Applied Data Science in Europe

Challenges for Academia in Keeping Up with a Highly Demanded Topic

T. Stadelmann, K. Stockinger, M. Braschler, M. Cieliebak, G. Baudinot, O. Drr and A. Ruckstuhl
Institute of Applied Information Technology
Zurich University of Applied Sciences
Winterthur, Switzerland
[stdm, stog, bram, ciel, baug]@zhaw.ch
AbstractGoogle Trends and other IT fever charts rate Data
Science among the most rapidly emerging and promising fields
that expand around computer science. Although Data Science
draws on content from established fields like artificial
intelligence, statistics, databases, visualization and many more,
industry is demanding for trained data scientists that no one
seems able to deliver. This is due to the pace at which the field
has expanded and the corresponding lack of curricula; the
unique skill set, which is inherently multi-disciplinary; and the
translation work (from the US web economy to other ecosystems)
necessary to realize the recognized world-wide potential of
applying analytics to all sorts of data.
In this contribution we draw from our experiences in
establishing an inter-disciplinary Data Science lab in order to
highlight the challenges and potential remedies for Data Science
in Europe. We discuss our role as academia in the light of the
potential societal/economic impact as well as the challenges in
organizational leadership tied to such inter-disciplinary work.
Index TermsData Science, Multi-disciplinary, Curriculum
development.

I. INTRODUCTION
Data Science seems to be the new magic bullet that could
solve many problems of todays enterprises and even societies.
The term is often used in close relationship with Big Data and
is ranked high on the agenda of CIOs across all industries.
Definitions vary, but have in common that Data Science
comprises a unique blend of skills from analytics, engineering
and communication aiming at generating value from the data
itself [1][2]. This makes Data Science inherently an applied
science, with the goal of applying various tools and techniques
to data in order to gain a data product, an exploitable insight
derived from collected facts.
It is interesting to see that especially in industry Data
Science is often considered as the new kid on the block even
though data-intensive sciences such as high-energy physics,
astrophysics or bioinformatics have been using certain data
science concepts already for decades for instance in the field
of scientific data management [3] and eScience [4]. These
communities are used to processing, analyzing and visualizing
terabytes and even petabytes of data to enable domain scientists
achieve scientific breakthroughs. Recent trends in social media,
however, allowed end-users to produce large amounts of data
that were originally only exploited by large web companies
such as Amazon, eBay, Google or Facebook. This is a strong

Institute of Data Analysis and Process Design

Zurich University of Applied Sciences
Winterthur, Switzerland
[dueo, rkst] @zhaw.ch
publicly visible sign of Data Sciences potential impact on
society and economy.
II. PERSPECTIVES TOWARDS INDUSTRY
For example, many companies realized that the changing
user behaviors and the higher affinity to online media of
younger generations have large impacts on their businesses. In
order for established enterprises to attract these new
generations, they must adapt their business models to the new
needs in order to avoid being taken over by web companies.
Why would customers go to physical shops if a majority of the
products can be bought online on Amazon - that in turn even
suggests articles that are bought by like-minded people? Why
would future generations go to expensive financial advisors of
established banks, when Google offers often better financial
advice by analyzing search behavior using Google Trends?
Understanding the needs of the new online society is key
for succeeding in todays business world, and Data Science is
one approach towards data-driven decision making as opposed
to using gut feelings. We currently see a shift in industry that
already happened in data-intensive science decades ago. Hence,
industry should learn from science a discipline that is used to
observing, modeling, simulating and explaining nature and the
respective natural laws: For instance, a recent success story has
shown that data mining techniques based on neural networks
that were originally designed for high-energy physics analysis
at CERN, have helped analyzing and interpreting the user
behavior of the German retailer Otto [5].
However, many industries are still struggling with so-called
Big Data-demands, as they are not quite sure how the new
trend fits into the current enterprise application landscape.
Analytical questions, as for instance in marketing, have
successfully been answered before using data warehousing and
business intelligence technology with data mining methods.
What is the additional benefit of Data Science? And how does
it scale from web economy problems to applications in
established European industries like industrial production, as
suggested by the concept of Industry 4.0 [6]? Data Science
education and research are needed to find answers to these
questions.
III. PERSPECTIVES TOWARDS TEACHING AND RESEARCH
Based on the demands from industry, the new job profile
data scientist has emerged whose interpretation differs widely

between various industries. Our goal is to foster a curriculum

as well as a research agenda for future data scientists that are at
the frontier of both applied research and industrial needs. The
skill set of these data scientists is multi-faceted as shown in
Figure 1. What follows portrays our work-in-progress and is
intended to start a discussion in Europe.

Fig. 1. Our interpretation of the Data Science skill set.

How do we ensure that we will be able to teach these

diverse skills? Our approach at Zurich University of Applied
Sciences is to address these challenges as part of the Datalab, a
multi-disciplinary virtual organization spanning several
institutes and departments for a collaboration between
researchers and practitioners of various areas that altogether
cover most of the depicted skill set. Data scientists require
entrepreneurial experience, pursue different research interests
from e.g. algorithms to visualization and come from
backgrounds like computer science, statistics or law.
Our experiences from various projects in both research and
industry have shown that truly multi-disciplinary teams are
most suited for solving complex problems that require
innovative out-of-the-box-thinking. Knowing one specific
discipline is often not enough. A broader perspective also helps
in picking up necessary domain knowledge faster. Hence, one
of the main aspects of our curriculum and applied research
agenda is to foster multi-disciplinary approaches, which
requires a broad education based on sound scientific methods
in computer science, statistical learning etc. as well as best
practices based on case studies from industry that pass the test
of real life. The ideal data scientist is thus a scientist with
entrepreneurial skills, who is used to asking the right business
questions, understands the techniques and is familiar with the
tools for solving them.
We believe that as a university of applied sciences we are
ideally positioned to address these dual challenges as most of
our research is already conducted within the confines of
projects executed with industry partners. However, in order to
make sure we are not trapped in local optima (i.e. country
specific adaptations), another goal is to establish the basis for a
European-wide Data Science curriculum and research agenda

that can be generalized to a wider range and is accredited by

the international Bologna system.
Our proposed curriculum and research agenda cover the
following topics:
Database and Cloud Computing technology for Big
Data
Data Mining, Statistics and Predictive Modeling
Machine Learning and Graph Analytics
Information Retrieval and Natural Language Processing
Business Intelligence and Visual Analytics
Data Warehousing and Decision Support
Communication and Visualization of Results
Privacy, Security and Ethics
Entrepreneurship and Data Product Design
The goal of the data science curriculum foresees that
students will work on coherent real-life projects and will apply
data science to various case studies from science and industry
to learn crafting business cases that have impact.
One of the challenges of inter-disciplinary education and
research is that different university departments need to work
more closely together to come up with joint agendas. What has
been shown to work successful in large-scale sciences such as
e.g. astrophysics or bioinformatics could be a role model for
data science centers: departments, universities, labs and
industry join necessary forces to cope with the high complexity
of the subject. This might require re-thinking established
scientific disciplines and organizational structures, being open
for applied research topics that at first glance do not fit directly
into the core competences of the respective departments.
However, the best scientific results with positive impacts on
society often happen at the crossroads between well-established
fields, i.e. where inter-disciplinary teams tackle larger problems
jointly rather than staying in the comfort zone of traditional lab
sciences with narrow focuses.
IV. CONCLUSIONS
Data Science is a new paradigm that has the potential to
revolutionize enterprises in a similar way as the introduction of
computers and IT changed business processes in the first place.
In order to make sure that this paradigm shift does not only
happen in Silicon Valley but also in Europe, a streamlined
curriculum and research agenda is required that adapts to the
European market. These new challenges offer great
opportunities but also responsibilities for universities and
higher education in general in order to stay abreast of
technological advances and to shape them for the benefit of
society.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

http://radar.oreilly.com/2010/06/what-is-data-science.html
http://online.liebertpub.com/doi/pdf/10.1089/big.2013.1508
https://sdm.lbl.gov/sdmcenter/
research.microsoft.com/en-us/collaboration/fourthparadigm/
http://www.big-data-europe.com/frankfurt-ruckblick-2012/
http://www.bmbf.de/de/9072.php