Big Data in Malaysia: Emerging Sector Profile 2014

Big Data in Malaysia:
Emerging Sector Profile
2014

Big Data in Malaysia: Emerging Sector Profile
ii
Big Data in Malaysia: Emerging Sector Profile 2014
Contact us at sandra@bigdatamalaysia.org, tirath@bigdatamalaysia.org
http://survey.bigdatamalaysia.org
Attribution
Excerpted material from this report can be cited in media coverage and institutional publications. Text
excerpts should be attributed to Sandra Hanchard and Tirath Ramdas. Graphs should be attributed in
a source line to: Big Data Malaysia 2014
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0
International License.

iii

Foreword by
Multimedia
Development
Corporation (MDeC)
Big Data Analytics is a crucial sector in growing Malaysia’s digital
economy. Multimedia Development Corporation (MDeC) is committed
to building a vibrant and dynamic Big Data Analytics industry as a key
component of ICT services in Malaysia. The Malaysian ICT Services
sub-sector has huge potential growth, with a projected share of
48%1
in the nation’s digital economy in 2020. These ICT services sub-
sectors include big data analytics, cloud, mobility, social media and
telecommunications. ICT services encompass software development
and IT services.
MDeC’s strategic intent is for Malaysia to be a hub for Big Data
Analytics services, both as a producer and consumer by 2020. Our goal
is to catalyse the adoption and usage of Big Data Analytics across all
private and public sectors. This reflects the Digital Malaysia initiative
1 IDC and MDeC Analysis
iv

v
which MDeC is driving, to advance the country towards a developed
digital economy by 2020. Digital Malaysia aims to create an ecosystem
that promotes the pervasive use of digital technology in all aspects
of the economy to connect communities globally and interact in real
time. This will ultimately result in increased Gross National Income
(GNI), enhanced productivity and improved standards of living.
TogamechangeMalaysia’sdigitallandscape,lastyearMDeCannounced
the Digital Malaysia 354 (DM354) Roadmap, which encompasses a
plan that addresses three ICT areas: access, adoption and usage across
five key subsectors: ICT Services, eCommerce, ICT Manufacturing, ICT
Trade and ICT Content and Media, targeting four important Digital
Malaysia communities. These communities are the B40 group (the
lower percent of the Malaysian population in terms of household
income), digital entrepreneurs, small to midsized enterprises (SMEs)
and youth, whereby elements of Digital Malaysia’s initiatives will be
proactively embedded in selected programmes and initiatives.
The DM354 initiative will roll out with the ICT services sector which
will see the growth of BDA adoption. With the participation of the
government and private sectors, BDA adoption will not only reduce
costs and raise productivity and competitiveness but also provide a
comprehensive and holistic approach that is needed to grow the Big
Data ecosystem in Malaysia. The DM354 roadmap was approved by the
Prime Minister, Najib Razak last November at the 25th MSC Malaysia
Implementation Council Meeting. This council meeting approved the
development of Malaysia’s Big Data framework, relevant government
pilot projects and funding for the private sector.
MDeC and Big Data Malaysia organisations share common goals such
as improving Big Data literacies in Malaysia and in helping to connect
supply with demand through networking opportunities. MDeC has
worked with Big Data Malaysia during the Big Data Week festivals in
Kuala Lumpur for the anchor events, “Big Data: Strategy, Practice, and
Research” in 2013 and the forthcoming “Big Data: Endless Possibilities”
in 2014. Last year we welcomed diverse participants from engineers to
executives, across all organisational levels and are looking forward to
an even greater response this year. We are delighted to support Big
Data Malaysia’s survey 2014 report, “Big Data in Malaysia: Emerging
Sector Profile”. We commend the goal of the report in raising the
profile of Big Data initiatives in Malaysia and identifying areas of need
and opportunities for growth.
Achieving Digital Malaysia’s 2020 aspirational goals is a challenge
for the whole nation. Issues of data management and governance,
talent development, funding, open data, policies and regulations,
and change management will continue to shape MDeC’s agenda in
supporting the Big Data Analytics industry in Malaysia. We welcome
the sharing of knowledge, successes and learnings in supporting each
other through this journey.
Chief Executive Officer
Datuk Badlisham Ghazali

vi

vii
Table of Contents
Foreword by Multimedia Development Corporation (MDeC) iv
Acknowledgements viii
Introduction 1
Who should read this report 2
Key Findings 3
Seven broad recommendations 5
Survey 6
About the respondents 6
Enabling organisations 8
End uses 10
Skills 12
Capabilities 16
Data sources 18
Profiles 21
Overview 21
Heislyc Loh, CoRate 24
End uses and data sources 24
Malaysian environment and skills recruitment 25
Tom Hogg, Effective Measure 26
Measuring online audiences 27
Regulatory and Malaysian business environment 27
Siim Saarlo, STATSIT 30
Practices 30
Infrastructure 31
Data environments 31
Robin Woo, Western Digital Corporation 32
Scope and context 32
Data governance 33
Resourcing and Malaysian environment 33
Ian Phoon, Malaysian bank 36
Data integrity and prediction 36
Tools of the trade 37
Industry benchmarking 38
Conclusion 39
About the authors 40
Sandra Hanchard 40
Tirath Ramdas 40
About Big Data Malaysia 41

Acknowledgements
The first portion of this report is based on a survey conducted in
October 2013. To encourage survey questionnaire completions, lucky
draw prizes were offered. We would like to thank the sponsors of these
lucky draw prizes:
• Olygen (organisers of the “Big Data World Show” conference
series)
• Merlien Institute (organisers of the “Market Research in the
Mobile World” conference series)
• Revolution Analytics
We would also like to acknowledge our questionnaire testers for
their invaluable feedback prior to the launch of our survey. Our
thanks go to Diana Jayne Gonzales (Big Data Philippines), John Berns
(Big Data Singapore), Ande Gregson (Big Data Week), Mark Smalley
(BrainControl), Daniel Walters (Experian Marketing Services), Jasper
Lim (Merlien Institute), Pak Mei Yuet (Multimedia Development
Corporation), Luke Lai (Olygen), Sivaramakrishnan Narayanan (Qubole),
Andrew Loo Hong Chuang (Taylor’s University), Joethi Sahadevan
(Taylor’s University), Szilard Barany (Teradata), Gopi Kurup (Telekom
Malaysia Research & Development), Syahrizal Salleh (Telekom Malaysia
Research & Development), Margie Ong (Thoughts In Gear), Rei Lim
(Transtel Technology) and Annie Hibberd (VLT).
For assistance with dissemination of our survey questionnaire to
their networks and other forms of support, we would like to thank
the Multimedia Development Corporation (MDeC), the Association
of Banks in Malaysia (ABM), Big Data Singapore, Big Data Philippines,
and EmTech Singapore. We would like to thank all respondents for
contributing their precious time so that we could have the data
needed to conduct the analyses. Furthermore, we would like to thank
respondents who shared with us quotable insights, including Jin Chuan
Tai (ChrysaSys Consulting) and Greg Whalen (Pivotal) and others who
preferred to remain anonymous. The second portion of this report
features profiles of Big Data thought leaders in Malaysia. We are very
grateful to have been able to interview Heislyc Loh (CoRate), Tom
Hogg (Effective Measure), Siim Saarlo (STATSIT), Robin Woo (Western
Digital Corporation) and Ian Phoon (from a Malaysian bank) for their
time. A note of thanks also to Raju Chellam (Dell Enterprise Solutions)
and Elena Tan (JobStreet.com) for participating on a panel dedicated to
this report during Big Data Week 2014.
For assistance with editing this report and providing feedback on our
survey questionnaire, we owe a huge debt of gratitude to Heoh Chin-
Fah (Storage Networking Industry Association Malaysia). Finally, we
thank everyone who has contributed to the development of Big Data
Malaysia. We value all contributions; simply attending a meetup or
posting something interesting to our Facebook wall generates value
for the group.
Sincerely,
Sandra & Tirath
viii

Introduction
Big Data has become part of everyday organisational parlance.
Increasingly, this awareness is being transformed into practice.
Support for harnessing data to amplify capabilities and achieve
organisational objectives at practitioner and senior management
levels is becoming aligned. A report by Forrester2
on Big Data
adoption in Asia Pacific in 2013-2014 observed a trend across all
industries of ‘using more types of data, from more sources, to enable
timelier better-informed insights’. Similarly in Malaysia, there is
growing cultural acceptance that Big Data can and should enhance
decision-making processes, although pathways for adoption are
not uniformly understood. Nevertheless, we are now observing
a transition phase from curiosity and enthusiasm to buy-in and
action across startup, corporate and government organisations.
This groundswell of interest fuels the basis of Big Data Malaysia,
a networking group for professionals with interest in all things
Big Data, including NoSQL, Hadoop, data science, visualisation,
business use cases, data governance, open data, and more. Our
community has welcomed participation from stakeholders ranging
from computer scientists to data journalists, reflecting a broad
societal interest in Big Data. Our mission is to encourage high
caliber knowledge sharing and to provide a space for professionals
with different interests to collaborate. This report grew out of a
need to understand in more detail the various networks that Big
Data Malaysia helps to connect. In order to support the Big Data
ecosystem that we see emerging, we identified critical questions
that required further investigation, in particular:
• What are the opportunities and barriers to Big Data activity in
Malaysia?
• Who is merely ‘interested’, versus who is actually committed?
• What is the current and future capacity for Big Data talent?
• Where are the critical gaps in training and skills?
• What are the soft inhibitors, including data access, regulation
and perception?
There are two parts to this report. The first includes results from
a questionnaire, while the latter features interviews conducted
in-person or via email. In collaboration with various partners, we
devised and distributed a questionnaire online across our networks
and collected responses during October 2013. In our final sample,
we collected responses from 108 individuals over 90 organisations.
As our report will show, these viewpoints represent a diversity of
organisational stakeholders and industries in the Big Data space.
We followed up with interviews of high-profile respondents for
richer insight.
2 Page 2, Big Data Adoption Trends in Asia Pacific: 2013 to 2014, John Brand and Michael Barnes,
Forrester, 2013.
1

Who should read this report
The objective of this report is to raise the profile of Big Data in
Malaysia. We conducted this research because we are passionate
about the potential for Big Data to transform the Malaysian
economy, provided the right conditions prevail to allow the
ecosystem to thrive. If you share this broad view, then this report
may be for you.
Although anyone with an interest in Big Data will find something
of value in this report, it was designed especially with the following
perspectives in mind:
• You are researching how to facilitate engagement with Big
Data initiatives across various organisation levels.
• You are preparing to jumpstart your organisations’ Big Data
initiatives, and are trying to anticipate internal and external
challenges.
• You are considering what use-cases you may target for your
own Big Data Proof of Concept (POC) projects and are trying
to identify what resources you will need.
This report shines a spotlight on the perspectives of senior
management and practitioners (e.g. analysts and engineers),
across producers and consumers in the Big Data ecosystem.
This is the first time we have run the Emerging Sector Profile
survey focused on the Malaysian market. Future goals include
longitudinal and cross-market analysis to provide market
measures and benchmarks. Our scope is limited to presenting a
‘grass roots’ profile of Big Data in Malaysia based on a thriving,
networked community; serving as a foundation to inform readers
of this report who may then go on to formulate recommendations
that are relevant to their organisation.
2

Key Findings
About the respondents
• Interest in Big Data is emerging from both a consumption
and production perspective. Slightly more than 50 percent
of respondents claimed to be in an industry vertical other
than Information and Communications Technology (ICT),
which indicates interest across the broader Malaysian
economy beyond any specific sub-sector.
Enabling organisations
• There is complacency in how managers view the
ability of their organisations to leverage data. Three
out of five managers said either they ‘Agree’ or ‘Strongly
agree’ that their organisation is effectively deriving
tangible benefits from their organisational data assets.
• Anticipated spend for Big Data in 2014 was bullish. Nearly a
third of managers said that they would be increasing their spend
on Big Data in 2014 compared to 2013 by more than 25 percent.
• High outsourcing intent amongst managers in ICT
industries suggests technical fragmentation in Big Data
industries. There are opportunities for boutique firms in
Malaysia to meet specialist technical needs for global markets.
End uses
• Customer-centric uses of Big Data dominates prioritisation.
This suggests that interest in Big Data is primarily driven
by the pursuit of top-line revenue and business growth,
rather than bottom-line factors such as efficiency.
• Managers need to align the perception of the value of Big
Data for both internal and external goals towards a data-
centric organisation. This means recognising both customer-
centric and operational end-uses of Big Data.
Skills
• Differences in desired skills acquisition between ICT and
non-ICT firms are likely to produce a sustainable Big Data
ecosystem. However some organisations may participate in
the Big Data ecosystem as both producers and consumers,
especiallyorganisationsintheICTandmarketingservicessectors.
• Thereisaneedtopromotethevalueoffundamentalapplied
math and statistics skills. These may be overlooked due to the
hype surrounding techniques perceived as more ‘sophisticated’.
3

• Specialised data analysis, modeling, and simulation
are the most commonly desired skill. Intermediate and
advanced respondents rated the following five skills as the
most desirable:
• Big and Distributed Data (eg. Hadoop, MapReduce)
• Algorithms (eg. computational complexity, CS theory)
• Machine Learning (eg. decision trees, neural nets, SVM,
clustering)
• Back-End Programming (eg. JAVA/Rails/Objective C)
• Visualisation (eg. statistical graphics, mapping, web-based
dataviz)
Capabilities
• Uncovering patterns in real-time is a strongly
desired capability. ‘Real time insights from real-
time data streams’ and ‘Uncovering patterns (e.g.
segments, correlations) from multi-structured datasets’
consistently placed in the top three in terms of priority.
• Data visualisation is the number one priority for managers.
Data visualisation is a highly effective way to reveal patterns
and trends that computer systems may not yet be able to
detect in a fully-automated fashion.
Data sources
• Most organisations have a single source of data rather
than combining internal and external sources. There are a
segment of users that would benefit from external consultation
ingettingstartedwiththebasicgoalofidentifyingdatasources.
• Open data from government sources is needed to
support the Big Data ecosystem. Two out of five
respondents said that they either ‘Agree’ or ‘Strongly
agree’ that having access to some government data does
or will create valuable Big Data opportunities for them.
• There is a general weariness by respondents of bureaucracy
andredtapeinachievingtheirBigDataprojects.Aboutonein
fiverespondentssaidtheybelievetheirlocallegalandregulatory
environment to be a hindrance to their Big Data initiatives.
• Public cloud vendors need to identify and address
concerns inhibiting adoption. This includes privacy,
resource cost, and perceptions specific to Malaysian
professionals. Only one in four respondents was willing
to upload internal data to a public cloud service.
4

Profiles
• There are pockets of maturity in the Malaysian Big
Data ecosystem. This includes the use of sophisticated
tools and techniques towards niche, high value use cases.
Areas of expanding interest we profiled included market
research (e.g. brand monitoring, demographic profiling,
and detection of purchasing intent), manufacturing (e.g.
root cause analysis), finance (e.g. risk management),
and many others (including ideas in web applications).
• Malaysians share the global trend of increasing concern to
do with data privacy. The Personal Data Protection Act 2010
(PDPA) has had an awareness-raising effect.
Seven broad recommendations
• Foster a data-centric organisational culture, where
organisational processes, decisions, and objectives are
supported by leveraging a variety of data resources.
• Equip your organisation to manage and make decisions
based on real-time data, in order to remain competitive.
• Cultivate multi-disciplinary data science teams,
consisting of specialists with advanced skills in
mathematics and statistics, domain experts, and others.
• Combine data from internal sources within your organisation
together with external sources (including public and
proprietary third-party sources) to spur innovation.
• Create a ‘data dictionary’ of data assets relevant
to your organisation, to minimise friction for
data science projects across your organisation.
• Large organisations, and especially government, should make
concerted efforts to release data to the public in machine-
consumable form to unlock value hidden within their
data assets.
• The PDPA should be frequently reviewed, and revised if
necessary, to protect individuals while simultaneously
encouraging data projects.
5

Survey
About the respondents
Substantial interest in Big Data from a consumption
perspective as well as a production perspective was indicated
by our respondent pool. Slightly more than 50 percent of
respondents claimed to be in an industry vertical other than
Information and Communications Technology (ICT). While the
ICT sector dominated our respondent pool, interest was indicated
by other major sectors such as marketing services, professional/
scientific/technical services, educational services, and media and
classifieds.
Figure 1: What industry vertical is your organisation primarily in?
n=108
Source: Big Data malaysia
Thesurveycapturedabroadrangeofprofessionalperspectives
at all levels of the organisation. We deliberately did not want
to limit our view to executives or purchasing decision makers
because our goal was to highlight the state of the overall Big Data
ecosystem. Consequently, the respondent pool reflects a balance
between decision makers and a class of respondents that we call
‘practitioners’, who have a ‘hands on’ role with Big Data.
Overall, 24 percent of respondents were C-level executives
(a combination of CEO or equivalent, CIO/CTO or equivalent,
and other C-level), a further 24 percent of respondents were
6

senior/middle management, and 40 percent of respondents
were practitioners (comprising software engineers, analysts,
data scientists, technical consultants, as well as scientists and
academics). It is encouraging that we also received some
responses from full-time students, indicating that awareness of
Big Data has reached into local universities and colleges.
Figure 2: Job title that is the closest fit to your current role:
n=108
Other = 5
Source: Big Data Malaysia
The respondent profile in terms of organisation size was
also reasonably balanced. For respondents who fell into the
‘practitioner’ or ‘senior/middle manager’ role categories, the
organisations they represented ranged in size between very small
organisations (including organisations with less than 5 employees
as well as the self-employed) and large organisations (with over
1000 employees). However, our C-level respondent pool was
heavily skewed towards small and medium sized organisations,
with almost 90 percent of C-level respondents being with
organisations of less than 75 employees.
CEO, or equivalent
CIO/CTO, or equivalent
C-level Other
Senior/middle manager
Software engineer
Analyst
Data scientist
Academic or Scientist
Technical consultant (e.g. pre-sales)
Full time student
14%
6%
4%
24%
16%
11%
6%
5%
4%
7%
7

Enabling organisations
Big Data approaches are becoming infused into organisational
cultures in Malaysia, reflected by resources being dedicated
to Big Data projects. We asked respondents with managerial
responsibilities to provide a ‘temperature’ of their Big Data
readiness and resourcing intent for Big Data in 2014. Overall,
we found that aspirations were high amongst managers, but
actual allocation to personnel was conservative. The results were
nevertheless encouraging in managers matching enthusiasm for
Big Data with resource commitment.
There was some complacency in how respondents viewed
the ability of their organisations to leverage data. Three out
of five managers said either they ‘Agree’ or ‘Strongly agree’ that
their organisation is effectively deriving tangible benefits from
their organisational data assets. Just under a third of managers
were neutral, while just over one in ten managers said they either
‘Disagree’ or ‘Strongly disagree’. However, organisations need to
ensure they are not over-estimating their capacity to leverage
their data. A discussion with Greg Whalen, Pivotal’s Chief Data
Scientist, Asia Pacific and Japan, revealed that many organisations
do not have a ‘data dictionary.’ Data scientists use this ‘inventory’
of enterprise data to identify new correlations that provide return
on investment through creative use of existing information. While
some managers might believe they are already using their data
assets effectively, newer approaches in data mining, exploratory
visualisation and comparison across external data sets may shed
new light on unknown opportunities.
Anticipated spend for Big Data in 2014 was bullish. Nearly a
third of managers said that they would be increasing their spend on
Big Data in 2014 compared to 2013 by more than 25 percent. This
large increase in investment suggests a small base of comparison,
supporting arguments that we are experiencing a turning point
in Big Data in Malaysia from ‘awareness to implementation’. Only
9 percent of managers said they were increasing their Big Data
budget by less than 5 percent. The remaining bulk of managers
(41 percent) said they were increasing their budget from 5 to
25 percent. There was a similar spread in spending intentions
between managers in ICT and non-ICT industries. In terms of
increasing headcount specifically around Big Data initiatives,
most respondents said they expected to grow their staff in the
4-10 range, with few anticipating growth of more than 20 staff in
the coming year.
Create a ‘data
dictionary’ of data
assets relevant to
your organisation,
to minimise friction
for data science
projects across your
organisation.
8

Figure 3: How do you expect your spend on Big Data to change
in 2014 compared to 2013?
n=75
Don’t know / prefer not to say = 16
High outsourcing intent amongst managers in ICT industries
suggested technical fragmentation in Big Data industries.
While 48 percent of non-ICT managers said they were either
moderately or extremely willing to outsource high-skills tasks
in their Big Data initiatives to external consultants, there was
greater willingness in ICT industries, with a comparative number
of managers accounting for 69 percent. This suggests increasing
technical fragmentation in Big Data. There are opportunities for
boutique firms in Malaysia to meet specialist technical needs.
Local firms can nurture local talent and export to global ICT
markets
Increasing by more than 25%
Increasing by between 10% and 25%
Increasin by between 5% and 10%
No change / Increasin by less than 5%
28%
17%
24%
9%
9

End uses
Applications of Big Data may be relevant to every organisational
process and objective, whether internally or externally focused.
While some uses of Big Data are clearly understood and valued,
others are still emerging. Greg Whalen, of Big Data platform
vendor, Pivotal, shared with us successful use cases ranging
from ‘identifying drivers of manufacturing quality to detecting
malware outbreaks to predicting customer churn to identifying
at-risk patients.’ While applications might be relevant to specific
domains, sharing case studies across industries can highlight
creative processes in extending data analysis. Ultimately, there is
a need by organisations to determine how Big Data strategies
align with their organisational values and objectives.
Customer-centric uses of Big Data dominated prioritisation
by respondents. Customer behavioural profiling was the
most highly rated end-use in terms of relevance. Improving
customer service and experience, retention, acquisition, cross-
sell and upsell were also identified as important uses of Big Data.
Respondents also valued Big Data for ‘competitive intelligence’ in
benchmarking against their industry. The high ranking of social
trends monitoring highlighted the general appeal of social data.
Of less relevance to respondents were uses rated to compliance
and regulatory issues, infrastructure and assets monitoring,
supply-chain monitoring, risk management and operational cost
management. This suggests that currently, interest in Big Data is
primarily driven by the pursuit of top-line revenue and business
growth, rather than bottom-line factors such as efficiency.
Figure 4: How relevant are the following end-uses for Big
Data capabilities from your own perspective?
There were differences in prioritisation between respondents
in ICT and non-ICT industries. Non-ICT respondents were more
likely to value social trends and brand monitoring, Conversely,
ICT respondents exhibited a swing towards internal operational
end-uses. The greatest differences between the two groups of
respondents were in infrastructure and assets monitoring, risk
management and supply-chain monitoring. ICT respondents also
Foster a data-centric
organisational culture,
where organisational
processes, decisions,
and objectives
are supported by
leveraging a variety of
data resources.
Customer behavioural proﬁling Very relevant 103
Customer service and/or experience 101
Competitive intelligence 105
Customer retention 104
Social trends monitoring 108
Customer acquisition 117
Customer cross-sell and/or up-sell 104
Forecasting supply and demand 110
Brand monitoring 109
Product and service innovation Moderately relevant 115
Operational cost management 115
Risk management 135
Supply-chain monitoring Slightly relevant 132
Infrastructure and assets monitoring 158
Compliance and regulatory issues Not all relevant 111
n=108 n=56 n=52
End-uses ranked by priority of relevance Industry swing by All
10

gave a higher prioritisation to product and service innovation.
There is a need for non-technical businesses to appreciate
how data-driven feedback can improve products and services
for competitive advantage. Differences in prioritisation should
prompt stakeholders formulating uses cases of Big Data to
consider whether they are too internally or externally focused;
there is an opportunity to think outside-the-box for applications
and to learn from other industries. Big Data vendors may choose
to tailor products to specialised organisational functions.
Managers need to align the perception of the value of Big Data
for both internal and external goals towards a data-centric
organisation. For example, data-driven approaches in change
management, product and service development and customer
building may be complemented with data-driven approaches
of increasing operational efficiency. In their perception of Big
Data uses, managers were more likely than practitioners to value
risk management, customer retention, customer acquisition,
product and service innovation, customer behavioural profiling
and forecasting supply and demand. Practitioners had a stronger
internal operational focus, being more likely to prioritise supply-
chain monitoring and infrastructure and assets monitoring. The
mantra of reducing cost while increasing value should be at the
core of applying Big Data in organisations.
11

Skills
One of the most frequently cited obstacles to delivering Big Data
initiatives is the difficulty in recruiting the required skills. We
presented respondents with a list of skills3
and asked them to
score each skill according to need from their perspective. Each
response scored on a 5 point scale, ranging from “No need” to
“Critical need”, then scores aggregated to produce the following
heatmap:
Figure 5: To deliver your Big Data initiatives, how much need
is there to recruit the following skills?
The most commonly desired skill across the combined
respondent pool was specialised data analysis, modeling, and
simulation. This was followed by distributed systems, deployment
and / or administration, and fundamental computer science and
software engineering. There was consensus across segments that
pure business skills (distinct from domain knowledge) was of low
importance, but the least relevant skillset was hardware/sensor
design. We included this skillset in our list as a proxy for the
“Internet-of-Things” concept that is often seen as closely related
to Big Data.
3 Our list of skills was derived in part from the Skills List presented in “Analyzing the Analyzers: An
Introspective Survey of Data Scientists and Their Work” by Harris, Murphy, and Vaisman, 2013. We
grouped together several related items from their list into composite items in our list.
Cultivate multi-
disciplinary data
science teams,
consisting of specialists
with advanced skills
in mathematics and
statistics, domain
experts, and others.
ICT only All respondents
Non-ICT
only
Specialised data analysis, modeling, and simulation
(e.g. operations research, machine learning, etc.)
3 Priority 1 Priority 1
Distributed systems (e.g. Hadoop) deployment and/or
administration
Priority 1 2 2
Fundamental computer science and/or software
engineering
4 3 3
Industry-specific/domain knowledge 2 4 4
Applied math and/or statistics 5 5 6
Web/mobile development and/or visualization 7 6 5
Research experience from any quantitative discipline 6 7 7
Business (strategy, marketing, product development,
etc.)
8 8 8
Hardware/sensor design 9 9 9
n=52 n=108 n=56
12

There is nuance in skills priorities between ICT and other
industries. The overall respondent pool cited specialised data
analysis, modeling, and simulation as their most pressing need,
with distributed systems (e.g. Hadoop) coming in a close second.
However, when segmenting by industry (ICT vs all other industries),
we see skews in preferences:
• Specialised data analysis, modeling, and simulation:
• 1st priority by non-ICT respondents.
• 3rd priority by ICT respondents.
• Distributed systems deployment and/or administration:
• Distant 2nd priority by non-ICT respondents.
• 1st priority by ICT respondents.
• Industry-specific/domain knowledge:
• 4th priority by non-ICT respondents.
• 2nd priority by ICT respondents.
The difference in domain knowledge may be expected (in
particular if ICT respondents aim to provide services to non-
ICT consumers, in which case they would need to acquire
target domain knowledge to provide adequate service), but the
difference in distributed systems versus specialised data analysis
may point to different aspirations by the respondent groups.
Differences in skills acquisition between ICT and non-ICT firms
is likely to produce co-dependencies that shape a sustainable
Big Data ecosystem. Broadly speaking, if we may assume that a
significant portion of ICT respondents are participating in the Big
Data ecosystem as solutions providers to non-ICT respondents and
that only a small portion of non-ICT respondents are participating
in the Big Data ecosystems as providers, then our findings may
indicate the emergence of a model of engagement with ICT players
providing Big Data infrastructure capabilities and solutions, with
some degree of industry customisation. Meanwhile, non-ICT
consumers are developing deep analytical capabilities pertinent
to their business needs on top of the provided infrastructure,
notwithstanding the role of small boutique ICT and analytics
consultancies that can fill gaps wherever opportunities arise.
Some organisations may participate in the Big Data ecosystem
as both producers and consumers, especially organisations in the
ICT and marketing services sectors.
There is a need to promote the value of applied math and
statistics skills for Big Data analytics. It is noteworthy – and
likely a point for caution – that respondents rated specialised
data analysis skills so highly but were lukewarm to applied math
or statistics skills. It is possible that the respondent pool have
already developed significant practical expertise in applied math
and statistics, but another likely scenario is that there exists a
lack of understanding of the utility and value of applied math
13

and statistics to Big Data analytics. The hype around more
novel approaches such as machine learning may be attracting
disproportionate mindshare over applied math and statistics,
which respondents may have perceived as “basic”4
.
The breadth of our respondent pool gives rise to a big range
in technical knowledge, which might lead to concerns that
some respondents may have not had sufficient understanding
to properly assess their need for some specific technical skills.
To account for this possibility, we followed up with additional
data collection. Firstly, we asked respondents, “How would you
assess your own degree of technical understanding of Big Data?”5
.
Respondents chose one of the following options: ‘Total beginner’,
‘Technical details are not relevant to my job role”, ‘Intermediate’,
and ‘Advanced.’ Respondents who answered ‘Intermediate’ or
‘Advanced’ (56 percent of our total respondent pool) were then
presented with an expanded skills list6
and asked to select only
their first, second, and third skills priorities. The responses for
each skill were weighted by priority and the resulting scores
aggregated. These results from skilled respondents reinforce our
earlier findings:
• ICT respondents are skewed towards distributed systems skills.
• Non-ICT respondents are skewed towards advanced
algorithms.
• There is lukewarm interest in math and statistical methods.
An additional finding is that despite the overall lukewarm interest
in statistical methods, there is one bright spot: temporal statistics.
This points to interest in time series analysis, which is an important
method applicable to many areas.
Despite differences between ICT and non-ICT perspectives
amongst intermediate and advanced respondents, it is clear
that there is significant shared interest in the top five skills:
• Big and Distributed Data (eg. Hadoop, MapReduce)
• Algorithms (eg. computational complexity, CS theory)
• Machine Learning (eg. decision trees, neural nets, SVM,
clustering)
• Back-End Programming (eg. JAVA/Rails/Objective C)
• Visualisation (eg. statistical graphics, mapping, web-based
dataviz)
Targetedtrainingintheseareas,especiallyifofferedincombination
with each other, is likely to see widespread interest and provide
4 In the “Profiles” section of this report, we will show that several thought leaders suggest that
skills such as statistics and a scientific background are of significant value to Big Data initiatives,
contradicting respondent perceptions.
5 This self-rating question only served as a basis to drill down into follow-up questioning; in no way
do we suggest it may be used as an accurate gauge of the extent of Big Data technical knowledge in
Malaysia.
6 Here we used the full Skills List from “Analyzing the Analyzers: An Introspective Survey of Data
Scientists and Their Work” by Harris, Murphy, and Vaisman, 2013.
14

a major boost to the development of the Big Data ecosystem in
Malaysia.
Figure 6: If you (or your employee/service provider) could
only develop one skill, what should it be?
Respondents presented with a skills list and asked to select their
first priority, second priority, and third priority only. Responses for
each skill scored according to priority, then aggregated.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Non-ICT (n=26)
Combined (n=61)
ICT (n=35)
Big and Distributed Data
Algorithms
Machine Learning
Back-End Programming
Visualization
Temporal Statistics
Data Manipulation
Graphical Models
Surveys and Marketing
Product Development
Front-End Programming
Unstructured Data
Systems Administration
Structured Data
Busines
Spatial Statistics
Classical Statistics
Simulation
Math
Bayesian/Monte-Carlo Statistics
Science
Optimization
Normalized priority
15

Capabilities
This section highlights specific capabilities that respondents seek
in their Big Data solutions. Capabilities are desired characteristics
and features that may cut across industry segments, be applicable
to multiple end-uses and/or rely on some combination of skills
to produce and consume. As such, we investigate capabilities
as a separate, though not necessarily independent issue. Some
capabilities are specific to certain types of data, but we still
regard these as unique capabilities rather than data source issues
because simply having access to the data does not produce
desired outcomes – it is the handling of the data that creates
value.
A meaningful opinion on this issue requires some degree of
technical knowledge, so we only included respondents who
answered either ‘Intermediate’ or ‘Advanced’ to the question, “How
would you assess your own degree of technical understanding of Big
Data?7
” (56 percent of our respondent pool). In addition to drilling
into ICT vs other industry segments, here we also investigate the
set of respondents who are “Managers”, i.e. respondents who
answered ‘Yes’ to the question, ‘Does your role in your organisation
include managerial/leadership responsibilities?’ (69 percent of our
respondentpool).Thesubsetofourrespondentpoolthatcomprise
managers with at least ‘Intermediate’ technical understanding of
Big Data make up 43 percent of our total respondent pool.
Uncovering patterns in real-time is a strongly desired
capability. Despite differences in perspective in the different
segments studied, ‘Real time insights from real-time data
streams’ and ‘Uncovering patterns (e.g. segments, correlations)
from multi-structured datasets’ consistently placed in the top 3
in terms of priority. These priorities were especially strong for
ICT respondents, who went on to indicate ‘Data discovery and
exploration across many data sources’ as their third priority. Taken
in combination, these are capabilities that are especially related
to Big Data infrastructure, once again illustrating an apparent
inclination for ICT respondents to focus on supplying capabilities
for external consumption, rather than for their own internal
consumption.
7 Respondents choose one of the following options: ‘Total beginner”, ‘Technical details are not relevant
to my job role”, ‘Intermediate”, and ‘Advanced”.
Equip your organisation
to manage and make
decisions based on real-
time data, in order to
remain competitive.
16

Figure 7: Which of the following Big Data capabilities are
relevant to you?
Note: Each response scored on a 5 point scale, ranging from “Not at all relevant”
to “Very relevant”, then scores aggregated to produce the above heatmap.
Datavisualisationisthenumberonepriorityformanagers.This
capability was also strongly desired by non-ICT respondents. This
points to a consumption need; non-ICT and especially manager
respondents may prioritise Big Data capabilities to support their
decision making processes, and visualisation of data is a highly
effective way to reveal patterns and trends that computer systems
may not yet be able to detect in a fully-automated fashion. In the
overall respondent pool, visualisation was only the third priority
because it was pulled down by ICT respondents, who only placed
it as the fourth priority.
Once again, there is no interest in “Internet of Things” (IoT).
Despite the opportunities for linking the two worlds (with IoT
serving as a data source and Big Data producing value out of the
raw data) it seems these fields, which are still relatively nascent in
Malaysia, are yet to instigate much interest in combination. Even
within the ICT segment of our respondent pool, there is virtually
no interest in IoT, which suggests that even the supply-side is not
yet ready to deliver these capabilities to the Malaysian market. As
the Big Data and IoT ecosystems continue to evolve in Malaysia
we may see this lack of interest give way to interesting startup
initiatives. In the meantime, awareness raising initiatives should
focus on this link.
ICT only All respondents Non-ICT only
Managers
only
Real-time insights from real-time data streams 2 Priority 1 Priority 1 3
Uncovering patterns (e.g. segments, correlations) from
multi-structured datasets
Priority 1 2 3 2
Visualizing/presenting insights 4 3 2 Priority 1
Data discovery and exploration across many data
sources
3 4 4 4
Statistical analysis on big working data sets (>100GB) 5 5 5 5
Automated decision making 7 6 7 6
Machine-generated data (e.g. log files, periodic
diagnostics)
6 7 9 8
Content and sentiment from online media (e.g. social
media)
9 8 6 9
Efficiently and safely storing large data sets on
infrastructure controlled by my organization
8 9 10 7
Image, video, and audio data 10 10 8 10
Physical sensor networks (e.g. "Internet of Things") 11 11 11 11
n=35 n=61 n=26 n=46
17

Data sources
Data is the new oil. Dimensions of data quality and value are
increasingly as important as volume, velocity and variety. The wider
social and business impact of Big Data initiatives is dependent
on the quality of data and information on which they are based.
Data environments whether open or proprietary are becoming
increasingly complex as new business models emerge. Demand
for open data for both private and public sector use is increasing
for endeavours such as data journalism. In Malaysia, complex
social problems such as traffic congestion or local crime could be
addressed by open data and crowdsourcing approaches, in both
diagnosing problems and in providing feedback mechanisms
from citizens. There are many issues related to data sources that
we identified during our survey. These include use of internal
and external data, access to open government data, regulatory
issues and appetite for using third-party infrastructure (e.g. cloud
services) for data management. The sourcing and management
of data requires consideration of complex relationships and
issues, including strategic partnerships, security, and distribution,
to name a few.
Most organisations have a single source of data rather than
combining internal and external sources. ICT respondents were
more likely than non-ICT respondents to use open access (e.g.
government) and proprietary third party data as the sole source
for Big Data initiatives. Non-ICT respondents were polarised in
terms of data resourcing where 43 percent said they were using a
combination of sources, while 17 percent of respondents were not
using any data for Big Data initiatives. There are a segment of users
that would benefit from external consultation in getting started
with the basic goal of identifying data sources. Respondents who
were likely to highly-value customer behavioural profiling were
dependent on third party data such as social media.
Figure 8: My organisation uses sources for Big Data initiatives
primarily from the following:
‘Don’t know/prefer not to say’ = 16
Combine data from
internal sources within
your organisation
together with external
sources (including
public and proprietary
third-party sources) to
spur innovation.
17%
7%
28%
27%
6%
20%
6%
11%
43%
36%
Non-ICT
(n=47)
ICT (n=45)
None yet
Internal data
Open-access third-party data (incl.
government)
Proprietary third-party data
Combination
Large organisations
should make concerted
efforts to release data
to the public in
machine-consumable
form to unlock value
hidden within their
data assets.
18

Open data from government sources is needed to support
the Big Data ecosystem. Two out of five respondents said
that they either ‘Agree’ or ‘Strongly agree’ that having access
to some government data does or will create valuable Big Data
opportunities for them. Just under half of respondents were
neutral, while 1 in 10 respondents said they either ‘Disagree’ or
‘Strongly disagree’. The large number of neutral respondents
suggests more education is required around how open data can
be used for Big Data projects. Respondents identified a number
of uses of government data. Topics that were frequently cited
included:
• Demographic and socioeconomic data
• Population (online and offline)
• Crime; Border security and migration
• Public and community services (utilities, health, education)
• Location (by utility); Geographic Information Systems (GIS)
• Financial; credit
• Weather
By far the greatest need was demographic and socio-economic
data. One respondent, head of decision science at a multinational
bank, said that “In order for us to understand the needs of
Malaysians, statistical data from the population census is important
to identify correlations with internal behavioural data.” There is
a need for government agencies to share data multilaterally for
social problems that have many dimensions. For example, access
to health care is also a function of transportation, location and
income. Ministries can utilise internal data management for service
optimisation. Jin Chuan Tai, Director of ChrysaSys Consulting said
that “Government has many different sets of survey data, collected
from various sources. For instance, Ministry of Health data can be
sourced from private and general hospitals, clinics or consultancies.
Big Data offers a mechanism to speed up consolidation of all this
information, without any processing delays to configure each and
every source.”
There was a general weariness by respondents of bureaucracy
and red tape in achieving their Big Data projects. About one in
five respondents said they believe their local legal and regulatory
environment to be a hindrance to their Big Data initiatives,
compared to 15 percent who did not. Most respondents (47
percent) were neutral. A further 15 percent said ‘don’t know’ clearly
identifying need for further education on emerging regulatory
trends in the era of Big Data. The most common hindrance cited
was the Personal Data Protection Act 2010 (PDPA), gazetted in
Malaysia in late 20138
. This was a common response highlighting
the need for balancing consumer protection with business needs,
and better education specifically on the PDPA. Other issues of
concern to respondents included data compliance, data risk and
loss of data.
8 The hindrance posed by the PDPA on Big Data initiatives may be more in terms of perception, rather
than actual legal definition. In the “Profiles” section of this report, we will show that one industry
thought leader believes that the PDPA is actually beneficial, by providing reassurance to the public that
their data is protected with a legal framework in place.
The PDPA should be
frequently reviewed,
and revised if
necessary, to protect
individuals while
simultaneously
encouraging data
projects.
19

Public cloud vendors need to identify and address concerns –
privacy, resource cost, and perceptions specific to Malaysian
professionals. Only one in four respondents was willing to
upload internal data to a public cloud service. While 37 percent
of respondents said they either ‘Disagree’ or ‘Strongly disagree’
with the statement, ‘I am willing to upload my internal data to
third-party infrastructure (e.g. public cloud’), just over a third
of respondents were neutral, indicating that there are a large
number of fence-sitters that have yet to be swayed either way.
Respondents who were more likely to invest in Big Data initiatives
were also much more likely to be willing to use public cloud
services, suggesting that maturity around data use and Big Data
literacy reduces the anxiety around uploading data to third-party
infrastructure.
Figure 9: I am willing to upload my internal data to third-
party infrastructure (e.g. a public cloud)
0%
14%
23%
37%
20%
6%
50%
100%
%ofrespondentsn=108
Agree
Neutral
Disagree
Strongly agree
Strongly disagree
20

Profiles
Overview
The previous section presented an aggregate view of the state of
Big Data activity in Malaysia. In this section we present a closer
look, from the perspective of five industry leaders:
• Heislyc Loh, Founder and Chief Executive Officer, CoRate
• Tom Hogg, Commercial Director, Effective Measure
• Siim Saarlo, Chief Technology Officer, STATSIT
• Robin Woo, Senior IT Manager, Western Digital Corporation
• Ian Phoon, Data Scientist, a Malaysian bank
Our objective in selecting these interviewees was to strike a
balance in backgrounds in end uses and to highlight diverse
industries. We have representatives from finance, information and
communication technology, manufacturing, marketing and social
media; across roles in Commercial, Digital Marketing and Product
departments. The areas of activity and expanding interest in Big
Data analytics we profiled included market research (e.g. brand
monitoring, demographic profiling, and detection of purchasing
intent), manufacturing (e.g. root cause analysis), finance (e.g.
risk management), and many others (including ideas in web
applications). Some of these use cases are niche interests and
specific to certain industries, but may nevertheless be of high
value and high impact.
The initiatives described by our interviewees point to the existence
of pockets of maturity in the Malaysian Big Data ecosystem.
Sophisticated tools and techniques – such as the application of
natural language processing, machine learning, full text search
and analysis, statistics, social science, data visualisation, and data
governance as well as general data quality measures – are being
applied towards tasks such as semantic analysis, defect root-cause-
analysis, financial early-warning systems, and general predictive
analysis solutions. In some situations there is already recognition
of the need for specific Big Data return on investment measures.
One respondent referred to the conceptual idea of a ‘value to bit’
ratio, which attempts to capture the objective of increasing the
value of data assets, while reducing costs associated with data
collection, retention, and management.
The organisations represented here are leveraging a wide variety
of data sources, including social media, website traffic and browser
interactions, traditional surveys, external third-party media (such
as news reports), and others. While in most cases many options for
data collection exist, several respondents expressed the view that
the Malaysian government should do better in terms of releasing
government data, for instance by making available regular census
information through promotion to organisations that would be
benefit from this data.
21

Beyond opening up government data, our respondents raised
several other points about the prospects of Big Data analytics
as a significant part of the Malaysian economy. Firstly, the issue
of talent availability was highlighted. Exacerbating this issue,
there is a perception that relevant training is relatively lacking in
Malaysia (e.g. as compared to Singapore). Until a local pool of
skilled specialists can be grown, it may continue to be necessary
to outsource the highest-value tasks to specialists elsewhere.
There are different opinions on the precise requirements and
priorities for data scientists, but a strong working knowledge of
statistics seems to be on everyone’s wish list. From the Big Data
consumption perspective, it is pertinent that the issue of privacy is
receiving significant interest in Malaysia. There is a global trend of
increasing concern to do with data privacy, and our respondents
indicated that Malaysians share these concerns. The Personal
Data Protection Act 2010 (PDPA), beyond its legal definition, has
had an awareness-raising effect.
As for issues unique to startup ventures, despite a favourable
view of Big Data initiatives by funding organisations, the usual
factors impacting all IT startups in Malaysia (in particular access
to post-seed-level-funding, and exit opportunities) remain. These
issues are made more acute by the longer runway required to
achieve sufficient scale – to produce sufficient data – such that
Big Data analytics can be meaningfully exploited. Some of our
respondents here share views based on substantial regional or
global experience as a point of comparison to their successes and
challenges in the local landscape. All are dedicated to seeing Big
Data succeed in Malaysia.
22

23

Heislyc Loh,
CoRate
CoRate is an early stage startup that aims to help people consume
information on the Internet more effectively, by providing a
smart article collection platform - like pocket with automatic
categorisation, organisation and summarisation capabilities.
These functions will be developed through a combination of
crowd sourcing and data analytics. CoRate founder, Heislyc Loh is
confident that their vision will be widely accepted by people who
need to deal with large amounts of information, ranging from
business analysts and researchers to students. He calls this target
group of people, ‘infoholics.’
End uses and data sources
Going beyond social bookmarking, the platform is an ambitious
yet classic example of Big Data, with a roadmap to utilise natural
language processing, machine learning, and other elements. It is
envisioned that the platform will leverage three major data source
categories: data and metadata within Internet articles themselves,
data captured on user interaction with the article (e.g. browser
extensions will allow users to highlight interesting passages), and
interactions and relationships between users (i.e. social data).
“We’re confident our
vision will be widely
accepted by people
who need to deal
with large amounts of
information.”
24

Loh recognises that this vision will be challenging to realise in the
best of circumstances, but is nevertheless optimistic. Since late
2013 a very early-stage piece of the product has been released in
the form of a Chrome web browser extension, but by-and-large
development work only commenced at the start of 2014, and
building-up of a team in Kuala Lumpur is ongoing. The current
team comprises mostly front-end developers with an emphasis
on User Experience (UX) as CoRate tries to achieve end-user
traction. At this early stage their technology stack consist of web
app mainstays MongoDB, Ruby-on-Rails and Javascript.
Malaysian environment and skills
recruitment
Loh believes that funding in Malaysia is not an issue at the early
stages, in part due to the availability of grants from various
government agencies, and Big Data initiatives in particular
are viewed favourably. However, in the medium-to-long-term
CoRate has its sights on Silicon Valley. This is primarily because
despite the availability of early stage funding, there is very little
high-capital investment (e.g. Series B and beyond) available in
Malaysia. Furthermore, given CoRate’s consumer-Internet focus,
exit opportunities are relatively thin outside of Silicon Valley.
He goes on to acknowledge that beyond the initial phases,
realising CoRate’s vision will require high-end specialised skills,
and finding individuals in Malaysia with relevant experience would
prove challenging. Nevertheless, he believes that in the long term
these capabilities can be developed at home because “there
are smart people here”. To jumpstart the development process
he plans to recruit specialists from overseas (primarily from San
Fransisco, where he spent a few weeks in late 2013 developing
the CoRate concept) with the needed skills.
25

Tom Hogg,
Effective Measure
Tom Hogg is the Commercial Director for Effective Measure
in Malaysia. Effective Measure is a leading provider of digital
audience, brand and advertising effectiveness measurement and
targeting solutions, bringing best practice online measurement
data to premium publishers, agencies, networks, advertisers
and researchers. Effective Measure’s client base in primarily
made up of publishers, agencies and networks, and they are
increasingly working with advertisers and insight agencies. They
are headquartered in Melbourne, Australia where they were
founded and developed their proprietary technology, which now
provides online measurement solutions globally. In South East
Asia, Effective Measure operates in Malaysia, Singapore, Thailand,
Vietnam, Philippines, Hong Kong and Indonesia. Top advertising
clients globally, based on digital spend, are typically in finance,
automotive, FMCG (food and beverage, homecare, health and
beauty), as well as travel and hospitality. In Malaysia, Effective
Measure covers website traffic from 38 million unique browsers
each month on its network (including users with multiple devices).
They currently offer 70,000 demographic profiles over the online
population nearing 20 million people in Malaysia.
“We need to showcase
the power of the local
audience... There’s a
place for everybody,
and room for everyone
to operate”
26

Measuring online audiences
In terms of the Big Data ecosystem in Malaysia, Effective Measure
provides detailed demographic and psychographic insights into
website audiences, reflected by Malaysian internet usage at a
census level. One of their core objectives is to help publishers
and networks validate who their audiences are, so they can
pitch to agencies and advertisers about the kinds of people on
their website. A main challenge for local publishers in markets
like Malaysia is the competitive influence of global operations,
who have large audience reach. Hogg emphasises that it is very
important to “help local publishers have a trusted and transparent
source of data that can be shared with agencies and advertisers.
We need to showcase the power of the local audience, in addition
to the audiences visiting these multinational giants. There’s a place
for everybody, and room for everyone to operate; the concern for
local publishers is that more and more of the AdEx [advertising
spend] is leaving local markets, with the risk that local publishers
lose their revenues, which ultimately means that local content is
non-sustainable.”
There have been vast changes in media measurement with the
advent of Big Data. Hogg notes that online environments offer
easier access to much larger data sets, whereas previously
advertising decisions were made on small subsets of the
population. An important development in online measurement
is the ability to compare the impact of advertising on members
of a population who have been ‘exposed’ to a digital campaign
versus members who have not. In post-campaign studies, clients
can link impacts back to core objectives such as consumer brand
advocacy, uplift in brand awareness, purchase intent and a range
of insights based on different business outcomes. Because more
advertising spend is going online, validation is required by third-
party online measurement providers to justify budgets.
Effective Measure presents aggregated, anonymous behavioural
data based on groups of people through a dashboard client
service. Hogg observes that most people “aren’t experts in
statistics, so it’s really important to make the data easy for people
to understand”. Data visualisation is increasingly important
in how clients consume data. The evolution of infographics
in Big Data follows general consumption trends for media i.e.
interactive, visually stimulating and appealing. In Malaysia, Hogg
has observed an increase in requests by clients for key summary
points to be pulled out in addition to their dashboard service.
Regulatory and Malaysian business
environment
Hogg says that Effective Measure has observed an increase in
sensitivity by Malaysian respondents to some types of survey
questions. Hogg believes this is part of a global trend of
27

increasing consumer concern with data privacy. In addition, the
crafting of the Personal Data Protection Act (PDPA) in Malaysia
has heightened awareness around data privacy; however, Hogg
believes the PDPA is likely to be ultimately beneficial in helping
people to understand that their data is protected with a legal
framework in place. Government can support the Big Data industry
in Malaysia by providing regular census information, insights
into online behaviour and information on infrastructure for the
purposes of corroboration. The more government data that can
be accessed, the better for the entire digital community. In terms
of comparing the data analytics market in Malaysia versus the rest
of South East Asia, Hogg doesn’t see a tremendous difference,
with the exception of Singapore. Hogg observes an equal interest
in data analytics in Malaysia as in other markets; however, there is
a range of data literacy within Malaysia. Education is required on
understanding statistical methodologies in how panels are used
to estimate population sizes; this can help both publishers and
advertisers understand Big Data methods better.
28

29

Siim Saarlo,
STATSIT
Siim Saarlo is the Chief Technology Officer of STATSIT, a social
media marketing technology company. STATSIT works with brands
and agencies in the Asia Pacific region to help them increase
advocacy and sales, to identify influences and to prove the value
of social media activities. In general, STATISIT extracts insights
from social media conversations. They define relevant subsets of
conversations and conversationalists for customers and projects.
Over the resulting stream of social data, they perform analysis
and produce insights in various forms. This includes basic volume
monitoring, tailored reports and action steps, influencer profiling
and predictions on content demand.
Practices
STATSIT relies on data science practices and the statistical
capabilities of social scientists, as well as machine learning
algorithms. New approaches and algorithms get tested first on
internal and one-off projects before reaching end products. Due
to their normal subject matter, STATSIT utilise natural language
processing (NLP) to some extent. However, STATSIT avoid more
complex NLP algorithms because these are usually language
“There is hardly a
profession in IT that
would not benefit from
some data awareness.”
30

specific while their main products cover a variety of languages.
Their methodologies for extracting insights from social data
depend on project goals. They rely on both standardised
quantitative elements and some qualitative analysis leading
to standalone products e.g. influencers discovery and analysis.
Qualitative analysis mainly relies on analyst work and is often
iterative, but is supported with different analytical tools, such as
full text search, entity and pattern detection etc.
Infrastructure
Saarlo believes that organisations have an opportunity to optimise
their resource usage in terms of infrastructure. He comments that
riding the big data wave often makes organisations collect more
than they analyse, leaving them with a lower ‘value to bit ratio’.
Recognising this challenge, STATISIT are looking for options to
change their data collection, processing and usage processes to
cut down costs for infrastructure and at the same time increase
the value of generated insights. This is not so much with the
aim for cutting cost, but to focus more on value creation before
data collection. This might mean reviewing their agreements
with infrastructure providers, employing new and more effective
technologies, being smart in reducing the amount of ‘active data’
and focusing analytical efforts etc. In summary, their objective is
to be more effective and smart with both technology and data
usage.
Data environments
STATSIT works mostly with data that is public and available for
free. From the perspective of a product developer, an ongoing
challenge is when platforms restrict access to data as business
models mature. Saarlo believes that general data aggregator
companies will get more dominant over time, and that data
collection and data analysis will become increasingly siloed
activities. Saarlo predicts greater specialisation in big data with
more companies tackling very specific domain problems in niche
fields. Additionally, he sees an increasing utilisation of open
source components versus approaches that port general machine
learning algorithms into completely proprietary solution stacks.
In order to support data intensive endeavours, Saarlo believes
that governments and organisations should make data available
as much as reasonable, keeping in mind an audience of curious
and innovative developers. He says that as data science is fairly
well popularised in professional circles, students should also
get taste of the hype, which would naturally increase interest in
subjects like maths and statistics. Saarlo says that there is “hardly
a profession in IT field that would not benefit from some data
awareness”.
31

Robin Woo,
Western Digital
Corporation
Western Digital Corporation is one of the largest computer
storage manufacturers in the world, and is one of the most
important manufacturers in Malaysia. Robin Woo, Senior IT
Manager at Western Digital Malaysia, is responsible for leading
datacentre operations management and systems administrations
at all manufacturing sites worldwide. He is also responsible for
leading web development for general IT applications globally,
and for leading the company’s IT roadmap strategy globally.
Scope and context
In Woo’s view, Big Data and Analytics are inseparable because
ultimately it is what you do with the data that counts. In his view,
the end-uses are essentially unchanged from the traditional
goals of analytics, which includes correlation, root-cause analysis,
predictive simulation, and triggers to name a few. Nevertheless,
there are some defining characteristics that distinguish the Big
Data challenge.
“One must know
how to ask the right
questions.”
32

First of all, the use cases have evolved to include unstructured
data, though still leveraging structured data in increasingly
complex ways. In this way, Big Data can provide greater richness
to enhance analytical capabilities. In addition, it can also provide
more timely results. Secondly, a major change is the computer
systems involved; dealing with a quantum of data several
times larger than traditional systems have been dealing with
necessitates the use of systems designed with parallelism as a
core engineering principle, which is why traditional solutions may
not always scale effectively. The key to success, however, may
be somewhat removed from core engineering and data science
algorithms. From Western Digital’s experience, the key piece is
data governance.
Data governance
Data governance is a rapidly evolving and poorly defined field.
Nevertheless, having a data-driven culture from the outset, top-
level management at Western Digital prioritised the creation
of internal data governance structures. This was in response to
increasing complaints from workers who were unable to find
the data they needed to do their jobs. Most often this was not
because the data was not captured and stored, but because there
was no systematic approach to finding and understanding data
assets scattered around the global organisation. Some specific
issues were the lack of a standard naming convention, and lack
of comprehensive metadata to describe the various attributes of
individual pieces of data.
In response to this, a ‘data council’ was formed, comprising key
representatives from business users and IT, which produced several
beneficial outcomes, including a maintained data dictionary. Woo
believes it was the involvement of end-users in the data council
that led to its success. Drawing on the Western Digital experience,
Woo argues that the key to success - regardless of the types of
data sets involved - is to have data governance structures in
place. To start, he recommends an organisation may implement
a data governance structure for existing structured data sets
(conventionally RDBMS systems). The next step would be to
extend this data governance structure to include unstructured
data (e.g. log files from equipment, as well as data captured from
social networking services).
Resourcing and Malaysian
environment
A wide skill set is required for an organisation to successfully
harness Big Data, but Woo suggests that it must start with domain
knowledge. It is crucial to be well grounded with the vision,
goals, and objectives of the core business. Otherwise, all that
data may simply prove too distracting to actually get anything
relevant done. Woo observes, “One must know how to ask the
33

right questions.” In addition to domain and business knowledge,
a scientific background is helpful. “Big Data experts are few and
far between”. Given the steep learning curve, this scenario is
unlikely to change very soon, and developing internal capacity is
especially difficult. However, a core analytics team is one that is
best grown at home, due to the need for a good grasp of evolving
business imperatives. Infrastructure and technology aspects, on
the other hand, are best outsourced to proven experts.
In general. Woo does not feel that Malaysian initiatives are
particularly disadvantaged, because for the most part, the
challenges around Big Data projects affect all companies
worldwide. However, one specific frustration is the relative lack
of relevant training. Woo believes that Malaysia is relatively
underserved in this regard, compared to other locations such as
Singapore.
34

35

Ian Phoon,
Malaysian bank
Ian Phoon works in risk management at a Malaysian bank
assessing credit risk. Part of their department’s role is assessing
loans and setting limits for commercial and corporate loans.
Phoon has worked at several local banks since graduating from
the University of Malaya. He has a background in Management
Information Systems including streams in computer science and
user experience / usability. His four year degree program included
one semester at the University of Melbourne in Australia, and an
internship at a financial software business.
Data integrity and prediction
At his current employer, Phoon works with a criteria rating
system. His department assesses the performance of clients in
determining loan limits; their objective is to improve loan quality
from a credit perspective. Phoon is responsible for data integrity,
helping to analyse, prepare and visualise potential data problems.
This involves preparing data, assessing data input quality and
providing design feedback on backend systems. Big data offers
opportunities for predictive analysis on prospects that are likely
“There is a clear need
for independent
benchmarking data in
the Malaysian financial
sector”.
36

to default through data modelling. The most valuable potential of
Big Data is in predictive analysis for ‘early warning systems’. This
involves using historical information to determine patterns and
triggers of company defaults on loans. External news reports and
sentiment analysis from the market in assessing the performance
of companies can provide valuable ‘predictive’ data.
Tools of the trade
In his current role, Phoon makes heavy use of traditional statistics
tools with a long established track record in the finance domain,
such as SAS and EViews. Nevertheless, he believes there is
scope for new and improved tools and solutions (e.g. in areas
such as data visualisation) and that the need for integration and
interoperability between these tools is important. Despite the
growing prevalence of analytics-as-a-service and other cloud-
based Big Data solutions, cloud services is not likely to be adopted
by the traditional banking industry anytime soon, particularly due
to perceptions on security.
Phoon believes the most important skills for a data scientist are
in managing databases and having expertise in statistics. General
Business Intelligence skills are also important. He believes data
scientists can help bridge the communication gap between
business and IT functions. As a data scientist, data quality is a key
concern in understanding how input integrity can affect prediction
models. He believes there will be a growing demand for data
scientist degrees, but first organisations need to understand the
value that data scientists can add.
37

Industry benchmarking
Phoon says “there is a clear need for independent benchmarking
data in the Malaysian financial sector”. Current regulation
prevents information sharing between financial institutions.
Commercial banks would benefit from industry overviews such
as the number of loans and customer segments as well as
guidelines on best practice. Benchmarking data can also assist
with predictive modelling. While newer business models such as
online credit services are emerging, Phoon maintains that the
customer-building strengths of traditional banks will give them a
competitive edge.
In Malaysia, banking, telecommunications and marketing are
important industries where Big Data can make an impact, but
Phoon sees that adoption of Big Data technologies is not regarded
as critical yet. He suggests that education around successes in
specific sectors should be shared more widely, even with other
industries, to help the rate of big data adoption. Campus talks
from Big Data practitioners would also help spread awareness
and drive the growth of talent supply.
38

Conclusion
In this report, we have sought to capture the state of Big Data
analytics in Malaysia in all its diversity and varying levels of
maturity. We know there is strong support for Big Data at grass-
roots and executive levels. But aspirations and enthusiasm for
Big Data is not uniformly matched with resource commitment.
There are several inhibitors that prevent organisations in
making headway with Big Data projects, including recruiting the
appropriate skills, the ideation of novel, high value use cases and
access to datasets. Our report raises many more questions. There
are varying requirements for the elusive role of the data scientist.
There are further questions of how Big Data budgets should be
optimally split between infrastructure, data, human resources and
other areas. We prompt stakeholders to consider:
• Are you committing resources aggressively enough?
• Are you driving a data-centric approach to all organisational
processes?
• Are you considering out-of-the box end uses to give you a
competitive edge?
• Are you prioritising the right blend of skills and capabilities?
• Are you participating in cultural change for Big Data
advocacy?
Big Data involves not just technological evolution, but a cultural
shift in the way we work and share knowledge. These values
include an increasing emphasis on transparency and continuous
innovation. The advent in Big Data brings rapid change but also
emphasises interest in classical disciplines such as mathematics
and statistics. The pitfalls in the use of Big Data should be at
the forefront in considering adoption. Statisticians warn us that
correlations do not equal causation, and that; the availability of
increasing variables does not guarantee greater clarity. The privacy
rights of individuals will also constrain how organisations collect,
manage and distribute data. However, we believe the productive
value of Big Data analytics in societies outweighs drawbacks if
these concerns are respected and managed carefully. We have no
doubt that the Big Data analytics industry will continue to mature
and evolve in unexpected ways in Malaysia that will support the
nation’s competitiveness in the global economy.

39

About the authors
Sandra Hanchard
Sandra is the Organising chair of Big
Data Malaysia. She is an industry
analyst with expertise in consumer
adoption of new technologies.
Formerly, Sandra was the Asia Pacific
lead for custom research with global
Internet measurement firm, Experian
Hitwise. She worked with leading
blue-chip and new media brands
to grow their businesses through
insights derived from Big Data. Her
commentary has featured prominently in the press in Australia.
Currently, Sandra is the recipient of an Australian Postgraduate
Award for her doctoral thesis which investigates social media
information use in Malaysia. She is a past speaker at the
Melbourne Writers Festival and a fellow of the Oxford Internet
Institute Summer Doctoral Program at Oxford University.
sandra@bigdatamalaysia.org
Tirath Ramdas
Tirath founded Big Data Malaysia while
working with Acunu, a London-based
Big Data startup. Though that was his
first role that explicitly wore the label
‘Big Data’, his first exposure to large
scale information processing occurred
many years earlier during his PhD in
computer architecture for applications
in computational quantum chemistry.
Other prior roles include systems
administration, embedded systems
development for VoIP and video applications, non-profit strategy
consulting, as well as business intelligence and data warehousing
conceptual application design. Currently based primarily in
Melbourne, Australia, he is a software development consultant to
information security startup Bromium.
tirath@bigdatamalaysia.org
40

About Big Data
Malaysia
The first Big Data London meetup was held in a co-working space
near London’s famous Leicester Square on the evening of May
25th 2011, featuring talks on Big Data use cases in interactive
TV, games, and business intelligence. Founded by London-based
startup Acunu, the group was created and developed to foster
awareness and understanding of the potential for Big Data to
transform businesses, industries, and society at large – while
always digging into the technology itself.
One year later, the first Big Data Malaysia meetup was help in
the Mercury room at Telekom Malaysia Research & Development
in Cyberjaya on the afternoon of May 16th 2012, featuring talks
on R, applications in genetics and air traffic modelling, as well as
marketing services. Also founded by Acunu, Big Data Malaysia
shared much of Big Data London’s ‘DNA’ - in particular core values
such as:
1. Maintaining a broad appeal (business executives in suits as
welcome as hackers in shorts).
2. Embracing the wide spectrum of issues to do with Big Data
(e.g. distributed systems and data journalism are equally
relevant).
3. Emphasis on interesting content (more than just a
networking forum).
Over time Big Data Malaysia has evolved in its own way to address
needs and opportunities in Malaysia, embarking on exercises such
as partnering with relevant organisations (such as community/
professional groups and commercial conference organisers),
conducting a group brainstorming workshop to discuss the
emerging impact of Big Data on society at large, coordinating
Big Data Week 2013, and producing the “Big Data in Malaysia:
Emerging Sector Profile” survey.
Although Big Data Malaysia has evolved to conduct activities
beyond meetups, running regular meetups is a core aspect of
the group, to allow people to network and learn. To that end in
our first 24 months we have run many meetups as listed here:
• “Hello World”: May 2012, sponsored by Acunu, hosted by
Telekom Malaysia R&D, Cyberjaya, with talks by:
• Julian Lee, Revolution Analytics
• Daniel Walters, Experian Marketing Services
• Hakim Albasrawy, Tandemic
• “Bigger, Faster, Messier”: July 2012, sponsored by Acunu,
hosted by iTrain, Kuala Lumpur, with talks by:
• Nicolas Yip, Innity
• Calum Halcrow and Kaveh Mousavi Zamani, RMG
Technology
41

• Joint meetup with SNIA Malaysia: September 2012,
sponsored by SNIA Malaysia and Acunu, hosted in Cyberjaya,
with talks by:
• SNIA tech tutorial on flash storage
• Rachelle Foong, HP
• Tirath Ramdas, Acunu
• “Big Data in Telecommunications”: October 2012,
organised and hosted by TM R&D, Cyberjaya with talks by:
• Zach Tan, Google
• Raymond Au, Oracle
• Shamsul Anuar Yahaya, Telekom Malaysia
• “Big Data in Social Media Analytics”: December 2012,
• Sandra Hanchard, Swinburne Institute for Social Research
• “Velvet Underground”, 30th January 2013, sponsored by
Acunu and Experian, a networking event coinciding with the
Market Research in the Mobile World (MRMW) conference,
Zouk, Kuala Lumpur.
• “Big Data in Market Research and Beyond”, 1st February
2013, sponsored by Acunu and Experian, coinciding with the
Market Research in the Mobile World (MRMW) conference,
• Colin Pal and Taufek Johar, Experian Marketing Services
• Tirath Ramdas, Acunu
• Shazri Shahrir, Telekom Malaysia R&D, and Nurazam Malim,
Twistcode
• Rebekah Lee, Integricity Interactive
• Attendees also participated in a group brainstorming
exercise, discussing how Big Data will impact society from
the perspective of small business, big business, government,
and individuals/consumers.
• Malaysia was part of Big Data Week 2013 in mid-April
2013. Big Data Malaysia assumed an oversight role during
this festival, working to encourage a balanced program as well
as fostering cross-promotion and minimising event overlaps.
All together 9 separate events were part of the program, each
of which was organised by different partner organisations:
• TRANSIT: Distillation
• Python Malaysia: Big Data Week Workshop
• Big Data in Telecommunications 2
• Big Data: Strategy, Practice, and Research
• SNIA Malaysia Tech@Break
• NoSQL Asia: Exploring The Technology Behind Big Data
Week
• GPU Technology Workshop Asia 2013
• The Data Revolution: Powered by the AWS Cloud
• Data Journalism: Storytelling in Malaysia with Big Data
42

• “Big Data Malaysia - Anniversary 1.0”: 30th May 2013,
hosted by iTrain, Kuala Lumpur, Emceed by Daniel Walters,
Experian, with talks by:
• Khoo Swee Chuan, Netapp
• Siah Hark Cheong and Mun Heng, from a multinational
financial services company, but speaking in their personal
capacity
• “Big Data Malaysia, with F-Secure”: 1st August 2013,
sponsored and hosted by F-Secure, Kuala Lumpur, emceed by
Ng Swee Meng, OnApp, with talks by:
• Keng Chee Chan, F-Secure
• Aizu Ikmal Ahmad, Malaysiakini
• Zakiullah Khan Mohammed, independent software
engineering management consultant
• “Big Data in Telecommunications 3”: 24th September
2013, sponsored and hosted by TM R&D, Cyberjaya, emceed
by Tirath Ramdas, Marklar Marklar Consulting, with talks by:
• Mohamad Ansahari Bin Abdul Kudus, Telekom Malaysia
• Szilard Barany, Teradata
• Rosmawati Binti Ab Raub, Telekom Malaysia Research &
Development
• “Cents & Security”: 20th February 2014, sponsored and
hosted by F-Secure, Bangsar South, emceed by Szilard Barany,
Teradata, with talks by:
• Foong Chee Mun, MoneyLion
• Goh Su Gim, F-Secure
• Pak Mei Yuet, MDeC
The first two years of Big Data Malaysia have uncovered a
significant and still rapidly growing interest and excitement for
the possibilities of Big Data in Malaysia, and we look forward to
seeing what comes next. We believe that a year from now, we
shall look back to see that this was the turning point where the
majority of discussion changes from “this is what we could do” to
“this is what we’re doing.”
Big Data Malaysia exists on a number of social media sites; our
Facebook group is especially active, and we also have a LinkedIn
group, a Meetup.com group, and a mailing list. Links to all
these properties may be found on our landing page at
http://bigdatamalaysia.org
43

Big Data in Malaysia: Emerging Sector Profile 2014

Big Data in Malaysia: Emerging Sector Profile 2014

More Related Content

What's hot

Viewers also liked

Similar to Big Data in Malaysia: Emerging Sector Profile 2014

Recently uploaded

Big Data in Malaysia: Emerging Sector Profile 2014