Download May 2014 - New Zealand Analytics Forum

www.pwc.co.nz
NZ Analytics Forum
Forming Successful Analytics
Partnerships
May 2014
Hosted by PwC
Agenda
 Welcome
 Analytics at PwC
 University Analytics Program Overviews
 Overview of Statistics NZ’s Integrated Data Infrastructure
 Introducing the Data Futures Forum
 Panel Discussion: How can the Government, Industry and Academia
work together better to enhance NZ’s analytics capability?
NZ Analytics Forum
PwC
May 2014
2
Analytics at PwC
NZ Analytics Forum
PwC
May 2014
3
Analytics at PwC
[Placeholder Slide]
NZ Analytics Forum
PwC
May 2014
4
Tertiary Education Analytics
Program Overviews
New Zealand Universities
NZ Analytics Forum
PwC
May 2014
5
University of Auckland
Teaching and Research
in
Analytics
Andy Philpott
[email protected]
27 May 2014
COMPUTER SCIENCE
•
•
•
•
•
•
•
•
COMPSCI 351 Fundamentals of Database Systems
COMPSCI 367 Artificial Intelligence
COMPSCI 752 Semantic Web Technologies
COMPSCI 753 Uncertainty in Data
COMPSCI 760 Data-mining and Machine Learning
COMPSCI 765 Interactive Cognitive Systems
COMPSCI 767 Intelligent Software Agents
Masters of Professional Studies in Data Science
(jointly with Statistics)
http://www.science.auckland.ac.nz/en/about/our-programmes.html
STATISTICS
• Mathematical statistics (theory)
• Applied statistics
• Statistical computing (including introductions
to R and SAS and data mining)
• Operations research and applied probability
• Data Science Practice (New 2015)
• Masters of Professional Studies in Data
Science (jointly with Computer Science)
http://www.science.auckland.ac.nz/en/about/our-programmes.html
ENGINEERING SCIENCE
(Model driven/computational business analytics)
• ENGSCI 255 Introduction to Operations Research
• ENGSCI 391 Optimization (Linear/MI programming)
• ENGSCI 355 Practice of optimization and simulation
• ENGSCI 760 Algorithms for optimization
• ENGSCI 761 Advanced integer programming
• ENGSCI 762-768
• Data-driven decision making: coming in 2015
http://www.des.auckland.ac.nz/uoa/course-details
INFORMATION MANAGEMENT
INFOMGMT 290 Business Analytics
INFOMGMT 291 Business Analytics and Applications
INFOMGMT 393 Data Mining and Decision Support
BUSMGT 712. Business Analytics
(in taught masters programme)
http://courses.business.auckland.ac.nz/SubjectListing/INFOMGMT/
RESEARCH
• Te Punaha Matatini
– Centre for Complex Systems and Networks
– One of six CoREs recently announced
– Analytics theme (funded for 6 years)
• National Science Challenge 10
– Science for Technological Innovation
– Smart Services Portfolio growing over next 10 years
Data Analytics Forum, Auckland, 27 May 2014
Machine learning and predictive data modeling of large
and stream data in KEDRI/AUT
Prof. Nikola Kasabov, FIEEE, FRSNZ, FIITP
Director, Knowledge Engineering and Discovery Research Institute (KEDRI),
and School of Computer and Mathematical Sciences
Auckland University of Technology, New Zealand
[email protected]
www.kedri.info
ALL NZ Universities
+
NZ based research and
commercial partners:
 PEBL
 FONTERRA
 TELECOM,
 Middlemore Hospital
 Mighty River Power
 Parrotanalytics
KEDRI’s National and
International Partners
Centre for Novel Methods
of Computational
Intelligence
Centre for
Neurocomputing and
Neuroinformatics
Centre for
Bioinformatics
Centre for Data Mining
and Decision Support
Systems
Centre for Adaptive
Pattern Recognition
Systems
Centre for the Study
of Creativity
 Statistics New Zealand
ASB Bank
Eagle Technology
Auckland Council
Sener International
[email protected]
International collaboration:
 China: China Academy of
Sciences Institute of Automation;
Shanghai JTU and Xinjian U.
 Japan – NiCT- Tokyo; Kobe
University; Kyushu Institute of
Technology; RIKEN
 Germany – TU Kaiserslautern.
 Italy – U. of Trento;
Polytechnico di Milano.
 Switzerland – University of
Zurich and ETH.
 UK – University of
Manchester; U. Lancaster,
University of Ulster, Imperial
College.
Vietnam - VNU
www.kedri.info
Temporal and Spatio/Spectro-Temporal Data Modelling and
Pattern RecognitionRIf STPR problems:
7000
E
Season
Season
Season
Season
6000
–NTRODU
Multiple time series prediction
– (e.g. FONTERRA milk volume prediction)
– Wind energy prediction
– Cyber-security event prediction
– Ecological event prediction
– Environmental event prediction, e.g. earthquakes
– Object movement recognition
– Audio/video data modelling
– Multisensor temporal data integration
– Brain signals (EEG, MEG, fMRI)
– Brain- computer interfaces
– Motor control for prosthetics
– Robot control
– Radio-astronomy data (SKA)
– Mobile-calls prediction (e.g. TELECOM)
– Business data mining
– Decision support systems
[email protected]
1
2
3
4
5000
4000
3000
2000
1000
www.kedri.info
0
0
50
100
150
200
250
The KEDRI Neurocomputing approach
SNNr
eSNN
 Neurofuzzy systems (e.g. DENFIS, 2002)
 Spiking neural networks and reservoir computing
(NeuCube, Kasabov, 2014)
 Adaptive, deep learning of complex spatio-temporal
patterns
 Fast , on-line operation
 Neuromorphic hardware (e.g. U.Manchester SpiNNaker
of 200,000 neuronal processing units available for
KEDRI): high speed and low power consumption
[email protected]
www.kedri.info
pa,b  C  e
 D 2 a ,b / 2
Example: Seismic fast stream modelling for on-line
early earthquake prediction
Actual ->
Predicted
Earthquake
Negative
Earthquake
14
2
Negative
0
32
[email protected]
PwC
PwC
PwC
Analytics at
Victoria University of Wellington
Vicky Mabin and Mark Johnston
NZ Analytics Forum
PwC
May 2014
20
Questions first, then DATA!
•
Think BIG: some key questions:
- What is the problem you are trying to solve, or the goal you are trying to achieve?
- How can we know our customers better?
◦ In an organized world, what would I be able to do?
◦ In an ideal world?
•
How will you achieve this goal?
- Making sense of too much data or complex interactions
◦ Looking for patterns, data mining
- Tailoring real-time actions in response to real-time data
◦ Tracking individuals, targeting messages in real time
•
What’s stopping you from doing this now?
•
Need to do it intelligently
- understand behaviours (e.g. of consumers)
- frame the problem wisely before diving into data!
NZ Analytics Forum
PwC
May 2014
21
Data and Models
Data
What data do you actually need?
Do you have it already or how will you collect it?
•
If you need statistical advice, ask for it sooner (before you collect
any data) rather than later
Models
Still a lot to be done around models and algorithms
There may be no off the shelf solution
Exemplars
Stories from various industries which may be transposed to your context
NZ Analytics Forum
PwC
May 2014
22
Working with Victoria Academics
People throughout university
• MSOR – mathematics, statistics, operations research
• Business school – management, information systems, marketing, tourism, HR,
accounting, …
• Psychology …
Initial contacts
• Vicky Mabin (Victoria Business School)
• Mark Johnston (MSOR)
• Dalice Sim (Statistical consultant)
- Email:
NZ Analytics Forum
PwC
[email protected]
May 2014
23
Analytics at
UC
NZ Analytics Forum
Marco Reale
Where at UC?
 Most
of the courses are offered by the
school of Mathematics and Statistics (MS)
with a broad range in pure and applied
mathematics as well as statistics
 Computational mathematics and
statistics and their applications are strong
both in terms of teaching and research
Where else?
 Other
courses relevant to Analytics are in
Computer Science, Health Sciences,
Psychology, Management Science,
Economics and Finance
New Initiatives
 Computational
applied mathematical
sciences
 Financial Engineering
PwC
PwC
PwC
University of Waikato
Analytics @ Waikato
Analytics related papers @ UoW
• Undergraduate level:
– COMP321: Practical Data Mining
• Co-taught with Statistics Department
• Graduate level:
– COMP521: Machine Learning Algorithms
– COMP523: Data Stream Mining
– COMP555: Bioinformatics
– COMP556: Computational Finance
32
Analytics research and software
• Honour’s, Master, and PhD research
– 4-year BCMS(Hons)
– ~20 finished, 4 on-going analytics-related PhDs
• Graduates both in Academia and Industry:
– LIC, 11Ants, Realtime Genomics, Pingar, Microsoft
Research, …
– Altoo University (Finland), UTS (Sydney), UoW, …
• Numerous open source software packages:
– Weka, Moa, Adams, Meka, Kea, WikipediaMiner
– http://cosi.cms.waikato.ac.nz
33
Weka-centered MOOCs
• Massive Open Online Course:
–
–
–
–
Based on text book by Witten/Frank/Hall
Video lectures, online activities, …
Part I: Data Mining with Weka
Part II: Advanced Data Mining with Weka
• First MOOC in NZ:
– Sept.2013, reruns in March and May 2014
– Next: eResearch NZ 2014, Hamilton, 30/6-02/7
– #WekaMOOC, weka.waikato.ac.nz
34
Analytics at AUT
Paul Cowpertwait
Head of Mathematical Sciences
School of Computer and Mathematical Sciences
School Centres of Research
Related to Analytics




Knowledge Engineering & Discovery Research Institute
Geoinformatics Research Centre
Mathematical Sciences Research Group
Statistical Consultancy Centre
Some Industrial Links:
Relating theory to practice






Statistics New Zealand
ASB Bank
Eagle Technology
Fonterra
Auckland Council
Sener International
Undergraduate Degrees with
Majors in Analytics
Bachelor of Computer and Information Science – BCIS
Bachelor of Mathematical Sciences – BMathSc
Bachelor of Science – BSc
Masters of Analytics (MoA)
 Currently with CUAP
 180 point masters
 Combines papers in statistics, computer science, and applied maths:
• time series
• operations research
• optimisation
• multivariate analysis
• financial maths
• data mining
• data warehousing
• stochastic modelling
Statistics New Zealand
Overview of the Integrated Data
Infrastructure
PwC
42
Integrated Data
Changing the landscape for analytics
Guido Stark and Bex Sullivan
May 2014
Integrated Data Infrastructure (IDI)
44
Consistently maintaining
Privacy
Security
Confidentiality
45
Accessing the IDI
Statistics NZ Microdata Access Services
Public interest
Legal compliance
Bona fide research
Non-regulatory
Proven researcher
Confidentiality
Suitable data source
46
Use of the IDI
Outcomes of tertiary study
Temporary migration
System-wide analytics
47
IDI Expansion
48
Questions
For more information contact
Integrated Data, Statistics NZ
[email protected] [email protected]
Microdata access, Statistics NZ
[email protected]
49
Introducing Data Futures Forum
NZ Analytics Forum
PwC
May 2014
50
Data Futures Forum
[Placeholder Slide]
NZ Analytics Forum
PwC
May 2014
51
Panel Discussion
How can the Government, Industry,
Academia work together to enhance
New Zealand’s analytics capability?
NZ Analytics Forum
PwC
May 2014
52
The value of an idea lies in the using of it…
This publication has been prepared for general guidance on matters of interest only, and does
not constitute professional advice. You should not act upon the information contained in this
publication without obtaining specific professional advice. No representation or warranty
(express or implied) is given as to the accuracy or completeness of the information contained
in this publication, and, to the extent permitted by law, PwC New Zealand, its members,
employees and agents do not accept or assume any liability, responsibility or duty of care for
any consequences of you or anyone else acting, or refraining to act, in reliance on the
information contained in this publication or for any decision based on it.
© 2014 PwC. All rights reserved. In this document, “PwC” refers to PwC New Zealand which is
a member firm of PricewaterhouseCoopers International Limited, each member firm of which is
a separate legal entity.