .

emcore:xyzCanonicalLabel rdf:type owl:DatatypeProperty ;rdfs:subPropertyOf emcore:alias ;rdfs:domain emcore:Concept ;rdfs:range rdfs:Literal ;rdfs:label ”XYZ Canonical Label"@en ;rdfs:comment "The ABC preferred label for a concept."@en ;rdfs:isDefinedBy <http://data.emii.com/ontologies/core> ;dc:rights <http://data.emii.com/business-groups/xyz> .

Page 15: LD@EM - what Linked Data does, what Linked Data needs

Labels management

• Consumer tailored labels• Late binding onto semantic predicates• No hacking into URIs

Page 16: LD@EM - what Linked Data does, what Linked Data needs

Go public?

• Linked Data doesn’t have to be no-profit• Public vs Proprietary datasets• Replicate model on a larger scale• Expose public domains for searchability and

KB augmentation as semantically rich endpoints

Page 17: LD@EM - what Linked Data does, what Linked Data needs

Linked Data public access layer

Page 18: LD@EM - what Linked Data does, what Linked Data needs

Linked Data public access layer: example

Page 19: LD@EM - what Linked Data does, what Linked Data needs

It’s all about quality, it’s all about people(Get in touch, we’re hiring)