[2024]Digital Global Overview Report 2024 Meltwater.pdf
Propelling the Potential of Linked Data in Enterprises
1. PROPEL: Topic and trend analysis
Javier D. Fernández, Sabrina Kirrane, Axel Polleres
2. Interviews & Workshop
23 interviews:
Domains
Consulting, Engineering, Environment,
Finance and Insurance, Government,
Healthcare, ICT, IT, Media,
Pharmaceutical, Professional Services,
Real Estate, Research, Startup, Tourism,
Transports & Logistics
Roles
Business Intelligence, CEO, Chief
Engineer, Data and Systems Architect,
Data Scientist, Director Information
Management, Enterprise Architect,
Founder, General Secretary, Governance,
Risk & Compliance Manager, Head of
Communications and Media, Head of
Development, Head of HR, Head of R&D,
Innovation Manager, Information Architect,
IT Project Manager, Management,
Managing director, Marketing Analyst,
Principle System Analyst, Project
Coordinator, Researcher, Technical
Specialist
Took place on the 10th of May 2016
• private sector (solution providers and
users)
• research sector (technicians and
strategic people)
Business barriers and drivers
Technological challenges and
opportunities
PROPEL 2
3. User Stories
4 out of 60 user stories we collected in the interviews:
Horizontals
• Business processes (e.g. product logistics and supply chain
management)
• Human resources (e.g. expert and resource management)
Verticals
• Media & Publishing
• Healthcare & Pharma
PROPEL 3
4. Technologies in need…
PROPEL 4
Analytics
Computational
linguistics & NLP
Concept tagging &
annotation
Data integration
Data management
Dynamic data /
streaming
Extraction, data
mining, text mining,
entity extraction
Logic, formal
languages &
reasoning
Human-Computer
Interaction &
visualization
Knowledge
representation
Machine learning
Ontology/thesaurus
/taxonomy
management
Quality &
Provenance
Recommendations
Robustness,
scalability,
optimization and
performance
Searching,
browsing &
exploration
Security and
privacy
System engineering
We pretty much
ended up in all
areas that SW
touches upon!
5. Business Processes
PROPEL 5
“I would like to be able
to exchange information
and coordinate
production and logistics
with suppliers and
customers…”
“…so that I can improve
efficiency, effectiveness
and flexibility of my
inventory management
and operations”
Analytics
Computational
linguistics & NLP
Concept tagging
& annotation
Data integration
Data
management
Dynamic data /
streaming
Extraction, data
mining, text
mining, entity
extraction
Logic, formal
languages &
reasoning
Human-
Computer
Interaction &
visualization
Knowledge
representation
Machine learning
Ontology/thesaur
us/taxonomy
management
Data
Quality &
Provenance
Recommendations
Robustness,
scalability,
optimization and
performance
Searching,
browsing &
exploration
Security and
privacy
System
engineering
6. Human Resources
PROPEL 6
“I would like identify
expertise within our
large organisation and
be able to pinpoint the
relevant experts…”
“…so that I can I can
identify top trends
within the organisation
and expertise for the
organisation as a
whole”
Knowledge
representation
Analytics
Computational
linguistics & NLP
Concept tagging
& annotation
Data integration
Data
management
Dynamic data /
streaming
Extraction, data
mining, text
mining, entity
extraction
Logic, formal
languages &
reasoning
Human-
Computer
Interaction &
visualization
Knowledge
representation
Machine learning
Ontology/thesaur
us/taxonomy
management
Quality&
Provenance
Recommendations
Robustness,
scalability,
optimization and
performance
Searching,
browsing &
exploration
Security and
privacy
System
engineering
7. Media & Publishing
PROPEL 7
I would like to display
personalized content as
precise as possible
So that my readers
stay as long as possible
on my website.
Analytics
Computational
linguistics & NLP
Concept tagging
& annotation
Data integration
Data
management
Dynamic data /
streaming
Extraction, data
mining, text
mining, entity
extraction
Logic, formal
languages &
reasoning
Human-
Computer
Interaction &
visualization
Knowledge
representation
Machine learning
Ontology/thesaur
us/taxonomy
management
Data
Quality &
Provenance
Recommendations
Robustness,
scalability,
optimization and
performance
Searching,
browsing &
exploration
Security and
privacy
System
engineering
8. Healthcare & Pharma
PROPEL 8
I would like to Integrate
disparate systems that are:
-Hard to integrate
-Widespread
-Contain the same data that
contradicts each other
So that I can gain insights
from other clinical trials
Analytics
Computational
linguistics & NLP
Concept tagging
& annotation
Data integration
Data
management
Dynamic data /
streaming
Extraction, data
mining, text
mining, entity
extraction
Logic, formal
languages &
reasoning
Human-
Computer
Interaction &
visualization
Knowledge
representation
Machine learning
Ontology/thesaur
us/taxonomy
management
Data
Quality &
Provenance
Recommendations
Robustness,
scalability,
optimization and
performance
Searching,
browsing &
exploration
Security and
privacy
System
engineering
9. Let’s take a step back…
What can we offer as a community?
PROPEL 9
Taking an introspective view
15. Semantic Web/Linked Data over time…
PROPEL 15
Early adopters:
MITRE
Chevron
British Telecom
Boeing
Ordnance Survey
Eli Lily
Pfizer
Agfa
Food and Drug Administration
National Institutes of Health
Software adopters/products:
Oracle
Adobe
Altova
OpenLink
TopQuadrant
Software AG
Aduna Software
Protége
SAPHIRE
17. LD Adopters - Companies
PROPEL 17
0
200
400
600
800
1000
1200
1400
1600
Google Oracle Yahoo SAP IEEE
Intelligent
Systems
Franz Bing Expert
System
IBM Research Poolparty
Occurrences
Companies
Conference Sponsors that appear in papers 2006-2015
19. LD Adopters - Domains
PROPEL 19
0
5000
10000
15000
20000
25000
30000
35000
occurrences
Domains
Topics grouped by domain 2006-2015
Well, they
publish in
other
venues…
E.g.:
20. Semantic Web/Linked Data over time…
PROPEL 20
The authors claim that "early research has
transitioned into these larger, more
applied systems, today’s Semantic Web
research is changing: It builds on the
earlier foundations but it has generated a
more diverse set of pursuits”.
22. Roadmap for Enterprise LD?
1. Linked Data security and privacy requirements
2. Determining and resolving data quality issues
3. Managing large amounts of data and associated metadata such
as provenance and temporal data
4. Visualisation requirements
5. Understand the different application areas and their maturity in
terms of real world deployment
6. Showcases in terms of startups, projects, applications and
systems based on Linked Data technologies
7. Linked Data and recommender system models
8. Dynamic linked datasets
22
https://www.linked-data.at/
Notas do Editor
From the interviews we derived approximately 60 user stories and mapped them to the 18 Foundations listed earlier
https://docs.google.com/spreadsheets/d/1x3s0r8Wlg5rt9paa5CwoP0MYwHPqlLj66WbClctbV9w/edit#gid=0
I selected 4 and added them here, however there are others incase you need them
From the interviews we derived approximately 60 user stories and mapped them to the 18 Foundation
https://docs.google.com/spreadsheets/d/1x3s0r8Wlg5rt9paa5CwoP0MYwHPqlLj66WbClctbV9w/edit#gid=0
I selected 4 and added them here, however there are others incase you need them
Horizontal – applicable to any business
Vertical – specific to a domain
Result of a stakeholder workshop we conducted with those people…
These are the foundations that we use for mapping the user stories from WP2 to the topics from WP3, therefore it is good to introduce them here
Analytics
Computational linguistics & NLP
Concept tagging & annotation (******Personally I’m not convinced about this one******)
Data integration
Data management
Dynamic data / streaming
Extraction, data mining, text mining, entity extraction (******Personally I’m not convinced the overlap with NLP******)
Logic, formal languages & reasoning
Human-Computer Interaction & visualization
Knowledge representation
Machine learning
Ontology/thesaurus/taxonomy management
Quality
Recommendations
Robustness, scalability, optimization and performance
Searching, browsing & exploration
Security and privacy
System engineering
****TODO Look for images that are relevant for this use case
High level conclusion - Generally speaking we don’t need NLP and Analytics for this use case
Analytics (Not needed for this use case)
Computational linguistics & NLP (Not needed for this use case)
Concept tagging & annotation (?)
Data integration (Different suppliers have different formats therefore this is essential)
Data management (Need to be able to handle provenance, temporal data, updates etc)
Dynamic data / streaming (Order may be streamed or batched so not essential)
Extraction, data mining, text mining, entity extraction (Entities will need to be extracted from line of business applications)
Logic, formal languages & reasoning (Not needed for this use case)
Human-Computer Interaction & visualization (This would be a nice to have but not essential, simple searching and browsing would suffice)
Knowledge representation (nice to have semantics may not be essential as exchange formats can be agreed upfront)
Machine learning (Not needed for this use case)
Ontology/thesaurus/taxonomy management (nice to have semantics may not be essential as exchange formats can be agreed upfront)
Quality (This is very important for line of business applications)
Recommendations (Not needed for this use case)
Robustness, scalability, optimization and performance (Not a big data use case but they still need to be able to handle reasonable data sizes)
Searching, browsing & exploration (The main use case is around the exchange but searching and browsing would be useful)
Security and privacy (This is very important for line of business applications)
System engineering (APIs, Services etc will be needed)
****TODO Look for images that are relevant for this use case
High level conclusion - Generally speaking we don’t Analytics for this usecase
Analytics (Not needed for this use case)
Computational linguistics & NLP (expertise may be in text form)
Concept tagging & annotation (?)
Data integration (Extertise profiles may be in many disparate systems)
Data management (Need to be able to handle provenance, temporal data, updates etc)
Dynamic data / streaming (Not needed for this use case)
Extraction, data mining, text mining, entity extraction (Expertise will need to be extracted from structured and unstructured sources)
Logic, formal languages & reasoning (Not needed for this use case)
Human-Computer Interaction & visualization (This would be a nice to have but not essential, simple searching and browsing would suffice)
Knowledge representation (Considering all the different potential data sources a common representation model would be useful)
Machine learning (Not needed for this use case)
Ontology/thesaurus/taxonomy management (Considering all the different potential data sources a common structure(s) would be useful)
Quality (This is very important for line of business applications)
Recommendations (Not needed for this use case)
Robustness, scalability, optimization and performance (Not a big data use case but they still need to be able to handle reasonable data sizes)
Searching, browsing & exploration (Searching and browsing is necessary)
Security and privacy (This is very important for line of business applications)
System engineering (APIs, Services etc will be needed)
TODO Look for images that are relevant for this use case
- This usecase could really benefit from the technology the community are working on
SK to provide reasons for the colouring (Tue AM)
TODO Look for images that are relevant for this use case
- This usecase could really benefit from the technology the community are working on
SK to provide reasons for the colouring (Tue AM)
A not-yet-very-scientific approach… still hope this is interesting, maybe a bit controversial to discuss here!
Outlines quite clearly what they thought back then the Semantic Web should be…
Terms to do with Agents and Web Services from conference/journal dictionary – there is no high level foundation so we might need to merge some terms
ontologies ontology management ontology engineering ontology languages
agents web services software agents services agent-based
… agents research definitly not going up, our community has largely been dominated by ontologies.
Might need data before 2006, semanttic Web services topic was already on the decline in 2006.
The year of ”Linked Data”
DBPedia
A lot of company use cases that have used SW mentioned:
Top 10 Companies plot from Sponsor Dictionary
Can look over companies per year FORM_
Top 5 OR 10 Domains plot from conference/journal dictionary
Here healthcare seems unimportant…
hmm, I think that I don't have enough data for that as of yet ;-)
… but that also gives me a chance to just throw in my own personal tast preferences!!!
What I believe to be hot topics/ the (no-brainers, if you want...)
* Move into messier data, don’t focus on just structured data and non-structured data ... I think there is a niche!
Combining Open and Closed Data Data Scecurity and Privacy
Archiving and storage temporal data, efficient indexing and efficient updates
“Andreas Braun (CIO Allianz): Hadoop to be faded out, privacy by design and by default top agenda item”