2. DS
RC
Outline
1. Goals of the DSRC
2. Embedding and organization
3. Realizing the DSRC
4. The roadmap
5. How to get involved
3. DS
RC
The goal
• Become a leading center on data science by
developing the new data science discipline
Leveraging our scientific excellence
Leveraging our tools and infrastructure
Reaching out
Educating talents
4. DS
RC
Our contributions to data
science
Visual
Analytics
Business
Analytics
Decision
Theory
Understand
and decide
Distributed
Processing
Large Scale
Databases
Store and
process
Software
Eng.
System /
Network
Eng.
10010001010010001010010101
01010001010101010101010101
01000101010101010101010101
00101010010101010101010101
Security
01001010100000101010101001
Privacy
01010100010100101010101000
11110101010101010101010101
Provenance
01010100010101010101010101
11110101010101010101010101
00101010101010101010101010
00101010101010101010101010
Data
Reasoning
Knowledge
representati
on
Analyze
and model
Multimedia
Retrieval
Modeling
and
simulation
Information
Retrieval
Machine
Learning
5. DS
RC
Goals in data science
research
Visual
Analytics
Decision
Theory
Business
Analytics
Insight
Impact
Understand
and decide
Distributed
Processing
Data
Conformance
Reasoning
Knowledge
representati
on
Large Scale
Databases
Precision
Analyze
and model
Speed and
Store
process
Efficiency
Software
Scalability
Eng.
System /
Network
Eng.
Recall
Model fit
Multimedia
Retrieval
Modeling
and
simulation
Information
Retrieval
Machine
Learning
6. DS
RC
Data Science
• Characteristics
– All are connected
– All driven by data and its use
– A holistic approach is needed
• Our answer
– The Data Science Research Center
7. DS
RC
Research in data science
Visual
Analytics
Business
Analytics
Decision
Theory
Understand
and decide
Distributed
Processing
Large Scale
Databases
Store and
process
Software
Eng.
System /
Network
Eng.
Data
Science
Data
Research
Center
Reasoning
Knowledge
representati
on
Analyze
and model
Multimedia
Retrieval
Modeling
and
simulation
Information
Retrieval
Machine
Learning
8. DS
RC
Embedding and relations
Center for content,
creation and technology
Department of mathematics
Network Institute
CWI
Center for
Digital Humanities
ILLC
HvA
Informatics Institute
UvA
DSRC
Department of Computer Science VU
E-science center
SURFsara
CLHC
Forensic Science
Amsterdam Business
School
9. DS
RC
Organization
Daily management team
Marcel Worring
Paul Groth
Sanne Veenenbos
Management board
Max Welling
Henri Bal
Bert Bredeweg
Ger Koole
Leading researchers
Cees de Laat
Sander Klous
Jacopo Urbani
Frank van Harmelen
Peter Boncz
Maarten de Rijke Arnold Smeulders
10. DS
RC
Realization: four aims
Research: a platform for research in data science
connecting people and methodologies.
Infrastructure: a data-driven infrastructure for
experimenting with realistic complex data sets.
Valorization: a channel between scientific research
and third party applications.
Education: data-science curricula with realistic data
experimentation throughout the program.
11. DS
RC
Research
• Focus on research with a holistic view on
data science
– Connecting the different disciplines
– From data to domain impact
• Start
– Seed projects: small projects bringing together two
or more ICT disciplines
• Workshops
– Domain workshops: with all stakeholders define
research topics leading to data science project
proposals
12. DS
RC
Infrastructure
“In a sense, the physical and technical infrastructure
becomes invisible and the data themselves become
the infrastructure – a valuable asset, on which
science, technology, the economy and society can
advance.”
[“Riding the wave” EU High Level Expert Group]
Shared domain driven tasks
Shared large
scale
datasources
Transparant access
to distributed computing
infrastructure
Common tools and
code bases
13. DS
RC
Valorization
• Joint full projects
– Within the DSRC
– With industry / govermental organizations
• Small-scale projects
– From data and problem to solution with quick
turnaround
• Competitions
– From data and problem to innovative solutions
worked on by a number of teams
• Spin-offs and startups
14. DS
RC
Education
• Infrastructure yields platform for education in
– Informatics
• Information Science, Artificial Intelligence, Software
Engineering, Computer Science, Business Analytics
– Domain specific courses
• E.g. Minor Data Science for X (your favorite discipline)
– Commercial courses
• The objective of DSRC
– to introduce a full data science program
• With hands-on experience
– On real data and real problems or innovations
17. DS
RC
Roadmap
Holistic data science
Year 1
Fully functional
Basic transparent
infrastructure
Start of new PhD projects
Seed projects
Infrastructure design
Domain workshops
Data acquisition
Invited talks
Year 2
Self-sustained
Data science program
Embedding
infrastructure
Start of projects
Courses
Domain workshops
Project acquisition
RDA
Minor(s) data science
Year 3 and 4
PIRE
18. DS
RC
How to get involved
• As a data science researcher
–
–
–
–
–
See what we can offer and what you can offer
Define one of the seed projects
Participate/propose workshops
Acknowledge DSRC in publications
Bring in (existing or new) projects
• Contact us via
– info@dsrc.nl
19. DS
RC
How to get involved
• As a data-driven application holder
–
–
–
–
Participate in the workshops and events
See how you can share your data
See whether we can develop joint projects
Link domain knowledge and data science
research
• Contact us via
– info@dsrc.nl
20. DS
RC
Overview of the talks
See some of the demos
Visual
Analytics
Business
Analytics
Decision
Theory
Understand
and decide
Henri Bal
Peter Boncz
Distributed
Processing
Large Scale
Databases
Store and
process
Software
Eng.
System /
Network
Eng.
10010001010010001010010101
01010001010101010101010101
01000101010101010101010101
00101010010101010101010101
01001010100000101010101001
01010100010100101010101000
11110101010101010101010101
01010100010101010101010101
11110101010101010101010101
00101010101010101010101010
00101010101010101010101010
Data
Reasoning
Knowledge
representati
on
Analyze
and model
Multimedia
Retrieval
Modeling
and
simulation
Information
Retrieval
Machine
Learning
Max Welling
Cees de Laat
Maarten de Rijke