Sie sind auf Seite 1von 141

dataviz

30
practical examples
to better understand data visualization

dataviz
Data visualization: Is it still only a concept, or is it a reality for organizations? To
find out, we conducted a study involving organizations from a variety of
business sectors (industry, media, transport, distribution, services, etc.). The
result is this book, in which more than 30 companies describe not just the
theoretical advantages of visually representing data, but concrete projects that
are already producing results.
Their insights and expert opinions will help you understand data visualization
and how to use it in your organization.

Introduction

hat do we mean when we talk about data? Numbers,

visualization eorts compare with traditional business

charts, tables? It's true that data is all around us; yet

intelligence approaches.

it is dicult to capture in a meaningful way. Its like


trying to hold sand; it always slips through your fingers.
We need data to guide us through decision-making
processes, but how do we avoid drowning in it, and how do
we use it as a real tool for insights and decision making?
Data visualization is the answer. Why? Because presenting
data visually enables us to understand that data better.

Next, we explore how to implement a data visualization


project. In this section we will review the criteria for success
and identify pitfalls.
Finally, our panel will envision the future of data
visualization. So lets begin our journey of discovery into the
terra incognita of data.

We asked 30 pioneering companies how they visualize their


data. Understanding their real-world experiences will help
you explore data visualization and how to use it to your
advantage. Along the way, SAS experts and other market
stakeholders and observers will shed light on dierent
aspects of data visualization as it relates to business
intelligence (BI).
The first part of this book is an overview of data
visualization. We will define it and examine its origins and
varied uses for a wide range of stakeholders across dierent
business sectors.
The second part uncovers the benefits of data visualization,
in particular the three building blocks of data visualization
data discovery, data analysis and information
communication. We will also see how organizations data
2

1
1

Data visualization and you:


An introduction

1
1.1 Defining data visualization
Data visualization is a hot topic these days. As with all buzzwords, it becomes a catch-all term, and not many of us know what it
truly means. Luckily, we wont be alone in our search for data visualizations meaning and value, because the expert
contributors to this book have agreed to help. So lets start by defining the term.
SAS defines data visualization as "the visual and interactive exploration and graphical representation of data of whatever size
(from small data to big data), nature or origin. It allows users to identify phenomena or trends that are invisible at first glance."

1.1 Defining data visualization


Our panel members share this view, judging by their own definitions below.
Data visualization: A collection
of definitions

We regard data visualization as a way of providing the user with information in a


visual form to facilitate access to it.

Caroline Denil, Data Warehouse Manager,


SPF Economie

Data visualization is the art of visual representation for encouraging the best
possible comprehension.

Jrme Tharaud,
Head of Research and Development,
Prisma Media
numbers. Guillaume Deschamps of "exploit"
"quicker, better undersatnding"
Lets examine a couple of
STMicroelectronics describes data Using data visualization to exploit It also fosters a communication
interesting terms from the
visualization as "a way of
data (for decision making) does not flow because visualization allows
quotes above
"visual form"
Representing data in a visual
format is the foundation of data
visualization.
"facilitate access"
Graphical representation should
always serve some purpose, the
first of which is to enable easier
access to the information
conveyed by the data. For most of
us, it is easier and more satisfaying
to look at a chart than a series of

translating a set of dry, abstract


data into information that can be
understood. Fostering a common
understanding is about giving
meaning to the data in order to
extract information from it.

just foster intellectual


understanding; it transforms raw
data into information that enables
action. Quicker data visualization
accelerates understanding.

data to travel beyond the BI or


statistical spheres into the entire
organization, becoming a unifying
platform for decisions and
collaboration.

"deduce insights"
Data visualization not only allows
intellectual comprehension, it also
transforms raw data into practical
information.

"sharing information"
Information sharing is also a
means of communication that
prevents data from remaining
isolated within the BI or strategy
realms and allows it to permeate
the whole organization and
become a tool for decision

1.1 Defining data visualization


Data visualization can inform an organizations operational approach as summarized below by Maguelonne Chandesris of
SNCF Innovation and Sandrine Noail of Coliposte. :
A means, not just an
end

Data visualization is not an end in itself; it is a means for meaningful production used to better

understand phenomena. In other words, we provide professionals with the ability to better grasp a
reality that they already know by suggesting they consider other points of view.

Maguelonne Chandesris,
Head of the Statistics Team, Econometrics and Data Mining,
Innovation and Research Directorate SNCF

I was manager of our KPI project, and as a result I managed the rationalization and optimization of our
performance indicators. The important thing with this project, that in time it will be generalized and deployed
across the sales teams, as a catalyst for the decision-making process and supported by data visualization. It will
help answer questions like, what can we do to perform better? In what areas do we need to intensify eorts to
quickly obtain specific results?

Sandrine Noail,
Head of Control and Analysis, Coliposte

1.1 Defining data visualization


In order to round out our definition of data visualization, it is important to highlight the three essential characteristics for its
practical use. :
Three characteristics of
data visualization
The value of data visualization is in its ability to respond to these essential requirements:
It can be interpreted.
The volume of data is snowballing with the exponential growth of Internet use (and in particular Google).
Unstructured data is undergoing the same evolution. If data visualization were based on data that could not be
clearly interpreted, it would serve no purpose.
It is relevant.
Relevance is linked to how data can be interpreted. Data visualization must focus on relevant responses to
questions within a defined context and aimed at specific objectives.
It is innovative.
Finally, data visualization is only of interest if it provides new information - something original - to give a new
perspective on the subject.

Dorina Ghiliotto-Young,
Head of Innovation,
Ingenico

1.1 Defining data visualization

1
A brief history of
data visualization

The graphical representation of statistical data has its roots in the work of the Scottish statistician
William Playfair (1759-1823). He invented tools that we still use today - the circular (or pie) chart, the bar
chart and the line graph. At a time when the Cartesian tradition of using mathematical functions
dominated, he highlighted the importance of graphs for representing data. Playfair asserted that simply displaying data in the
form of points and shapes on a chart can improve comprehension. It is one of the ironies of history that , Playfair remained
largely misunderstood by his contemporaries.1
In the 19th century, another trailblazer was Charles-Joseph Minard. He became famous for his "figurative" map of the
successive losses of men in the French army in the Russian Campaign of 1812-1813, published in 1869.2

Minard succeeded in representing (in a single diagram) data of dierent forms such as numbers of troops, geographical
features (distances, towns and water courses), temperatures and directions (toward Russia or in the opposite direction).
1.Howard Wainer : "William Playfair" (version 3). StatProb: The Encyclopedia Sponsored by Statistics and Probability Societies.
Freely available at http://statprob.com/encyclopedia/WilliamPLAYFAIR.html
2.http://commons.wikimedia.org/wiki/File:Minard.png

1.1 Defining data visualization


But we had to wait until the 1960s and the work of
statistician John Tukey at Princeton University for data
visualization to be finally recognized for its true value.
Among other contributions, Tukey invented the box-andwhisker plot (also known as the box diagram, Tukey box or
box plot). This plot allows series of numbers to be
represented graphically using five characteristics:
minimum value, first quartile (Q1), median, third quartile
(Q3), maximum value.
The box plot is often added to illustrate the disparity of
distribution between the median and mean.

In the 1970s, Tukey collaborated with Edward Tufte, a


professor at Princeton University, on a series of seminars
on the graphical representation of data. These seminars
would later be included in a book by Tufte seen as a
landmark for data visualization, The visual display of
Quantitative Information.
More recently, we have the Swede Hans Rosling to thank for
his animated bubble chart, which he presented at the 2006
TED conference. More than 5 million people have view a
video of his talk.4

The box plot is also able to highlight misleading data.3


- An example of the box-and-whisker plot used at
STMicroelectronics:

Hans Rosling at the TED conference

3. Sources : http://en.wikipedia.org/w/index.php?title=Box_plot&oldid=551581640 and http://fr.wikipedia.org/w/index.php?title=Bo%C3%AEte_


%C3%A0_moustaches&oldid=90308055
4. http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen.html

1.1 Defining data visualization

1
Why does a graph say more about the
data than the data itself?
In 1973, statistician Francis Anscombe created his eponymous Anscombe's
quartet to show that a graph is more capable of showing the reality behind the
data than the data itself. The quartet takes the form of four sets of X and Y data
as shown below.
When looking at the tables, it is hard to understand what events and properties
are being represented?
To see more clearly, you can perform some statistical calculations to identify
properties. And then, lo and behold, you can see that the four sets have identical
properties. It would therefore be tempting to believe that the four sets represent
similar data.
If you plot the data sets as points on a graph, however, you will notice that they
represent completely dierent models!
Property

Value

Mean of each x variable

9.0

Variance of each x variable

10.0

Mean of each y variable

7.5

Variance of each y variable


Correlation between each x and y
variable
Linear regression line in

3.75

III

II

IV

10.0

8.04

10.0

9.14

10.0

7.46

8.0

6.58

8.0

6.95

8.0

8.14

8.0

6.77

8.0

5.76

13.0

7.58

13.0

8.74

13.0

12.74

8.0

7.71

9.0

8.81

9.0

8.77

9.0

7.11

8.0

8.84

11.0

8.33

11.0

9.26

11.0

7.81

8.0

8.47

14.0

9.96

14.0

8.10

14.0

8.84

8.0

7.04

6.0

7.24

6.0

6.13

6.0

6.08

8.0

5.25

4.0

4.26

4.0

3.10

4.0

5.39

19.0

12.50

12.0

10.84

12.0

9.13

12.0

8.15

8.0

5.56

7.0

4.82

7.0

7.26

7.0

6.42

8.0

7.91

5.0

5.68

5.0

4.74

5.0

5.73

8.0

6.89

II

III

IV

816
y=3+0.5x

The Anscombes quartet goes at the right hand side of the page, the entire
text on this page on the left hand side!
Data visualization allows you to ask new questions about series of data whose
diering characteristics could otherwise remain unnoticed.

10

1
1.2 What is data visualization used for ?
Now that you have a better understanding of what data visualization is, lets examine its practical uses. Data visualization is
applied to a wide array of activities among the organizations we interviewed.
Our panel represented the following business sectors:
FMCG or Fast Moving Consumer Goods (LOral and Ubisoft)
B2B services (Accenture)
Manufacturing (Ingenico, Philips, Schlumberger and STMicroelectronics)
Media (BFM Business, France Tlvisions, InfoPro and Prisma Media)
Marketing and advertising (IFR-GfK, Lagardre Active, PagesJaunes and SM-Marketing Convergence)
Scientific research (IMEC)
Public service (SPF conomie)
Telecom (Alcatel-Lucent, Orange and SFR)
Transport and logistics (Aroports de Paris, ID Logistics, SNCF and Voyages-sncf.com)
All of these organizations use data visualization either to improve certain things that they were already doing (optimization) or
to supplement their activity around value-added tasks (innovation).
Regarding optimization, Lagardre Active for example was able to accelerate production, and STMicroelectronics was able to
enhance the performance of its manufacturing processes thanks to data visualization.
To innovate, PagesJaunes identified and closed gaps in its indexing thanks to data visualization. Voyages-sncf.com introduced
an innovative service to the transportation and logistics market (Mytripset). And Alcatel Lucent is envisioning tomorrow's
mobile applications.

11

1.2 What is data visualization used for ?


The following pages present five use cases in more detail. These case studies illustrate the goals of optimization and innovation
by using data visualization to:
Improve management of an organizations resources.
Improve client relations.
Define a company's product.
Contribute directly to company profits.
Empower citizens.

Improve management of an organizations resources


One of the primary uses of data visualization is more ecient control of organizational activities and performance. The focus
here is on the operational side of the business.
Data visualization used to
manage an organizations
resources
Our company was being swept aside by the competition.
We had to use more modern decision-making tools in order to anticipate where
our competitors stood and the adverse eect of this competition on our client
portfolio. We wanted to make data visualization a strategic decision-making tool
to be used by line managers to control organizational performance.

Sandrine Noail, Head of Control and Analysis,


Coliposte

We wanted to provide the board with a real management tool. In other words, we
wanted a dashboard with all the strategic indicators required to make decisions.

12

1.2 What is data visualization used for ?

Source: SAS

Analysis of calls by equipment type

An example of data visualization in the


telecom sector
The main components of a communications providers network infrastructure have an impact on call quality and, in general,
the quality of services, all of which aect revenue.
Switches, the principal components of a telephone network, play a vital role in routing calls, analyzing and monitoring trac,
and maintaining service quality. Having the ability to see the dierent types of equipment in use is important in maintaining a
healthy network.
This shows the trac indicators (call volume and interrupted calls) for dierent manufacturers and pieces of equipment.

13

1.2 What is data visualization used for ?


IMEC, a Flemish micro- and nano-electronics research center, provides an interesting example of data visualization.

Data visualization as a
reporting tool
We have been using SAS for years for scientific projects. A few years ago we also started
using it for our company reports.
We started o with a simple parameter (telephone use by our sta) and then added new fields in succession
in order to gradually enter all information in the BI environment.
We use the solution mainly to represent the indicators we need to produce for the Flemish government, for
example, with regard to the number of publications produced by our researchers, the number of doctorates
or even our activities that lead to companies being founded (spinos).
In addition, we also produce reports to visualize our scheduling and performance in terms of how our
resources are used.

Yves Daelmans, ICT Project Manager


IMEC

14

1.2 What is data visualization used for ?


IMEC cost of telephone per employee

Talent management (employee breakdown by


nationality at IMEC in a given month)

12-month summary of time spent by an employee


on corporate activities

15

1.2 What is data visualization used for ?


How data visualization can help organizations move forward by looking into the past.

Using data visualization or visual


ideas to create a strategic
direction
The idea of data visualization arrived at LOral at a time when all digital teams, both
for brands and in the countries, were equipping themselves with tactical dashboards;
our strategic dashboard project therefore brought together all the local campaigns and allows us to have a
real strategic vision.

Laurence Kerjean, , Global Digital Manager


LOral
A strategic dashboard at LOral

16

1.2 What is data visualization used for ?

Improving client relations


Marketing and client relationship management are two areas where you can apply data visualization, as the examples below
from SFR and Lagardre Active illustrate.

Visualizing client data to improve


multichannel CRM

Our activity is focused around three main channels: points of sale, customer call
centers and our online boutique at sfr.fr. We are developing our dataviz solutions

with the view to optimize the management of these channels. Above all, our eorts are aimed at improving
management and better understanding our KPIs. In this sense, our projects are very focused 'activity
indicators'... Our data visualization development is therefore associated with bringing together numerous
sources of data in order to find relationships and causes in them.

Franois NGUYEN,
Director for IT and Relationship Marketing, SFR

17

1.2 What is data visualization used for ?

Data visualization and the multichannel


customer relationship
Analysis of customer relationship centers

Multiple communication channels are an additional challenge for

Source: SAS

customer relationship centers tasked with maintaining customer


satisfaction. The volume and complexity of data (often compartmentalized
by silos) requires analysis in order to increase eciency and optimize costs.
The above screenshot shows the main performance indicators (quality of
service, uptake time and close rate) with a focus on the first-contact
resolution rate.

A good customer relationship also requires clear answers to these questions: . What are their interests and behaviors? How do
you segment them? The ability to examine the data using data visualization enables you to get those answers. More about this
later in part two.
Data visualization for better
qualification of customers

At Lagardre Active, we have one of the largest databases of subscribers to our

magazines, with more than 3 million customers. We would like to develop the analysis
capabilities of our CRM by integrating the dataviz tools into it in order to benefit from a refined overview of
our range of clients.
Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active

18

1.2 What is data visualization used for ?

Better product definition


For product development, data visualization provides companies with the tools to better define their products and features.
Exploring customer data and the ability to test options are particularly valuable.

SNCF uses data visualization to


develop products in dierent fields

The sources of data at SNCF are extremely varied, with many possible fields of
application. We first worked with dataviz back in 2010 and are currently running

several projects aimed at various topics: information, traveler frequency and economic impacts, yield
management and energy-related matters.

Maguelonne Chandesris,
Head of the Statistics Team,
Econometrics and Data Mining,
Innovation and Research Directorate, SNCF

Applying data visualization to yield


management

Applying data visualization to the electricity


consumption of trains

19

1.2 What is data visualization used for ?

Contributing to the business


Data visualization can have a direct eect on a company's business. Charles de Rau explains below how Ubisoft has enabled
game developers to better understand its market.
A better understanding of the
competitive positioning
We set up a project allowing us to compare our brands with those of the competition
using more than a hundred criteria. For this, we relied on the data collected from thousands of gamers by an
external research institute.
Where before we could only compare our brands against a theoretical average brand, we can now compare
them instantly against all other brands in a graphical way, based on those consumer ratings.

Charles du Rau,
Consumer & Market Knowledge Director,
Ubisoft

Ubisoft's brand comparison mapping

20

1.2 What is data visualization used for ?

Geolocalization associated with dataviz is an indisputable asset for marketing. It allows us to highlight our
catchment areas or to set up a mapping of groups and companies we are targeting. There is no doubt it is a
powerful tool that we use in our competitive analysis.

Clment Delpirou, Associate, InfoPro Digital


Emmanuel Lechypre of BFM Business explains how data visualization adds value for journalists.

Data visualization, a source of


value and dierentiation
In my domain, economic journalism, I have two main sources of information:
numerical data (studies, reports, databases and ocial statistics) and wire service reports that are fact-based
and available to the whole profession. Or people we welcome in the studio and whose insights provide us with
more qualitative material. My interest in analysis and exploratory tools such as data visualization is to obtain
the same insightful and qualitative nature that we can with interviews but based on data. In other words, in the
world of data journalism, data visualization is a way to identify topics, to oer the journalist new angles to deal
with the news and even create the event. In this sense, it is a tool that adds a great deal of value to the
profession and is introduced to develop it.

Emmanuel Lechypre, Editor,


Head of the Analysis Center and Forecasting,
BFM Business

21

1.2 What is data visualization used for ?


Data visualization can also provide insight into how customer behavior is changing and how to adapt to it. The same techniques
can be applied to advertising investments.

Immediately identify advertising


investment opportunities

We had a project to show the investment by our


advertisers by theme in the form of a radar chart.
In particular, the latter could enable trends for
Year N to be compared against Year N-1.
Visualization by advertiser would also enable a
more refined overview of how its investments are
developing in other words, changes made in
terms of subject focus (e.g., dropping one trend
from one year to the next) and benefiting from a
more focused view of trends.

Radar chart representing advertising investment at


Lagardre Active
Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate,
Lagardre Active

22

1.2 What is data visualization used for ?


Data visualization can even directly aect a companys economic model. Philippe Nieuwbourg uses the example of online
selling to help us understand this concept.

When data visualization directly influences a


company's business model
If we take the example of distance selling, companies such as La Redoute
or 3 Suisses obtain the majority of their turnover online these days. Now, if La Redoute did not have a way of
knowing which pages of its paper catalog captured the most attention, or how much time was spent on
average on each page, it can now have this type of information for visitors to its e-commerce site. Thanks to
connection logs, it can identify the most visited pages, how much time a visitor spends on each page, how
they move from one page to another, etc. All this is a potential gold mine, providing we know how to exploit
these logs, to make them talk.
Data visualization is one way of extracting the value of these logs.

Philippe Nieuwbourg,
independent analyst-journalist in decision-making IT

23

1.2 What is data visualization used for ?

Empowering citizens
Data visualization also has its uses outside the business realm. It can help better inform citizens and empower them to act. Two
examples illustrate this nicely: data visualizations available to the public by the Belgian Ministry for the Economy via the Federal
Public Service Economy and the Pariteur" application placed online by France Tlvisions.

Giving people something


to think about
Our dataviz projects come in three formats:
Be Stat, mobilizing data available for more than four years and enables the building of multidimensional
dashboards (SAS cubes). We are proposing a dynamic interface bringing together Java and SAS that allows
citizens to create their own dashboard by selecting the measurements to be shown, the dimensions
according to which they want to view the data and the filters to be applied to the data. Be Stat allows the
results to be exported in CSV and PDF formats.
An economy focused dashboard that produces a static view with interchangeable graphs grouped under a
collection entitled EcoZoom. These graphs are likely to indicate the evolution of petrol price indices, for
example.
A mapping system (or dynamic graphs) is currently in a pilot phase. This is a particularly ecient system for
visualizing densities (population, price, etc.), which are visible thanks to dierent color shadings. We see a
probative application for visualizing the development of property prices over time.
The launch of these dataviz projects aords us the opportunity to present in a dierent way data we have
collected and are already using, that enables citizens to perform their own analysis.

Caroline Denil,
Project Manager, SPF Economie

24

1.2 What is data visualization used for ?


Citizen data visualization projects can take a more activist turn, like France Tlvision's Web application "Pariteur".

France Tlvision's "Pariteur"


Currently, a woman earns 25 percent less than a man in the private sector and 18 percent in the
public sector, based on like-for-like job and equivalent hours. This is an average calculated by
Insee (French Statistics Oce) for 2010 (DADS 2010, Dclaration Annuelle de Donnes Sociales
Annual Statement of Social Data) on salaries surveyed according to age, gender, region and
position (standardized job nomenclature). Based on this information, France Tlvisions
Nouvelles critures and Francetv info developed an innovative and entertaining online and
mobile app, "Le Pariteur," allowing you to study your own pay inequality: "And if I were paid as a
Antoine Allard, Head of Communication and Marketing Social Media,
France Tlvisions
25

1.2 What is data visualization used for ?


Data visualization can be grouped into three main categories, as explained by Dorina Ghiliotto-Young, Head of Innovation at
Ingenico :

Why use data visualization?


Three reasons:
To confirm or overturn assumptions
Data visualization can then take the form of a dashboard, allowing a decision to be made based on an overall
view of the market.
To educate
Internally, Ingenico principally uses data visualization for reporting tasks or brainstorming sessions. It can
also complement other creative approaches such as "gamification" because it develops intuition.
To explore
This is the most forward-looking aspect of data visualization that needs to be developed. Dataviz can help to
construct predictive models. This leads us to the domain of data analysis.

Dorina Ghiliotto-Young,
Head of Innovation,
Ingenico

26

1
1.3 An innovation or a revolution?
Strictly speaking, graphical representations of statistical data are not new. They can be traced back to the 18th century. Skeptics
would use its history to argue that data visualization is in no way revolutionary.
However, Claude-Henri Mldo, co-founder of Aldecis (experts in dashboards and organizational performance systems)
reminded us that the term revolution in itself contains the concept of revolving or returning.
And if you insist that using graphics to represent data is nothing new, then that is simply because we have recently returned to it
and found a power and potential that was never before discovered.
Data visualization marks a watershed in the history of business intelligence. Modern data visualization is based on a series of
major innovations that enable it to reach its full potential.
In the following pages, youll read about the big data movement, and how advancements in neuroscience and the development
of new terminals and new interfaces for accessing information have made data visualization an important feature of the
changing face of data analysis.

Big data
Big data refers to both the current data-explosion phenomenon and the technology for handling it.
Its three primary characteristics are often described as the three Vs: Volume, Variety and Velocity .
Several studies have clearly highlighted these characteristics in particular Big Data Analytics (TDWI Research, August 2010),
Big Data: The next frontier for innovation, competition and productivity (McKinsey, June 2011), or even The Deciding Factor:
Big Data & Decision Making (Capgemini, June 2012).
27

1.3 An innovation or a revolution?


Two additional characteristics should be included:
Variability of data (the eects of peaks or troughs and seasonality in the quantity of data produced.
Complexity (for example, the number of sources from which data originates, its diversity figures, diagrams, relational
sets, text or its format).
The exponential growth in
the quantity of data

580
annual
growth

75

2009

67%

2007

0,8 ZO*

2020
35 ZO*

2011

The number of exabytes produced per year.

Projected growth in the digital universe, 2009-2020

Source IDC 2008

Source IDC Digital Universe Study,


sponsored by EMC, May 2010

1 exabyte = 1 billion gigabytes


1 zettabyte = 1,000 trillion gigabytes
According to IDC, the total amount of information
produced by companies saw an annual increase from
2007 to 2011 of 67 percent, rising from 75 exabytes to
580 exabytes.

But this growth is nothing compared to what is awaiting


us. The amount of digital information created or
replicated each year (what IDC calls the digital universe)
will be multiplied by 44 in 2020 (compared to its 2009
level) to reach 35 zettabytes.

Source : Qualit des donnes, quelle(s) vrit(s) pour les entreprises,


EBG, 2012

4. http://www.sas.com/big-data/

28

1.3 An innovation or a revolution?


Big data can, if required, be handled by conventional BI tools, but they are insucient for truly understanding it.
Data visualization, however, bridges the gap by introducing a fourth V (value) that Philippe Niewburg sets out in the following
formula:

Velocity
+
Volume
+
Variety

x Visualization = Value

Data visualization to the rescue for


handling big data
Data visualization solutions are a response to the complexity encountered when
handling large volumes of data. In this sense, visualization itself allows a targeted analysis which passes
through pattern detection (those recurring diagrams reflecting trends), impossible to detect using traditional
reporting tools (e.g., an Excel dashboard). In other words, dataviz allows large volumes of data to be explored
visually and trends to be highlighted that are not necessarily visible by just listing figures.

Frdrique Pain,
Experience Strategy Director, Alcatel-Lucent Bell Labs

The challenge we have with big data is information that is more and more voluminous, and therefore we
really must address the complexity for it to stand out. We also have a challenge with big data in how it is used
for enhancing analysis. Visualization must rightly allow a benefit to be obtained by allowing new questions
about the data to emerge.

Fabrice Benaut,
CIO global IFR, head of IS and Operations, IFR Monitoring

29

1.3 An innovation or a revolution?


As these examples show, it is analysis that is at the heart of creating value from big data. Data visualization provides value
because it enables exploration of these enormous data sets.

Data visualization, a tool that


democratizes analysis
Within our profession, the term analysis refers to statistical methods of analyzing data:
descriptive statistics, predictive analysis, optimization, forecasting, scenario testing, etc. With
regard to business analytics, it is how analysis is applied in the corporate world. More specifically,
it refers to tools such as those for calculating correlations, regressions, analysis onto main
components, factor analysis, neural networks, etc.
Business analytics aects all professions. Lets look at two examples. By studying the type of
clients and their behavior, predictive analysis allows a direct marketing company to calculate the
likelihood that a certain email will lead to a particular individual making a purchase. This is how
the company knows which individuals to target with the email.
Another example: the marketing of new medicines. Numerous tests take place from when a new
substance is developed to when the medicine itself is placed on the market. The aim is to ensure
that the new product really adds value, whether in terms of eectiveness, tolerance, ease of use,
etc. It is theory testing that plays a part here.
It should be noted that analytics not only applies to large companies. It applies to all companies
and business divisions where data plays a large role. For example, analytics will be fundamental
for a small-to-medium size social gaming company (social media-based games) because
analysis of usage logs for its games is the only way to understand its customers and, hence, its
Jrme Cornillet,
Business Solutions Manager
SAS
30

1.3 An innovation or a revolution?

1
An example of data visualization in
pharmaceutical research

Matters of patient health and safety require strong oversight in order to identify risks and avoid public health disasters. The
intensity of required monitoring is constantly rising. Monitoring the safety and ecacy of medicines in the approval phase
and once they are in the marketplace is critical. Measures in the field require careful and thorough analysis.

Source: SAS

Analysis of side eects

This graph displays the frequency and severity of certain side eects. In particular, the correlation matrix shows a strong link
between a specific medicine, Thiazolidinedione, and chest pain.
To illustrate the importance of analytics in the pharmaceutical industry and the role that data visualization can play, think about
the role of regulatory oversight for patient health and safety in order to identify potential risks and avoid public health problems.

31

1.3 An innovation or a revolution?


Accentures Andrew Musselman confirms big data visualizations added value in the field of analytics.

Analytics, data visualization and big


data: The winning combination
The data visualization department has collaborated on commercial projects and
proposals with the analytics department within Accenture to create a more modern approach with regards to
big data. It seems that these two aspects are more linked than in the past - not only by eliminating duplicated
tasks, but, more importantly, by capitalizing on the advantages of each. Big data has much to oer analytics,
and vice versa.

Andrew Musselman,
Data Scientist, Engineer and Architect in Accenture's Big Data Department dAccenture
The movement footprint (as exhibited at the Architecture and Heritage Exhibition, from April to August 20127).

7. http://www.orange.com/m_fr/sponsoring/culture/Cite-de-l-architecture-et-du-patrimoine

32

1.3 An innovation or a revolution?

1
The movement footprint, at the crossroads
between art, business and big data

Our work with data visualization is part of our overall big data approach. In this context, implementing
dedicated tools enables processing large volumes of data (and specific operator data). Even though data
visualization is a tool for evaluating data that can be used independently of big data, it is clear to us that the
two are closely linked.
Our projects are aimed at customers with the view to oering them, in time, products or services in line with
their requirements. An initiative developed at group level, l'empreinte de movement (movement footprint),
analyzes and adds value to movement information collected for one or more individuals.
With this approach, we start from physical points and initially integrate a short-term window with data being
captured at regular intervals (for example, every seven minutes). The result is a type of surface area or
footprint that we are able to represent visually.
This visualization reflects the degree of inactivity (one way of reading it) and highlights the ground covered,
but also shows the habits of the individual from the dierent routes. The tool uses logarithmic representation
in such a way as not to obscure the details. By increasing the observation time (over several days or weeks)
and including a longer-term window, we increase the relevance of the individual footprint. The resulting
analysis is enhanced in the same way.

Catherine Ramus,
Team Design Project Leader, Orange

33

1.3 An innovation or a revolution?


The growth in big data has made data visualization a necessity. In return, interest in data visualization is growing because of its
ability to analyze and explore big data.
At first sight, explains Mouloud Dey, Director for Solutions and Emerging Markets at SAS, there often appears to be no
connection between data visualization and big data, but we view it dierently: These two sectors are well and truly linked. In
fact, data visualization is particularly interesting when you have large volumes of data that you need to give context to in order
to make sense of it."

Progress in understanding how the brain works


Today, science allows us to understand the human brain in much more detail. It reveals that data visualization is a particularly
eective method for representing, analyzing and interpreting data.

Data visualization, a tailor-made


tool for your brain

Your vision, controlled by the visual cortex located at the back of the brain, is very
fast and eective. You can see things immediately with little eort. Conscious thought,

which is controlled mainly by the cerebral cortex at the front of the brain, is a lot slower and less eective.
The conventional way of representing and making sense of data requires conscious thought for almost the whole
process. However, data visualization shifts the performance balance to the visual side of perception, taking
advantage of the speed of our eyes as much as possible.
Source :
Few, Stephen (2013): Data Visualization for Human Perception.
In: Soegaard, Mads and Dam, Rikke Friis (eds.).
The Encyclopedia of Human-Computer Interaction, 2nd Ed.
Aarhus, Denmark: The Interaction Design Foundation. Available online at http://www.interaction-design.org/encyclopedia/
data_visualization_for_human_perception.html

34

1.3 An innovation or a revolution?


As explained by Noah Ilinsky8 in a LinkedIn Talks conference9 posted to the Web in April 2012, "the brain is primed to recognize
recurring patterns. These patterns are a way of comparing dierent information by assessing the contrasting elements. In fact,
the mind is much more capable of identifying trends, dierences or changes than we once believed."10

New interfaces for accessing information


Data visualization is without a doubt the way to analyze and explore data in the post-PC era. The means for accessing
information are changing. While IDC claims that global sales of computers have plummeted by 14 percent in the first quarter of
201311, Gartner forecasts that sales of tablets will overtake those of PCs by 201712.
Current data visualization methods make the best possible use of tablets' touch-screen interfaces by allowing graphics to be
manipulated directly, dierent views to be selected, filters to be applied, etc. This table interactivity is now a mandate by data
visualization users, as related by a revealing anecdote from Pierre-Olivier Sicamois from Lagardre Active: "The president of the
board gave us only one constraint: The solution has to work on the iPad.

8. Holding a master's degree in communication technology from the University of Washington and a degree in physics from the Reed
Faculty, Noah Illinski (@noahl) is an IT developer who focuses on how to eectively approach visual representations of information.
9. http://www.youtube.com/watch?v=R-oiKt7bUU8
10. The full conference is available at: http://fr.slideshare.net/Cooperatique/noah-iliinsky-prsente-la-dataviz
11. http://www.01net.com/editorial/592687/chute-historique-des-ventes-mondiales-de-pc-au-1er-tri-mestre-2013/.
12. http://www.journaldunet.com/ebusiness/internet-mobile/les-ventes-de-tablettes-depasseraient-celles-des-pc-0413.shtml

35

1.3 An innovation or a revolution?


Data visualization on a tablet, the data analysis
tool that CEOs were waiting for
As soon as we show executives SAS Visual Analytics on tablets, they
unanimously agree that this is what they needed. Executives do not want to go into the data in depth; they
want intelligent dashboards at their disposal that can be used anywhere, any time.
Ill use our steering committee as an example. Previously, they had a portal containing all available
dashboards: progress of sales, turnover, quality of customer service, etc. Today, they can access all of these
performance indicators on the iPad. Updates on the SAS Visual Analytics server are automatically
synchronized on my iPad. If I want to analyze statuses from home in the evening, I can. As of now, we are all
free and mobile, while remaining optimally connected to the data that allows us to run the company.

Serge Boulet, Director of Marketing and Communication,


SAS France
In some cases, having a small number of employees equipped with unique hardware can seed interest for data visualization
throughout the organization. This is what happened, for example, at ID Logistics.

When the iPad acts as a


gateway to data visualization

The spread of our data visualization solutions came about in a trivial way, due to a
restricted group (that we want to expand). That is, those who use an iPad. And more

specifically, members of the operations and general management teams.


Emmanuel Vexlard, DSI,
ID Logistics
36

Example of data visualization on the iPad at ID Logistics.

1.3 An innovation or a revolution?

And this is just a start. The progress in making more and more intuitive man-machine interfaces (MMIs) is opening new
horizons for data visualization.
Renewed interest in MMIs - a source of
innovation for data visualization

For us, data visualization has a bright future, in particular with the soon-tocome touch tiles, which promises even more interactive graphical

representations, and the boom that tablets and 3-D are bound to create. These technological advances will
change the ways of representing things. As proof, INRIA is currently experimenting with multiscreen
visualization, involving the user's gestures.

Frdrique Pain, Experience Strategy Director,


Alcatel-Lucent Bell Labs
37

1
1.4 For whom is data visualization intended?
Lets clarify a point of confusion data visualization is not just for big data. While data visualization makes sense when dealing
with large volumes of data, it would be wrong to suggest that data visualization is only for large organizations or very large data
sets. As highlighted by SAS Jrme Cornillet (in section 1.3), smaller organizations are just as likely to process large volumes of
data and want to do so in the easiest way possible, without having to bring in statistical analysis experts.
If you look at the profiles of those who use data visualization, you will notice that they are wide-ranging, but they can be
grouped into three broad categories:
The general public.
Internal stakeholders (management, workers or service providers).
External partners.
Some use cases of data visualization by the general public have already been mentioned e.g. France Tlvision's Pariteur
app and data access for citizens provided by the Belgian Ministry for the Economy. In these examples, you can see that the same
application can be targeted at dierent audiences.
This is particularly true with data provided by the Belgian FPS (Federal Public Service) Economy: "On the one hand, our data is
available for the general public who access it via an open data portal, and on the other for users at the National Statistics Oce,"
explains Caroline Denil.
The challenge is in determining the most appropriate methods of access and visualization for each group (more about this later).

38

1.4 For whom is data visualization intended?


Another example of data visualization for
the general public: Social Ecorama

The Social Ecorama

project provides visitors to the


BFM-Business website with an indicator of sentiments in
households and companies. This confidence indicator has
several original features.
It is produced by analyzing messages posted on social
media sites such as Twitter and Facebook. The technology
partners of this project, SAS and Inbox, collect and analyze
a large amount of text-based data in order to identify
positive or negative sentiments based on dierent criteria.
Social Ecorama13

For households, messages dealing with unemployment,


consumption, housing, purchasing power, etc., are monitored. With regard to companies, the key topics are
growth, profits, recruitment and exports (among other subjects). For each item, a "positivity indicator" is
calculated to estimate whether the sentiments expressed on the social media channels tend more toward
optimism or pessimism.
The second feature is that these indicators are followed daily, while conventional opinion polls are
undertaken with a frequency that is counted in months.
The third original feature of Social Ecorama is that the results are presented using data visualization. Internet
users are able to view how the sentiment is progressing for households and companies by way of line charts
by selecting the key words that they are interested in, as well as the period. The project was launched on April
3, 2013, and already months worth of data are available to explore. The combination of these three original
features makes Social Ecorama a global first.

Emmanuel Lechypre, Editor,


Head of the Analysis and Forecasting division, BFM Business

13. http://ecorama.inbox.fr

39

1.4 For whom is data visualization intended?


Data visualization adds value to InfoPro Digitals business processes:

Of course, data visualization allows our content to be evaluated, but above all I see definite value for
responding to three essential objectives.
With regard to audience, it is the way to make our content available to everyone (e.g., ranking engineering
schools) or our mapping of BTS.
Secondly, it is the ability to bring real added value to the way we manage our subscriptions, which is not
negligible in view of the challenging media landscape that we face.
Furthermore, it is no longer a question of traditional subscriptions to magazines, but real subscriptions to
information systems; With Industry Explorer, for example, we oer access to magazines, but we also oer a
file-by-file consultation of information about companies. This is how we are able to enhance our editorial
product while at the same time developing the revenue from our subscriptions.
Finally, we use dataviz to develop specific business lines. With the mapping feature, in particular, we have at
our fingertips a value that goes beyond the value of the subscription alone which was impossible with
conventional technological resources. For example, you can consult a map of all the factories in Aquitaine (a
province in France) that are planning to invest more than 10 million euros next year.

Clment Delpirou, partner, InfoPro Digital

40

1.4 For whom is data visualization intended?


Data visualization can be used at all levels of an organization, from management to teams in the field and beyond. Its use is
sometimes widespread, especially at huge multinational organizations like Philips.
Menno Hakima, Senior Global Manager at Philips, confirms this: "The majority of people involved are managers or business
leaders, whether in-country or at headquarters level."
As stated before, it is the ability to obtain (in real time and while on the move) an overall view of the organization's performance
indicators that generates the interest of an organization's management or board.

Making decisions
based on key
indicators

With business intelligence in mind, we have


implemented an iPad-based solution that is
meeting management's expectations
because it incorporates a dashboard with key
indicators that can facilitate analysis and
optimize the decision-making process.

Data visualization on the iPad at Prisma Media


Jrme Tharaud,
Head of Research and Development,
Prisma Media
41

1.4 For whom is data visualization intended?


Along the way, data visualization project managers sometimes run into pleasant surprises. Occasionally, information (in its
traditional format) can and has been totally ignored by management. Data visualization comes in action to highlight
information parts so they can finally be used.

Data visualization for sharing available information


that had remained unnoticed
With the arrival of our new general manager four or five months
ago, the head of our advertising department and the general manager organized a meeting of the Executive
Committee. During the meeting, we reviewed current projects. This is what we found:
"The advertising department uses multimedia, but the statistics showed no multimedia data." This comment
somewhat surprised us as, quite rightly, our reporting system was capable of highlighting such data by
means of the cubes that we had implemented. Given this finding, we committed ourselves to provide the
chairwoman and general management with these multimedia statistics within one month.

Pierre-Olivier Sicamois,
Head of Advertising Production, Technology Directorate,
Direction des Technologies Lagardre Active

42

1.4 For whom is data visualization intended?


Business divisions are taking special interest in data visualization because they are typically the first users. Without them, those
solutions could not be eectively implemented. The business divisions know the needs, the meaning of numbers and how to
interpret visualizations. Hence, they are both the consumers and the co-producers of the data visualization.

Business Division teams, users


with co-producers of data
visualization

In time, once our system has been put into production, it will be aimed at both
technical teams and business teams. In fact, it is the latter that contribute largely to

the development process because they are responsible for setting the parameters, qualifying the data and
feeding the engine. The system will help them with reporting and monitoring the quality of the service
oered to online users.

Eric Lajarige,
Head of Search Project, PagesJaunes

43

1.4 For whom is data visualization intended?

1
PagesJaunes automatically detecs badly categorized requests

PagesJaunes search engine allows


online users to search for professionals
across France by completing two fields: What and Who.
Classifying a request as What or Who is therefore particularly important.
Incorrectly classifying the requests significantly reduces the chances of a professional being found during a
search. If a visitor types "Acacias, Toulouse", it could be thought that they want to buy acacias. This means
that "acacias" are classified as an object (What). But what if the user was actually looking for the address of a
restaurant called "Les Acacias" in Toulouse? "Les Acacias" needs to be understood as the name of a company
(Who).
In one case, the expected response is a list of nurseries; in the other, the address of a correctly identified
restaurant. Knowing that the PagesJaunes group's value comes from oering visibility to professionals and
companies, this could be significant.
But, in the stream of daily requests, how do you detect unsuccessful requests that are the consequence of an
incorrect classification?
To achieve this, the agency 55 | fifty-five.com has developed an interface for PagesJaunes for exploring
requests. It allows a list of requests to be filtered according to dierent, interdependent criteria. Thanks to this
interface, it is now possible to visually identify requests for which there is a strong presumption of a
categorization error.

Martin Daniel,
Head of Data Visual Project, 55

44

1.4 For whom is data visualization intended?

1
PagesJaunes search analytics interface

(The closer the figures in the columns Anage or Localit are to 1, the greater the
probability that the professional has been assigned the wrong labels in the database.)

45

1.4 For whom is data visualization intended?


Most departments within an organization can benefit from data visualization: marketing, sales, finance, manufacturing, R&D,
etc. Laurence Kerjean of LOral, agrees.
Data visualization: Relevant
for all functions

Because the "base" data is quantitative, data visualization is a real means of exploiting it.
All business divisions or areas are aected. Further, even qualitative data can be put into

images; it just needs to be converted into numerical values (e.g., not happy equals one point, very satisfied
equals seven points). Data visualization is very eective for marketing in particular, as it allows you to present
quantitative data in a visual form to visual thinkers, and is therefore easier and quicker for them to digest.

Laurence Kerjean,
Global Digital Manager, LOral
Sometimes you can encounter business units not very attuned to the field. Until now, these business units have not had the
right reporting or analysis tools. Data visualization is able to bridge this gap.
Data visualization, a tool for understanding the
competitive environment

The solution is accessible to all employees, but actually it is above all

the marketing departments in Europe that use it, as well as development


studios around the world.
Charles du Rau,
Consumer & Market Knowledge Director,
Ubisoft

In order to analyze competition, our tools are aimed principally at the purchasing, sales and marketing
departments. In this regard, they provide targeted sector-based intelligence useful for defining our strategy.

Clment Delpirou, partner, InfoPro Digital

46

1.4 For whom is data visualization intended?


Other contributors focus less on this reporting aspect and instead emphasize understanding the business itself. Visual
representation provides an overview of how the processes making up the work come together, something that was not always
possible previously.
Better understanding and controlling the
business thanks to data visualization
For us, reporting is not the most innovative part of data visualization. Even if
it now allows us to filter more precisely and provide a summary, reporting already existed in conventional
decision-making (BI) IT tools. Our approach to data visualization focuses more on calling up, at any moment
of the purchasing cycle, all the information in the data repository (supplier accounts, item prices, catalogs by
suppliers and product groups, purchasing recommendations, information from the supply chain, etc.) in a
visual and interactive manner. This gives us a dashboard with an overview of several business divisions
processes that up until now were disjointed. For us, this is where real innovation lies.

Franck Jouenne,
Director, IT & Services Strategic Sourcing, Alcatel-Lucent
Thanks to its practical side, data visualization leaves the oce and brings the power of analytics to the field. Operations teams
are able to use it to make decisions quickly, as is the case with STMicroelectronics.

Data visualization, a tool


of choice in the industry

When implementing our dataviz solutions, we aim them at all of our employees: from the

typical worker, the engineers to the head of department. In this respect, a worker will more
likely use control cards and machine status charts, for example, while the head of department will be more
interested in performance trends.
Guillaume Deschamps,
FMT/Central Functions YTI/EDA, STMicroelectronics

47

1.4 For whom is data visualization intended?


In the manufacturing sector, data visualization enables certain complex technical situations to be understood and the best
decisions to be made. Engineers will find this useful and relevant.

Data visualization for


engineers
I feel that what dataviz has to oer can be particularly interesting for engineers in areas
where targeted expertise is required, but also for decision makers. The management of recovering energy
produced by the dierent trains is a particularly good illustration of this: A braking train produces energy that
is dicult to store. This is why it is better to use it as soon as it is produced by sending it to another train. The
ability to locate these recovery points on a map is significant: Being able to exactly visualize energy
recovered in the system helps to calculate electrical consumption and identify associated costs as well as set
out actions to be undertaken as a result. As it happens, data visualization allows our expert engineers to better
visualize at what point or in what zone they need to design a storage unit... This is also the way for those
who make decisions to identify those sectors where it is valuable to invest.

Maguelonne Chandesris,
Head of the Statistics Team,
Econometrics and Data Mining,
Innovation and Research Directorate SNCF

48

The energy sector is another industry where data visualization has turned out to be very useful.

Data visualization and the energy sector

Source: SAS

1.4 For whom is data visualization intended?

Managing Smart Meters


Deploying smart meters enables the optimization of electricity distribution in the network by closely monitoring trends and
profiles and forecasting demand (for predictive maintenance purposes). The volume of data to be analyzed is an important
consideration, but the speed and capacity of analysis is the overriding factor. This graph illustrates the consumption of
electricity according to dierent profiles (hourly and daily) as well as the continuous monitoring of the demand. The bubble
graph above shows consumption by geographic area.
49

1.4 For whom is data visualization intended?


As you have seen, data visualization can be used for multiple purposes or industry sectors. Data visualization is finding its way
into more and more organizations at an increasing pace.

Fraudsters beware here


comes data visualization!
Some employees may exhibit fraudulent behavior or contribute to a fraud scheme or an
organized fraud network (misappropriation or misuse of company benefits, insurance fraud, etc.). This type
of fraud is dicult to detect when you consider an isolated transaction such as repairing a vehicle or selling
eyeglasses, for example.
However, when you can correlate a large number of insured parties benefiting from the same fraud, it is
possible to highlight the collusive networks of involved parties. This is dicult, if not impossible, just by
studying spreadsheets of raw data.
A graphical representation allows you quickly to identify patterns, outliers or unusual occurrences that may
flag a potential fraud. Data visualization is used to detect fraud by private insurers as well as in the case of
public claims (unemployment claims, family benefits, etc.). Let me give you a flagrant example of a case that
happened with one of our clients. The County of Los Angeles came up with the idea of overlaying the graph
of data representing benefit claims with claimants addresses on the map. They were surprised to discover
that a high number of fraudulent claims coincided with certain residential areas. Thanks to the combination
of multiple dimensions (corporate data, data on insured parties, geographical details, even conversations on
social media), suspicious claims can be identified more quickly.

Mouloud Dey, Director of Solutions and


Emerging Markets, SAS France
50

1.4 For whom is data visualization intended?


Another target audience for data visualization is corporate clients. IFR GfK is a particularly good example. The group studies the
distribution of consumer goods using three methods:
Retailer panels.
Ad hoc surveys (questionnaires).
An inventory of products available on store shelves.
From these analyses, the main product delivered by IFR GfK to its customers is data. Although certain clients remain loyal to
traditional methods of reporting, an increasing number of corporate customers are eager to use more visual tools and analysis.
Use of data visualization techniques can be grouped in four clusters, as identified below by Frdrique Pain of Alcatel-Lucent.

ways to use
data
visualization

Data visualization generally is applied in four ways:

Analysis of data for the purposes of reporting.


Exploration of data for explaining the world (e.g., medical studies).
Optimization of processes (e.g., logistics).
Forecasting or predictive analytics, where the study is driven by the analysis of patterns for anticipating
future trends.

Frdrique Pain, Experience Strategy Director,


Alcatel-Lucent Bell Labs
If data visualization can be applied to so many types of organizations, it is because it oers a broad range of business benefits.
Lets explore them in the second part of this book.

51

2
2

The benefits of data


visualization
The wide range of sectors and use cases represented by our panel converge on a clearly identified
set of benefits of data visualization for their organizations, their users or their customers.
This common areas include:
Understanding of data.
Communicating of data.
The decision-making process.
User motivation.
The ability to innovate.

52

2
2.1 Data visualization makes data easier to understand
From a theoretical standpoint, the example of Anscombe's quartet (in part 1) illustrated that a spreadsheet of raw data is not
always easy to interpret and that a graph often speaks for itself.
From a practical point of view, the panel we interviewed confirms the theory: When our goal is to give data a meaning, then
graphical representation is a major ally.

Using data visualization to understand the


behavior of advertisers
We have chosen to use dataviz to simplify access to information and
bridge gaps in our understanding of certain data. In fact, reading spreadsheets is useful when the numbers
themselves are of overriding importance (e.g., turnover). Otherwise a pie chart says more. It allows us to
visualize the distribution of dierent media, the predominance of one media over another or even to
visualize the spread of these media by type of advertisers.

Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active

53

2.1 Data visualization makes data easier to understand


Data visualizations advantage over raw data in a spreadsheet is intensified by two factors - the increasing complexity of the
corporate environment and the wide variety of available data.

Simultaneously and intelligently visualizing


dierent types of data
It is not unusual for a customer query to return up to eighty route
alternatives. It is up to us to define the optimal solution for the client in this maze of alternatives. For
example, we need to determine the best way to represent timetables, prices and geographical data - all at the
same time.

Frdrique Ville,
Innovation Director, Voyages-sncf.com

Mytripset, Voyages-sncf.com

54

2.1 Data visualization makes data easier to understand

2
Mytripset, Voyagessncf.com

Our unique advantage is that we can simultaneously aggregate and visualize data from dierent
structures production data (trains) and commercial data (how many empty seats are left in the trains?). We
are making the complexity of the system visible.1

Maguelonne Chandesris,
Head of the Statistics Team, Econometrics and Data Mining,
Innovation and Research Directorate, SNCF

We created new information that was not previously available, just by combining dierent data sources.

Yves Daelmans,
ICT Project Manager, IMEC
Data visualization allows complexity to be represented and interpreted more easily interpreted in order to extract the most value
from it. This is what Bernard Lebelle, founder of the Bleu-Cobalt agency, explained in his books Convaincre avec des
graphiques ecaces (Convincing with eective graphs) (Eyrolles, 2009) and Dites-le en images: des ides? Un crayon! (Say it
with pictures: Any ideas? A pencil!) (Eyrolles, 2013).

1. For more details and examples, go to http://senseable.mit.edu/trainsofdata/

55

2.1 Data visualization makes data easier to understand


Controlling complexity in order to identify areas of growth

Visualizing complex dimensions (time, geographic, relational) allows us to better identify growth areas of
productivity, competitiveness or margin. In the above example, we have four elements (turnover, market
segments, changes over time and two dierent stakeholders) represented on a single graph. By enhancing
and optimizing the way data is represented graphically, we avoid information overload (where the addition
of too much numerical data to new dimensions overloads the mind's ability to process it). Suddenly,
significant amounts of multidimensional data is easy to interpret.

Bernard Lebelle, founder of Bleu-Cobalt

56

2.1 Data visualization makes data easier to understand


The economy is complex and moving fast. You need to tackle this complexity, faster.
Understanding
more quickly
When everything goes faster and we no longer have the time to read long reports full of details,
data visualization is a way to understand and analyze information more quickly. It allows us to benefit
immediately from a global view of the company's key indicators.

Emmanuel Vexlard, DSI,


ID Logistics

Confronted with this new constraint of speed and complexity, organizations need to face facts: it is unrealistic to put all data
into equations or to rationalize everything. The real world cannot be put into theoretical models. Data visualization, however, is
well suited to observation without theoretical preconditions.
Claude-Henri Mldo said, "Data visualization techniques, particularly with regard to big data, brought something new. We have
seen the emergence of graphs that not only come after the analysis - for the purpose of explaining or communicating - but that
come before them - for exploring it."

57

2.1 Data visualization makes data easier to understand

2
Discovering data

With traditional BI, the user is in query mode (selection, then data retrieval) and needs to ask the
question first. In this sense, traditional BI 'reduces the data' in a certain way and may deprive the user of
information that they would have spotted by visually retrieving all of data.
With traditional tools, something implausible may not be readily apparent, and we run the risk of missing it.
If we look at things in the context of big data, this is even more striking. However, with data visualization, the
user is in a full discovery mode of all the data. An exploratory analysis can be undertaken we can 'walk
through the data.'
Lets imagine a data series that includes some misleading data; perhaps its a problem with data quality or
because it highlights an exception that is worth noticing. I challenge you to identify it with a traditional BI
tool unless you develop multiple queries. With a visual interface, this particular piece of data will grab your
attention almost instantly.
Data visualization brings a lot to the table. A number of our clients already have a lot of experience with BI or
statistics and, for them, data visualization oers an opportunity to revisit business questions they would have
missed with traditional BI. Really, a whole new world opens up for them.

Serge Boulet, Director of Marketing and Communication,


SAS France

58

2.1 Data visualization makes data easier to understand

2
One example of data visualization in
pharmaceutical research.

Source: SAS

Data visualization and purchasing dashboard

This screenshot shows part of a purchasing dashboard. It displays


how the gross margin and cost developed over time. The treemap
shows the distribution of parts cost by supplier.

59

2.1 Data visualization makes data easier to understand

2
Two examples of monitoring enhanced
by data visualization

Source: SAS

Data visualization and financial dashboard

The graph at top shows part of a financial dashboard. It illustrates how


revenue has developed over two years (actual and forecast). The histogram
shows how monthly profit has developed over the same period.
Lets discover how several organizations plan to use data visualization to explore their data.

60

2.1 Data visualization makes data easier to understand

2
Data visualization for
exploring trends

Dataviz is, without doubt, a targeted way of mining data, which is of increasing interest for
marketing: filtering or the ability to zoom into a time, for example, are ways of exploring trends. They allow
assumptions to be validated and also experiments to be performed with regard to the user experience.

Frdrique Pain, Experience Strategy Director,


Alcatel-Lucent Bell Labs

Ideally, we would love to highlight information that was invisible without data visualization. Eventually, our
aim is to extract trends from huge volumes of data. Trends that would be dicult to identify without visual
representation. We would like to explore this in more depth, more specifically for our marketing data.

Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active
With its ability to highlight atypical data and make it stand out, data visualization, when used for data exploration, is also an
excellent way of identifying problems and fixing them.

61

2.1 Data visualization makes data easier to understand

2
Exploring data to fix
a problem

We needed to explore our data set in order to better understand the relevance of the results we
send to online users. To date and in order to improve the quality of our indexing we only had statistical
information available. It was dicult to exploit when you take into account that we process tens of millions
of requests each month. With this volume it is particularly dicult to identify requests to which our search
engine gave an incorrect response. A tool like data visualization will give us another way of detecting
problems more quickly.

Eric Lajarige,
Head of Search Projects, PagesJaunes

Data visualization allows the mining of structured or unstructured data that up until now has remained untapped and poorly
understood. This is all the more important if you look at the explosion in data volume, especially with the increasing use of
social media.

62

2.1 Data visualization makes data easier to understand

2
Data visualization for mining online
content and social media

The case of content published on the Web is speaks volumes: the


analysis and linking of themes, words, sentences and documents (and potentially, behind all this, links
between individuals) is a real gold mine, not only for marketing departments but also those involved in
economic intelligence in general.
More specifically, there are links that we are unable to see with the naked eye but that we can identify using
sophisticated graphical analysis. In particular, this is the case with social graph analysis that allow noninsignificant links between individuals to be highlighted.
For me, this is one of the main contributions of data visualization. To put it another way, we need dataviz to
highlight what cannot be seen straight away, and the problems with big data are going to reinforce this need.

Mouloud Dey, Director of Solutions and


Emerging Markets, SAS France

Its clear data visualization is inviting us to take another look at the traditional distinction between data, information and
knowledge. If data is raw units that reflect reality, then transforming it into information gives it coherence and meaning.

63

2.1 Data visualization makes data easier to understand

2
Data visualization and the decision-making process the benefits of an analytical approach

In simple terms, the Anscombe example1 perfectly illustrates the benefits of data visualization
using an approach that could be regarded as analytical (with the additional need of
implementing solutions in a decision-making process based on data). In the example, data
visualization calls for reconsidering the very (too) simple linear adjustment model which is
clearly insucient for describing and explaining observed data in a suciently relevant manner,
and also to look for more precise and suitable alternative models.
There are some simple terms for expressing these additional approaches HINDSIGHT,
INSIGHT and FORESIGHT. These form the main pillars of an experimental approach suited to an
analytics-based decision-making process aimed at giving meaning to data and justifying
optimum decisions.
A recent study showed that during the last 10 years, companies in the S&P 500 who use analytics
at the heart of their decision-making processes outperform their competitors by about 64
percent. The use of analytical approaches focused on data are in fact finding their worth in a
growing number of professional processes that produce stunning results.
Revenue management, also known as yield management, has been widely used by airlines and
Mouloud Dey, Director of Solutions and
Emerging Markets, SAS France
Data visualization allows users to take their information, process it and extract its value. But to maximize this value, it has to be
shared throughout the organization. This is the subject of the next chapter.
1.See I.1, Definitions of dataviz
2.See I.2, inset "Dataviz for developing SNCF's product in dierent fields.

64

2
2.2 Data visualization improves communication
Data is not just a tool for managing the activities of an organization. It is also a communication tool. Unfortunately, not many of
us speak the language. Most of us need an interpreter to help get at the intelligence held in the data.
This is where data visualization comes in.
Data visualization can be used for communicating issues both internally and externally within companies and public
agencies, as well as government. This is why SPF conomie uses it when communicating with the Belgian public.

Improving communication
with the public
The benefits of our dataviz program are felt by the public at large. They do not just
need graphs showing progress (e.g., progress of those liable for VAT) but also to be able to export their data as
they see fit. This is what we are now able to oer them.

Caroline Denil,
Data Warehouse Program Manager, SPF Economie

65

2.2 Data visualization improves communication

Data visualization can be used in many ways. For example, it can be used to navigate a map to
find the mean or median price of housing and export the results to a spreadsheet.1

Managers dont enjoy tables of complex figures any more than the general public does.
In fact, senior management doesnt want to leaf through long, complicated reports.
What they need is for the information to be set out clearly and accurately to allow them
to make decisions. This is where the graphical representation of data finds its home.

Reports that are easier to understand for


busy decision makers.
Our approach to data visualization is seen in our continued eorts with regard
to business intelligence. An important component: making reports and data sexy. In other words, making
them easier to understand. At the moment, our eorts are aimed at making data better looking and easier to
access. Actually more accessible to people who have little time to really get a grip on the content of the
information. For this reason, we say that currently our dataviz is mainly targeted at reporting.

Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active
1.http://economie.fgov.be/fr/statistiques/chires/economie/construction_industrie/immo/prix_moyen_appartements/

66

2.2 Data visualization improves communication


Organizations need to apply good visualization principles when communicating with clients who are also faced with problems
of limited time and resources. In this context, data visualization is the logical continuation of the BI story.
Data visualization, the next step in data
communications
Data visualization is the latest chapter in a long history of reporting. In
the not-too-distant past, the data we collected was stored on paper, and we used to send files to our clients. A
whole room was set aside for our archives.
The next evolution consisted of communicating this data in a digital format. Since then, we have started to
convert the data into graphs to make it more digestible.
The current stage is one where we need to collect data that is more and more complex and bulky. At the
same time, our clients have less and less time to process the information we provide them, with teams that
are smaller and smaller.
Three years ago, and particularly since 2013, this lead us to introduce a new way of accessing our data in the
form of data visualization. This service is already available to all of our clients around the world.

Fabrice Benaut, CIO Global IFR,


Head of IS and Operations, IFR Monitoring

When data is well understood and properly communicated, it becomes a catalyst for action. You will see how in chapter 2.3.

67

Data visualization optimizes and accelerates the


2.3
decision-making process
As you read earlier, data visualization is often used for management purposes. This is easy to understand if you consider the
flexibility and eectiveness that it oers during the decision-making process. Jrme Tharaud of Prisma Media agrees: "There is
no question about the contribution made by dataviz, whether in terms of business strategy or decision making at the highest
level."
The reason is simple: Thanks to these new tools, managers have rapid access to easy-to-understand information that allows
them to make decisions.
Analyses on request for
making decisions
Now, whatever type of analysis you want to do is done at the click of a button. One or
two clicks is enough to obtain the desired result without needing to send a series of requests.

Sandrine Noail, Head of Control and Analysis,


Coliposte

We are now able to extract data and pull out the indicators from it at the last minute before our managers
need it, whereas before, this required a protracted process.
Yves Daelmans,
ICT Project Manager, IMEC
68

2.3 Data visualization optimizes and accelerates the decision-making


Making analyses available much more quickly is one of the most significant returns on investment of data visualization, as we
can see in this example from Pages Jaunes.
An ROI measured in time
savings
We measure the return on investment of our approach in time freed up for our sta.
Previously, managers used to put half a day or even a full day aside at the start of the week to obtain the
production figures. Now, the reports are done in two minutes. The return adds up fast. In addition, managers
can have this information updated at any moment and not just once a week.

Richard Core, Senior Business Analyst,


PagesJaunes
Decision makers now have useful KPIs at their disposal. They no longer need to ask for them or order them in advance.

Data visualization: key indicators at


management's fingertips
Whereas traditionally managers are not too keen on using tools, they recognize
the benefits of dataviz for two reasons. First, it is perfectly in line with the company's sales action plan and
strategy. Second, it is easy to use and they can use it without assistance.

Sandrine Noail, Head of Control and Analysis,


Coliposte

69

2
2.4 Data visualization helps motivate staff
Data visualization is becoming part of a movement towards worker autonomy, improved working conditions and accountability
as expected by employees, according to Yves Daelmans, ICT Project Manager at IMEC: "Users appreciate being able to edit
reports by themselves."
This is even truer for digital natives, who are accustomed to being creative in their private lives using digital tools and dont
understand why they cant in the workplace. This explains the emergence of a new role within organizations, the power user.
Not specialized in BI, but more operationally focused, they use data visualization tools to develop real expertise that everyone
can benefit from.

The role of the


power user

The new BI that data visualization allows provides users with a high level of self-

suciency. In this context, one type of player has a major role the power user. For me, the power user has
two main characteristics: autonomy and expertise. The power user is a member of sta within an operational
department who has sucient command of BI or analysis tools to be able to manage on their own due to the
previous experience or because its their passion and they made themselves an expert. In other words, they
know how to do a lot because of their job type or because they learned on the job due to circumstances. It is
someone who does not require any assistance and instead is able to help others.

Serge Boulet, Director of Marketing and Communications,


SAS France

70

2.4 Data visualization helps motivate staff


The case of IMEC's research center is an interesting one because power users are integrated into a well-defined process.
Power users: Complete autonomy at a local level, integration
into the IT process in the event of wider use
Our BI skills center is comprised of 36 power users. They
have access to the source data in our data warehouse and are able to produce their own reports that they can
disseminate to a restricted audience. If they find that these reports could be used by a wider audience within
IMEC, they send them to IT and use the traditional processes of development and dissemination.

Yves Daelmans,
ICT Project Manager, IMEC
But the benefit to work teams is not only for power users. When applied to management (individual performance indicators,
monitoring of annual objectives, etc.), data visualization improves transparency. Managers, executives, foremen, workers and
employees are able to use the same dashboards when holding discussions and adjust their focus quickly if required. In a real
sense, data visualization becomes a tool for corporate dialogue.
Data visualization as a
management tool
The faster and more frequent creation of reports allowed by dataviz contributes to the local
management that we wanted to install. Where the manager has more information on the activity of those he
or she manages, and where this information is reliable, both parties find it easier to talk. It creates trust. This
is particularly noticeable with the continual monitoring of annual objectives for each member of sta over
the year.
Richard Core, Senior Business Analyst,
PagesJaunes
71

2.4 Data visualization helps motivate staff


Because data visualization can be made available very quickly and because it is of immediate use to the teams, it is a way for
managers to encourage employees. It gives them tools that are easy to use and that helps them with their work in a meaningful
way. This type of encouragement is even more important when the organization is faced with dicult issues such as a
reorganization or a competitive struggle. These situations put the teams under pressure, and providing them with a resource
such as data visualization shows sta they are valued.

Data visualization, helping to reassure


teams during turbulent times
Another benefit is that these very practical tools that can be used directly by the
teams have helped reassure sta during the phase of reorganization that we went through. That of course
causes concern, as do all such periods. Making fast and usable data visualization available has allowed
everyone to be shown that even if reorganization involved changes to practices that were not necessarily
easy to manage, the management could attend to the teams by trying to provide them with the tools to
improve their work life.
Richard Core, Senior Business Analyst,
PagesJaunes

Ultimately, when teams have access to data and they see that it is useful, they will feel that their management is moving in the
right direction. The feeling of empowerment that comes out of it tends to motivate them.

72

2
2.5 Data visualization encourages innovation
This chapter on the advantages of data visualization would be incomplete without looking at its potential for innovation. In fact,
data visualization can lead an organization to look at new possibilities such as providing new ways of interacting with users.

Data visualization gives technology accessibility


and meaning
The activity of the Alcatel Lucent group involves varied skills such as
creating algorithms or computer science. When the end user is involved, we move into the field of user
experience. This promotes thought about technology and how to give it meaning. Its about making the
content of this technology accessible to users. In other words, transforming it into useful services with
financial value.
This is where our data visualization activities come in. And, as part of this, I make my skills as an industrial
designer and ergonomist available to enhance the user experience. This is why I regard data visualization as
a means of giving meaning to data analysis by making it more accessible to users.

Frdrique Pain, Experience Strategy Director,


Alcatel-Lucent Bell Labs

73

2.5 Data visualization encourages innovation

2
A project led by the Acceleration Platform of Alcatel-Lucent Bell Labs representing
mobile applications used over time and by many users

Researchers are at the front line of this innovative work, and they make it a point of pride to bring data visualization out of the
labs to demonstrate that it has practical application.

Pragmatic innovation
The aim of our work is not just to look into the future, but above all to provide internal support to
fixed-term operational solutions. In this context, dataviz allows:
Insight by revealing the potential wealth of data mining. Here, what we are doing is really educational.
Food for thought by opening up the range of possibilities.
Forecasting applications and interfaces for tools to help the decision making of tomorrow.

Maguelonne Chandesris, Head of the Statistics Team, Econometrics and Data Mining, Innovation and Research Directorate, SNCF

This advanced approach helps spread innovation within a team and provides ideas for future customer oerings.
74

3
3

Implementing data
visualization projects
Now that you have seen what data visualization is and its uses and advantages, next you will read
about the best ways to implement it. In fact, shifting to data visualization is a project in its own
right with all the normal demands of any other large project, but with two built-in advantages
speed and ease of use. In this respect, data visualization fits in well with flexible project
management methods, as the panel will explain.
You will also find that data visualization can, in certain cases, shift focus between the IT
department and business divisions (the panel will provide tips for a making it a win/win
relationship).
This IT department and business division partnership is one of the success factors for a data
visualization project. You will also read about other keys to success as well as the pitfalls that may
crop up (and ways of overcoming them).

75

3
3.1 Fast and easy projects
One of the main attractions of data visualization projects is that they are, on the whole, quick to implement. Their advantage is
in reducing the time between launching the project and debuting the first operational version.

First versions quickly


available
We produced an initial base version in less than five weeks in order to prove the
commercial directorate's interest in the procedure. Boosted by the positive reactions received, we kicked o
the project based on 20 days of production in April/May 2013. All this managed by a small, but expert, team of
only three people.
Sandrine Noail, Head of Control and Analysis,
Coliposte

Within two weeks, we had designed version zero of our solution.


Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active

76

3.1 Fast and easy projects


Several factors are behind this speedy space. One is that data visualization projects use material and software resources that
have little eect on existing architectures. They do not need lengthy financial evaluation, and the necessary equipment is
readily available.

Few technical resources


required
Recent technological developments, such as the development of JSON - type formats
(DS.JS3), put us in direct contact with the data. Thanks to these formats, we are able to retrieve varied data
from all sectors by using standard applications. A real human/data interface is in the process of being
installed. Our targeted Web or design resources are capable (without the need to be a developer) of very
quickly making use of the technology.

Frdrique Pain,
Experience Strategy Director, Alcatel-Lucent Bell Labs

Another reason is that data visualization lends itself to proof-of-concept experimentation and test-error iteration procedures.
Data visualization projects have an empirical side to them completely in sync with current agile development approaches such
as SCRUM.

77

3.1 Fast and easy projects

3
Agility at the heart of the
approach

The project was spread over a year. We had a flexible approach to meeting the users during
the exploratory phase. Then we imagined dierent solutions. Over the last quarter, we took on a continued
improvement approach.
With regards to the flexible method, I will add that I chose to deliver quickly and often. As for us, as soon as
we had something ready, we showed it to the users so that it could be used straightaway. It is not just an
eective method for obtaining feedback and amending the solution, it is also a powerful way of motivating.
In other words, the flexible methodology and the mindset of the project team (a real team spirit) really
contributed to the success.

Richard Core, Senior Business Analyst,


PagesJaunes

With regards to procedures, we chose to innovate in the sense that we reduced the traditional V cycle. By
their nature, these projects are shorter, and we already have the necessary data at our disposal. We opted for a
more collaborative working method with a system of evolving deliveries and reviews.

Franois NGUYEN,
Director for IT and Relationship Marketing, SFR

78

3
3.2 The impact of data visualization on the relationships
between IT and business divisions
Data visualization projects have this unique feature, as you saw earlier, in that they oer users a large degree of autonomy.
This is why the relationships between IT and the business divisions must be nurtured.
Friction sometimes occurs when the IT department feels it is giving up areas of competency.
Business divisions, are naturally
quicker with data visualization topics

The IT department, who traditionally had a hand in the decision-making


process for BI projects, found it hard to watch the business divisions develop

decision-making tools that were just as successful. We are making a lot more ground than they are with
these topics. The IT department is unable to be as successful as the business division with a data
visualization tool because it does not have the professional knowledge. An IT department still needs to
receive a requirements document, then write a specification, while the business division, by virtue of its
experience and expert knowledge of the benchmark data, is aware of everything and is able to produce
things easier and without restriction.
Sandrine Noail, Head of Control and Analysis,
Coliposte
Unfortunately, these tensions have been fed by certain suppliers of data visualization solutions that only focused on the
business divisions and bypassed the IT departments in order to gain sales.

79

3.2 The impact of data visualization on the relationships between IT

3
The risk of
circumventing IT

Certain solution providers show up with tools that can fulfill the requests from business
divisions that the organizations IT department is unable to provide. Because investment costs

were not high, the business divisions were able to equip themselves using their own budgets.
The problem with this practice is that it contributes to the phenomenon of shadow IT with the gradual
development of parallel IT departments beyond the control of, and potentially incompatible with, the overall
IT system.
Serge Boulet, Director of Marketing and Communication,
SAS France
Resistance and resentment can arise, particularly within traditional BI teams who could feel unsettled by this new approach to
BI.
Data visualization benefits from traditional BI;
it does not call it into question
We actually encountered minor problems internally within my own
team. This was because we attempted to break from traditional BI without bringing it into question but
instead by focusing on what we had obtained in order to make us stand out . . . We had to explain that it was a
way of exploring a voluntary break with existing BI tools without eliminating what we have obtained. In
reality, it is a question of focusing on what we currently have and complementing it at the same time. And in
order to calm growing concerns, we insisted that the project be based on the foundations of BI.

Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active
80

3.2 The impact of data visualization on the relationships between IT


Fabrice Benaut of IFR-GfK agreed, saying that the requirements for traditional reporting and those for data visualization will
continue to coexist.
Data visualization does not render other
forms of reporting obsolete
We are living in a hybrid world. We still have clients who expect readymade dashboards to manage their operations. These are clients who still expect the same type of report and
to use it in pretty much the same way, for management reasons, to validate the process. The other emerging
BI approach is to answer questions that have still not been asked. The trick is to draft answers for all
requirements: Some are very static, simple and even basic; others are more advanced.

Fabrice Benaut, CIO global IFR,


Head of IS and Operations, IFR Monitoring

Data visualization
at IFR-GfK

81

3.2 The impact of data visualization on the relationships between IT


So lets steer clear of discussing the loss of competences by IT departments or BI teams. Data visualization can, however, lead to
new questions about how data is represented, how roles and responsibilities are distributed across the business divisions and
the IT department, and how projects are led.
IT departments understand this. Even those that initially have reservations will come to realize that data visualization is not a
threat, and they will eventually soften their stance.
IT departments recognize the
value of data visualization tools
Initially the IT department had significant reservations because our dataviz tool was
not a corporate one. The IT department changed its position 18 months ago. Now there are no concerns, and
the solution has become a corporate tool.

Richard Core, Senior Business Analyst,


PagesJaunes

In reality, data visualization is an opportunity for the IT departments and BI teams. It will relieve them of time-consuming tasks
by allowing them to focus on activities with the greatest added value. It will also enable them to invent a new form of BI and
forge a better relationship with their business divisions.

82

3.2 The impact of data visualization on the relationships between IT

3
An opportunity for IT
departments and BI teams

At SAS, we believe that data visualization is an important asset for an IT department. It oers it a real
opportunity to invent a new service for the business divisions and to reposition itself as a business partner.
Based on this new partnership, we can imagine two forms of business intelligence based on their
requirements:
Traditional BI to respond to need for conventional reporting.
New BI. To be specific, it will be designed to respond to new business questions from the users that
cannot be met by traditional reporting because they want to analyze and explore their information for
business purposes.
The BI teams would implement a server dedicated to a data visualization tool like SAS Visual Analytics,
completely scalable, with the ability to add an additional resources to business divisions without needing to
write a statement of needs. It would provide a self-service BI environment in which business divisions could
access their data in order to explore it and quickly produce any reports they required, including ones they
had never imagined.
They would have a much healthier governance model at their disposal. At the BI manager level, these ideas
have been well received. It would also remove a thorn from their side. Nowadays, having to meet the specific
reporting requirements for all departments entails quite a lot of work. Data visualization allows BI to be
relieved from a part of this load and be able to oer more value to the business divisions.

Serge Boulet, Director of Marketing and Communication,


SAS France

83

3.2 The impact of data visualization on the relationships between IT


Some IT departments even use data visualization projects as a way to strengthen their standing in the organization by
demonstrating their ability to innovate and add value. This is how Charles du Rau, Director for Consumer and Market
Knowledge at Ubisoft, explained it: "This tool was also used as a test run for our IT department to allow it to show its potential by
demonstrating its knowledge in this field of data visualization. It has now become a sort of showcase for it."
The buzz phrase is "value creation," as analyst Philippe Nieuwbourg explains below:
The key to a smooth relationship
between the IT and business divisions
Generally speaking, my view of the IT department is that it is at the service of
the business divisions. In my opinion, the IT department is not there to launch projects that have not been
requested in advance, or at least validated, by the business divisions, because they are the ones that create
value. This does not prevent the IT department from driving proposals by showing the business divisions the
dataviz tools in existence and by proposing to work with them to see how these tools can create value.

Philippe Nieuwbourg,
independent analyst-journalist in decision-making IT

Numerous organizations are already at this stage with their data visualization projects. The more advanced ones are clearly
those in which the IT departments and business divisions are already working in partnership1.

1. With regards to the relationships between the IT department and the business divisions, refer to the Guide for IT Practices, 2013 edition, 80 companies sharing their experiences and their projects for tomorrow, EBG, April 2013.

84

3.2 The impact of data visualization on the relationships between IT


The IT department and business partnership on the
subject of data visualization: Business as usual at ID
Logistics

The eorts provided with regards to data visualizations


were initiated by the IT department, but they are based on

cooperation with the business divisions operating in project mode, a way of functioning that we have
systematically adopted. Because of this, the focused dataviz approach did not have any significant impact on
the IT department/business division relationship.
Emmanuel Vexlard, DSI,
ID Logistics
Certain organizations are structured around partnerships between IT and business divisions, with heads of business divisions
aware of IT matters and IT project leaders aware of business division problems. This type of organizational relationship is a
major plus, particularly for data visualization projects.
A cooperative eort between
IT and business divisions

I work for a business division and not for the IT department. I represent users while
having a sort of alter ego within the IT department a business analyst that supports me.

Because of this relationship, progress related to data visualization is the result of this collaboration.

Guillaume Deschamps,
FMT/Central Functions YTI/EDA, STMicroelectronics
Our solution is the result of a joint initiative between the IT department and the marketing team. Its success
is based on real cohesion between the two departments.
Jrme Tharaud,
Head of Research and Development,
Prisma Media
85

3.2 The impact of data visualization on the relationships between IT

Even without going as far as having this partnership, a data visualization project can, by its very nature, improve the relationship
between business divisions and IT departments.
"With this type of solution, the users feel a great deal of responsiveness from IT, said Olivier Brousseau, Business Intelligence
Program Manager at Schlumberger. The question of time-to-market is finally resolved. This contributes to good relationships
with the business divisions."
In the end, as Boulet and Nieuwbourg said, data visualization is an opportunity for the IT department and BI to rid themselves of
tasks that add little value in order to focus on a relationship with business divisions that create value.
This is how data visualization can become part of a win/win scenario. To enable data visualization to play its role to the fullest, a
proper data governance procedure needs to be implemented within the organization.

86

3.2 The impact of data visualization on the relationships between IT

3
When data visualization highlights the
importance of data governance

The multitude of opportunities oered by data visualization and the


various possibilities companies have to implement it (cloud, in-house, departmental, corporate, etc.) also
urgently highlight the question of data governance. At a time when each company needs to consider
information as an asset, defective governance can have serious consequences: under-exploitation of data
(caused by a lack of sharing) or incomplete exploitation of data (due to low quantity, incorrectly accounting
for real time).
A centralized data governance approach guarantees improved availability, integrity, security and real-time
value of the data for business divisions. And even when the data itself is decentralized for often legitimate
reasons (which is more and more often the case), particularly when the company uses decentralized
infrastructures (departmental, cloud etc.). A company's capacity to use information as a strategic asset is
directly related to its maturity in terms of data management and governance.

Mouloud Dey, Director of Solutions and


Emerging markets,, SAS France
One important idea to keep in mind is that data visualization, even if it provides business divisions with a great amount of
autonomy, still creates a strong connection with IT departments and BI teams if for no other reason than their common interest
in the data.
The autonomy of the business divisions is useful if it is implemented intelligently; especially if it creates even more value for the
IT department and BI. Getting business divisions to speak the language of data will contribute to increased awareness of the
datas value.

87

3
3.3 Key factors for success
For a data visualization project to succeed, three factors need to be considered. First of all, good project management is required:
preparation and planning, choosing the right scope, implementing appropriate methodologies, etc.
The second factor is to ensure that the data is a prime consideration: who it is targeted at, its quality, respect for confidentiality
and access rights are essential.
Last, but certainly not least,, the user experience and graphical intelligence play a vital role in data visualization acceptance and
its eective use.

Proper preparation of the project


The first secret of success is preparation on two levels: the content of data visualization, and the progress of the project.
The panel members remind us that data visualization is not a panacea and that to be of use, it must be well thought out.

88

3.3 Key factors for success

3
Its critical to define objectives before launching a
data visualization project

We only push the visual dimension if it really means something to


the client.

Clment Delpirou, Associate, InfoPro Digital

We need to be thinking about the data to be actually able to visualize it. In other words, we need to ask
ourselves if there is interest in showing a figure, to visualize a trend or a break from the trend, etc.

Caroline Denil,
Data Warehouse Program Manager, SPF Economie

Regarding the methodology, preparation consists of implementing a viable, timely and flexible process for collecting, analyzing
and representing data.
"It is important to take time to construct the right procedures in order to start with a road map that is both lasting and able to
accomodate changes applied over time," said Yves Daelmans, ICT Project Manager at IMEC.

89

3.3 Key factors for success

Targeting data and visualizations based on user profiles


A second factor for success is the ability to use data visualization as a communication tool. However, to communicate
eectively with someone, you need to be able to respond with clarity in a way they will understand.

Providing users with the information they need


according to their preferences and situation
We have methodical clients (who prepare a long time in advance
and who are trying to optimize dierent criteria), restricted clients (who do not really have any choice about
a given journey), opportunist clients (who, above all, are on the lookout for a good deal and can be flexible),
etc. Our application needs to meet the very dierent expectations of these dierent profiles without
overwhelming the online user in an attempt to flesh out the application. We must therefore oer good views
and good profiles.

Samuel White,
Head of User Experience, Voyages-sncf.com

90

3.3 Key factors for success

3
Identifying targeted users and their profiles.

Three key questions to ask in order to select


the best representation
What question do we want to answer? Not all the graphs allow the same analyses to be represented
(distribution, trends, breakdown, etc.). This is why it is important for the designer to think about what he or
she really wants to illustrate.
Who are we speaking to? Are they experts or typical users? What will they do or decide with the
information (e.g., keep it for a later decision or make a decision right away)?
What is the context of the contact?
Getting the desired visualization not only depends on the graph itself but also on the intellectual and visual
capacity of the person reading it. All the designer can do is try to anticipate this capacity to a lesser or greater
extent in order to select the most appropriate representation. For example, stress (noise, crowded
environment, etc.) alters perception by reducing the field of awareness and hiding the peripheral landscape.
In the same way, the media used (a tablet, projection on wall screen or paper) will aect the user's ability to
perceive the visualization.

Claude-Henri Mldo,
co-founder of Alcedis and member of the International Institute for Information

91

3.3 Key factors for success


Education is vital when addressing the general public.

Keeping it simple is essential


Simplicity is fundamental. For Social Ecorama, we took the very
traditional criterion of positivity, widely used in the majority of opinion polls; from a graphical point of view,
curves are used for representation that are very easy to understand. It is really important that the tool can be
used by non-professionals. And the quest for simplicity is never over. I am convinced that there is still more
for us to accomplish on this subject.
Emmanuel Lechypre, Editor,
Head of Analysis and Forecasts,
BFM Business
The problem is greater when data visualization needs to be aimed at dierent audiences. You need to consider how to tailor
visual representations for each audience. Consider, for example, the following case of the Belgian SPF conomie, which is
aimed at both the general public and professionals.

92

3.3 Key factors for success

3
Making data and information available to
dierent types of users

Our data is available to the general public via our website or via the
Belgian Open Data portal in the form of aggregated data, but its also available at the Belgian National
Institute for Statistics (this audience has a set of much more detailed data available). Researchers and
university students also have access to much more detailed information for their research studies.

Caroline DENIL,
Data Warehouse Program Manager, SPF Economie
But even in a strictly professional environment, target audiences can be dierent. The contexts and diering stakeholder
requirements need to be accounted for. In manufacturing, for example, a worker or a foreman will need data visualization to
solve a specific problem on the production line. A manager will often need a more global view.

Dierent visualizations for dierent


profiles and dierent contexts.
Most of the time, our data visualizations encompass a static dimension
(particularly useful for workers and heads of departments). But interactivity is important to engineers.

Guillaume Deschamps,
FMT/Central Functions YTI/EDA, STMicroelectronics

93

3.3 Key factors for success

Starting off with a limited scope for learning


Another good practice is to experiment with a restricted set of well-known data, because this allows you to learn without
being overwhelmed. Your initial solution can then be used as a proof of concept. Once you achieve initial success, it is easy
to apply what youve learned to larger projects and to gain the confidence of your targeted users.

We started o with small projects involving infographics or Google Maps, and we developed those projects
step by step.

Clment Delpirou, Partner, InfoPro Digital

We started o with a pilot project to learn about the SAS technology.


Yves Daelmans,
ICT Project Manager, IMEC

We decided to launch the approach in only two areas, gardening and IT, representing half a percent of
requests. We knew that we had labeling problems in these two areas and, at the same time, they were noncritical areas in terms of business. It was the ideal ground for testing the approach and producing a proof of
concept.
Martin Daniel,
Head of Data Visual project, fifty-five

94

3.3 Key factors for success


Starting small is a good way to avoid procrastination or attempts at perfection.
As soon as a simple solution has been produced (sometimes called a minimum viable product), its in the organizations interest
to present it to the decision makers and users.
This baptism by fire allows you to gather valuable feedback and comments in order to improve the solution.

Those who dare, win


When an organization has planned to use data visualization to innovate, it needs to stick with it.
I normally advise starting with a small statistical analysis without waiting for the whole platform to be
defined or all standards to have been set.

Caroline Denil,
Data Warehouse Program Manager, SPF Economie

95

3.3 Key factors for success

Ensuring the quality of data


With BI, you reap what you sow. In other words: garbage in, garbage out. If you want data visualization to communicate the
right message, allow informed decisions to be taken or unknown territory to be charted, there is one condition that must be
met before all others: having reliable data.

Data quality: a prerequisite for data


visualization
Above all, data must be well structured. Cleaned and prepared data is the starting
point for good data visualization. Data needs to be consolidated to four or five essential axes, while ensuring
that it remains correct and its integrity is maintained.

Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate, Lagardre Active

Having a well-structured data warehouse has really helped our power users in creating their own reports.

Yves Daelmans,
ICT Project Manager, IMEC

96

3.3 Key factors for success


Data visualization enables data quality improvements, but it requires developing internal discipline in that area. Poor data
quality will stand out immediately.
Incorrect data may appear in the form of unexpected points on a graph, for example. If problems regarding data quality are left
unattended for too long, they will jump o the screen in a data visualization environment.

A lesson for data quality


We can see that company procedures around data visualization always
consider data quality as being key, whether on an ad hoc or systematic basis.
For this reason, we can say that dataviz tools require more attention on data quality than traditional BI tools.
But the approach is the right one because it allows companies to foster a strong data quality environment.

Pierre-Olivier Sicamois,
RHead of Advertising,
Technology Directorate, Lagardre Active

97

3.3 Key factors for success

Building cooperation between departments


Another ingredient for success is the cooperation between project stakeholders. In section 3.2, you saw how data
visualization could be an opportunity to strengthen partnership between IT and business units. But in general, it is a hybrid
process at the crossroads of IT, BI and communication. This is why the most spectacular results come when all three groups
work together.

Collaboration is crucial
for success
The editorial sta contributes a lot to our dataviz approach. Besides, everything started with
initiatives of our journalists. Nowadays, the content of our data visualization is the result of collaboration
between two traditionally dierent teams that now operate very cleverly with us the technical team and the
editorial team.

Clment Delpirou, Partner, InfoPro Digital

Data visualization helps to break down barriers within an organization and leads to much greater interaction.
"The solution allowed us to cross the existing frontiers between dierent parts of the organization, said Franck Jouenne,
Director of IT and Strategic Sourcing Services at Alcatel Lucent, and it provides now anyone having to make a decision the
elements needed to do so."

98

3.3 Key factors for success

Team training
Data visualization does not really require any end-user training, and it is most successful if it is adopted quickly. To
accomplish this, simplicity and intuitiveness are essential. But creating a proper report can be complicated, so training those
who produce visualizations is a must. This is why the team at IFR-GfK has a user-experience expert working with them.

Learning to produce good data


visualization
I regularly send my teams for training in the tool but also to learn the rules of
design and good practices in terms of visualization. There are standards and good habits to acquire. Some
are universal, others depend on local preferences. This must be learned.

Fabrice Benaut,
CIO global IFR, head of IS and Operations, IFR Monitoring

99

3.3 Key factors for success

3
A gap in university training

Visual grammar exists. The problem is that it is not taught in universities. In the 70s, Jacques Bertin
published a reference work, Semiology of Graphics, a real grammar book for graphics. Today, the experts in
the subject are people like Claude-Henri Mldo or Stephen Few, and they continue to define and develop
these visual grammars, which still need to be learned and applied. However, graphic visualization is not
taught in any syllabus in management, statistics and business schools.
Students are taught to represent things correctly, but not necessarily in a way that is easy to communicate.
We find that managers who completely master standard Excel charts are unable to export them to another
tool to improve the look and feel that visualizations oer.

Philippe Nieuwbourg,
independent analyst-journalist in decision-making IT

Aesthetics as a way to collect information


Data visualization cannot be reduced to an aesthetic representation of data. You can make it look pretty, but still have the
information be perfectly useless. This does not mean that aesthetics has no role to play. If it is used well, artistry becomes
essential for eective data visualization.

100

3.3 Key factors for success

3
Do not forget the aesthetic side of data
visualization

The graphical aspect and the artistic nature of dataviz, often regarded as
toys, are far from being useless: They make the business divisions want to explore their data. They help to
bring value to the business divisions.
Pierre-Olivier Sicamois,
Head of Advertising,
Technology Directorate Lagardre Active
Aesthetics enable data visualization. If this is important for the user of the data visualization tool, then it plays an even more
essential role when data visualization is used for the purposes of data mining. Mining requires availability and concentration,
and the design of data visualization solutions is crucial in that area.

The importance of design for encouraging new


questions and answers
Users need to feel comfortable in the mining environment that we
oer them, in order to get them to play with the data. This strength is what allows emerging, unforeseen
insights to appear.

Fabrice Benaut,
CIO global IFR, head of IS and Operations, IFR Monitoring

101

3.3 Key factors for success


In order to make aesthetics eective, you must use your imagination to go beyond traditional Excel type graphs.

ColiPoste's performance
indicators

102

3.3 Key factors for success


Innovative features correctly implemented are the key to data visualization acceptance.

Change your approach for more


operational value
We need to go beyond the standard concepts and move away from curves, pie charts
and histograms. Innovative visual representations will sometimes allow you to see operational realities that
were hinted at but could not be easily demonstrated.
By using a bar chart format where the monthly performance for each member of a sales team is represented
by a specific vertical bar, it is easier to understand how performance is distributed.

And, we need to be ready to spread the word and give teams the tools for reading the new graphs. Consider
data visualization in both its forms as a means of dynamic analysis and exploration, and as an exceptional
means of conveying information.

Bernard Lebelle, founder of Bleu-Cobalt


103

3.3 Key factors for success

3
The world is not painted
red and green

Perhaps a passenger had a bad experience, while other passengers had a good experience. Data
visualization helps our passengers to understand that they do not share same needs. And our satisfaction
ratings must account for this dimension.
This is why in terms of colors used in our representations, we did not opt for the classic red (negative) and
green (positive). We are in fact comparing global expectations to individual experience.

Philippe Saon,
Head of Media, Aroports de Paris
Degree of service satisfaction on the
mobile site for Aroports de Paris
But Claude-Henri Mldo warns us against creating original features for originalitys sake. Dont
innovate in terms of cool graphics and sexy design if these dont add real business value.

104

3.3 Key factors for success

3
Data visualization design fallacies

The motion chart or animated graph, with its awe-inspiring movement,


is generally an inappropriate approach. Insofar as it is not possible to follow all the movements, the user
needs a way to replay the animation. This is where the first problem arises: the need for a reset button if we
want to give the user the ability to restart the animation.
The second problem is the hypnotic fascination caused by the movement that leads the user into losing
concentration. Movement makes it dicult to absorb all the information available to the point where a note
is required to explain what needs to be focused on and what needs to be retained.
And 3-D, to a certain extent, could be the death of dataviz.
Because it mirrors nature, 3-D appears more real than a flat graph. But 3-D presents major problems with
regards to the precise evaluation of measures. Its perspective makes deciphering the size of a symbol very
dicult. We do not know if the symbol is small because it represents a low value or because it is supposed to
be farther away from the viewer. Another problem is where there is a particularly large object in the front
because it may hide smaller items behind it (visual occlusion).
And even if users know how to operate a three-dimensional device to move around the graph (e.g., a mouse
or a joystick), we cannot be sure that they will analyze the 3-D graph from all possible angles to identify the
weak signal that they are looking for.

Claude-Henri Mldo,
co-founder of Aldecis and member of the International Institute for Information Design

105

3.3 Key factors for success

Using contextual help to accelerate learning


The direct and intuitive nature of data visualization does not rule out the risk of misinterpretation. It may be that the graph
itself is misleading or that the user does not understand the underlying data or the conditions for implementing data
visualization (timeline, filters, scale, etc.).
In order to avoid confusion, the designers of data visualization should provide legends or explanations with their graphs.
Because user manuals are rarely read (and because the subject does not warrant a manual) a contextual tool is the most
eective solution and should be placed next to the graph itself. Alternatively, you could place explanatory texts under the
graphic or use pop-out balloons that appear when you hover over the graph, or even embed short videos.

Knowing how best to


educate
We are working on version 2.0 of BeStat, which is more graphical and user friendly. Faced with
complexities presented by the software solution for users, we plan to systematically explain how it works, in
particular with practical videos regarding the use of graphs.

Caroline Denil,
Data Warehouse Program Manager, SPF Economie

106

3.3 Key factors for success

With our project, we took particular care to ensure that it is easy to use. For example, we implemented an
automatic save for the last variables used allowing the application to be manipulated more quickly.
Introductory texts and a glossary highlight the meaning of dierent graphics and the way they were put
together in order to allow users to understand them well and more easily interpret them.
The results have been positive because we see complete self-reliance by teams that have never been trained.

Charles du Rau,
Consumer & Market Knowledge Director, Ubisoft

Pay attention to legal and confidentiality issues, especially for public uses.
This piece of advice is important: Data visualization does not ignore the requirement to respect data confidentiality, the right
to privacy and the conditions for using the data. As Caroline Demil from SPF conomie rightly pointed out: "We need to take
care not to publish the data too quickly in order to avoid problems of confidentiality and to be aware of the legality regarding
methods of publication."

107

3.3 Key factors for success


All of the factors for success mentioned previously are important. Here is a small summary provided by Dorina Ghiliotto-Young
of Ingenico of three factors that she believes are particularly important.
Three prerequisites for
good data visualization

1. Having quality data


Data must be complete, reliable and used well! For example, data that is in no way connected should not
enter the mix.
2. Understanding the data
Data visualization must be based on a good understanding of the relationship between the data being
analyzed. In order to avoid any risk of confusion in interpretation, we may be forced to remove data from the
project scope. For example, for a market, it is not necessarily relevant to compare the data for developed
countries with that of developing ones. This could lead to incorrect conclusions because the countries have
dierent situations.
3. Being able to interact with data visualization
The options for interaction can range from the choice of colors to much more sophisticated options such as
3-D. Nothing is prohibited, provided that it means something and that it assists with understanding.

Dorina Ghiliotto-Young,
Head of Innovation,
Ingenico

108

3
3.4 Pitfalls to avoid
If certain good practices maximize the chances of a projects success, you should likewise anticipate diculties, as with any
project. Based on their experiences, panel members share some fundamental pitfalls and provide pointers for overcoming them.

Reduce information overload


One of the advantages of data visualization is that it increases the density of information.

More information in less space


thanks to graphics
If the amount of information transmitted per pixel is calculated, there is always much
more information in an image than in a text document.

Philippe Nieuwbourg,
independent analyst-journalist in decision-making IT

A graphic condenses the information. It allows important things to be seen quickly such as patterns and
exceptions to those patterns.

Claude-Henri Mldo,
co-founder of Alcedis and member of the International Institute for Information Design

109

3.4 Pitfalls to avoid


Be careful, however, with information overload! Charles du Rau from Ubisoft believes ". . . there should be no more than about
40 data points on the screen; otherwise, the user will get lost. Andrew Musselman, from Accenture agrees.

When to avoid data


visualization
We do not recommend that you try to visualize absolutely everything, because
sometimes, visualization only produces confusion. For example, if you try to add points representing
millions of users, you will get a nice picture, but extracting information from it will be impossible. Analysis of
tables of numbers or curves will often be a more powerful way of revealing connections.

Andrew Musselman,
Data Scientist, Engineer and Architect in Accenture's Big Data Department
Fortunately, certain concepts of data visualization can help avoid this "graphical information overload." Its normal, for example,
to go from the general to the specific by allowing the user to zoom or filter if they want more detail.

First, the overall view and then


the details
The progressive disclosure of data is one of the key success factors for using dataviz. It
is a question of determining beforehand the crucial information to be highlighted first is and then what is the
secondary data that needs to be shown.

Frdrique Ville,
Innovation Director, Voyages-sncf.com

110

3.4 Pitfalls to avoid

3
Set the scene first and the
details follow

In order not to overwhelm the user, we have adopted an approach of overview first,
details on demand. In other words, business users are able to refine their searches by means of gradual filters
to track badly categorized queries. It should be noted that applying one filter automatically modifies the data
available in the other filters. In this sense, we are adopting the good practice of progressive disclosure
recommended for dataviz.
Martin Daniel,
Head of Data Visual Project, fifty-five

Managers sometimes fail to see the value of data visualization


One pitfall may be a lack of enthusiasm by management. Initially, managers might say that data visualization adds nothing
because your organization already has reporting tools. "We haven't really encountered any resistance to implementing our
data visualization solutions, apart from right at the start, said Jrme Tharaud, Head of Research and Development at Prisma
Media. We already had reporting tables or reports, although when I first mentioned our project to business directors, they
couldn't really see any tangible benefit in it."
So an educational campaign was needed to show the commercial directors all the advantages of data visualization and how it
diers from traditional reporting. What was the process that was finally used at Prisma Media? "In the end, it was aesthetics
that convinced everyone, Tharaud said, and we had real success."
This proves, again, the importance of graphic quality as a success factor.

111

3.4 Pitfalls to avoid

Skepticism regarding performance


Paradoxically, data visualization performance can be its own worst enemy. Today, the available solutions sharply contrast
with the traditional BI solutions in terms of reduced processing speed and smoothness of interaction. So users can be
dubious. Sometimes, when data visualization delivers bad news, the temptation is to doubt the reliability of even high-quality
solutions. Other times, fear of the new leads some to question data visualizations value. This can confer fear, uncertainty and
doubt to new solutions.
The counter-measure? Giving users easy, direct access to the underlying data.

Data visualization too good


to be true ?
Faced with the ease of use and the speed of a data visualization tool, those introduced to it (particularly if
theyre decision makers) tend to be skeptical and question the results. It is as if the tool was too user-friendly
and fast and perhaps hid a crucial flaw. It may also be a defensive reaction to a graphic reflecting bad
performance. It would be easy for someone to say that the visualization proves nothing because we have
incorrect data behind it. The solution we have found is to allow the data visualizations to speak for
themselves by allowing the smallest detail to be accessed directly by the client. Faced with something this
enabling, reservations disappear.

Sandrine Noail,
Head of Control and Analysis, Coliposte

1. http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt

112

3.4 Pitfalls to avoid

A final pitfall: forgetting traditional methods


Another risk is falling into the trap of making visualizations that are too simple. Of course, ease of use is important; for
example, presenting general data before going into details. And aesthetics is another important success factor, but it can lead
to misinterpretation of data if its not based on the foundational pillars of serious BI and data reliability.

Confusing simplicity with


simplification
The future will, without doubt, bring more simplicity of use. However, be careful not to
confuse simple and simplistic. We must never forget that all the basics of data analysis remain true, in
particular the importance of the quality of the data.

Fabrice Benaut,
CIO global IFR, head of IS and Operations, IFR Monitoring

As you can see, data visualization already oers numerous advantages. Yet in some ways, it is only just
beginning. The fourth and final section considers the outlook for data visualization and its users.

113

4
4

Outlook
This section imagines how data visualization may develop in the future. First, youll read about
projects in development (or in planning) by our panel members. Then the panel will consider how
they think the future of data visualization will pan out.

114

4
4.1 Future projects
The majority of our panel saw data visualization extending its scope in the future.

In view of initial successes,


the desire to continue
Our data warehouse continues to grow, which explains why we have BI data visualization
projects in our sights. We have already scheduled a project to implement financial reporting. We are currently
in the process of determining the KPIs associated with health and safety in the workplace.
Yves Daelmans,
ICT Project Manager, IMEC

Management asked us to work on a new dataviz application for evaluating suppliers: systems for calculating
bonus/malus to account for the quality of suppliers; but because these vary, we now want to standardize our
evaluation and ensure traceability. The tools we learned to use with our first dataviz project lend themselves
well to this.
Richard Core, Senior Business Analyst,
PagesJaunes

We plan to use data visualization as opportunities arise. And we also intend to transfer what we used for the
sales team to the finance department in 2013.
Jrme Tharaud,
Head of Research and Development,
Prisma Media

115

4.1 Future projects


Beyond these data visualization projects, Fabrice Benaut of IFR-GfK and Olivier Brousseau of Schlumberger think that data
visualization will change traditional reporting. Brousseau said: "Data visualization is moving forward. It is innovating BI, but
traditional BI will need to catch up with dataviz." This doesnt mean that more traditional reporting is being called into question.
Rather, reports are likely to adopt some data visualization concepts.

Rethinking traditional
dashboards
In the future, access to simple, traditional dashboards would be via a data visualization tool.
It would be a tragedy not to rethink them in the light of this new paradigm, instead of saying they are fixed,
that they will never evolve, while they continue to represent 80 percent of data usage .

Fabrice Benaut,
CIO global IFR, head of IS and Operations, IFR Monitoring
To get to this stage, new skills must be acquired. In tomorrows world, IT departments and BI teams will employ more people
having nontraditional career paths and a variety of skills.

116

4.1 Future projects

4
What IT job profiles are required to lead the
data visualization projects of tomorrow?

We will need to convert enthusiasm into good practice and new uses. It also
allows us to see that IT profiles will change. There will be ideal candidates equipped with knowledge of
modeling and also an appetite for graphics and usability.
The next stage will also certainly include scripting and storytelling. In other words, the process of
transforming a series of photos into a video that tells a story.

Franois NGUYEN,
Director for Analytics and Relationship Marketing, SFR

My career path is in mathematics, and the mathematics and statistics skills of many of my colleagues are
turning out to be very relevant for data visualization. Those who have studied IT, physics, biology and
chemistry also know how to visually represent their work.

Andrew Musselman,
Data Scientist, Engineer and Architect in Accentures Big Data Department

117

4.1 Future projects


But where do you find these atypical job candidates?
If data visualization is still not fully integrated into syllabuses (as pointed out previously by Philippe Nieuwbourg), initiatives are
nevertheless being led by schools, universities and organizations to give training in this discipline.

Training young people in data


visualization
Conscious of the growing need for these new job profiles in its ecosystem, SAS
decided to develop an innovative training program specifically to respond to the numerous recruitment
demands of its client and partner companies as well as for its own recruitment of qualified young people.
Out of this was born the SAS Spring Campus 2013 a full course syllabus together with a mandatory
internship at the end of the cycle with the final aim of acquiring a first job placement. The first cycle brought
together 20 students who were given several projects by client or partner companies.
Encouraged by the success of this first year (250 applicants received), SAS Spring Campus will be renewed
and extended in 2014 for a new course, because companies needs for these new profiles will only grow in
the coming years.

Ariane Liger-Belair Sioufi, Academic Director,


SAS France

118

4
4.2 Looking forward
According to some of our stakeholders, in the future, data visualization needs to fill the gaps in expectations that are still not
being fully met (for example, the portability of solutions between dierent environments). Jrme Tharaud of Prism Media said:
"What is currently blocking us is the fact that neither of our solutions is multi-platform, nor really multi-browser. For us, this
would really bring added value."
Our contributors are, on the whole, happy with the current state of data visualization. Some go as far as imagining how the data
visualization of tomorrow could be. The IT Director at ID Logistics, Emmanuel Vexlard, focuses on accessibility: "If we try to
imagine the future of data visualization, the ideal thing for us would be the ability to obtain graphical representations from
simple queries in everyday language, just as easy as getting results from Google."
Others point to the booming quantified self movement1 individuals providing personal data linked to their lifestyle, hobbies
and how they eat, exercise, sleep, etc. All these facets can now be measured by sensors or devices, from the smartphone to an
electronic fork. These create a huge amount of data that cannot be easily mined without data visualization.

1. http://en.wikipedia.org/wiki/Quantified_Self

119

4.2 Looking forward


Frdrique Pain, of Alcatel-Lucent Bell Labs, foresees "the future of a new dataviz experience, individualized and contextualized.
The fields of fitness and health will probably be at the cutting edge of this development."
Will these still fairly cutting-edge uses of data visualization become commonplace? Some members of our panel seem to think
so. Data visualization would signal a new paradigm. France Tlcom's movement footprint project is significant in this respect.

Heralding a new era


Our data visualization procedure allows us to think beyond the data, starting with
movements within a closed space or an open environment. By overlaying the dierent graphic reports, we
are able to achieve specific applications: optimization of routing according to road trac, improvements to
city planning, etc. It all has to do with cross-referencing data data regarding individuals, their
consumption, their habits as well as external data about the town or transportation route. Some of this data is
not under our control, particularly external data. But by considering it and linking it with data over which we
have more control, we begin to see new perspectives emerging that shed light on potential solutions for us.
It is a real reflection on the change in behavior that we are seeing emerge. ... For example, regarding a more
rational management of energy, we are asking ourselves about how ways of optimizing the user's behavior
may lead us to look again at consumption habits. Changing these habits may, in the end, result in saving
energy. As far as the France Tlcom group is concerned, this optimization can be a benefit both internally
and externally.
In the end, we are witnessing the start of a movement that can teach us things about ourselves, our ways of
doing things and make us aware of our failings. This is why in a philosophical, indeed sociological approach,
we come to question established behavior.

Lamia Delenda, Project Management Oce, Orange

120

4.2 Looking forward


How data visualization will look in the future is dicult to predict because data visualization straddles dierent disciplines
(usability and IHM1, IT, design, analytics, communication, etc.). In large part, it is the innovations in each of these fields that will
contribute to the development of data visualization.

121

Conclusion
D

ata visualization, while growing in popularity, remains a largely uncharted (but fertile) land. Data visualization has
passed the concept stage, and the more than 30 organizations represented here are able to bear witness to its real-world
application with completed, successful projects and new projects they are developing and pursuing.

Data visualization is a radically dierent approach to BI and, at the same time, it is extremely loyal to good BI principles. It has
advantage of being able to use graphical representation technologies that allow it to quickly exploit huge volumes of data. It
oers a new way of analyzing, communicating and exploring data of all sizes. It empowers your data and is a catalyst for its
transformation.
By focusing on data, data visualization reveals its true nature as a reservoir of knowledge for the organization, a real, tangible
asset a source of value.
Use the information in this book as a guide to help you take full advantage of the value that is in easy reach.

122

Contributors

Contributors
contributors
SAS presentation
EBG presentation

123

Contributors

Accenture
Andrew Musselman Data Scientist, Engineer and Architect in the big data practice
Trained mathematician and IT engineer (Central Washington University) Andrew Musselman first
worked as a Web developer (at Real Networks), then as a software development engineer (Real
Networks and Rhapsody International). As he gained experience, he gravitated towards big data
analytics. He became a developer/analyst at WhitePages before joining Accenture as a Hadoop/big data architect,
his current position.

Aroports de Paris
Philippe Saffon Head of Media
With a degree in Telecom Management (INT), Philippe Saon has more than 15 years experience in
the field of communication, marketing, commerce and new media. Before joining Aroports de Paris
in 2010, he worked at Bouygues Tlcom as head of marketing/commercialization of services and at
Kodak as head of digital marketing. As part of his role, he developed the e-commerce activity at both groups.

Affini-Tech
Frdric Gonnet Associate Director
As an engineer from the ESIEA, Frdric Gonnet started his career at service and IT companies. Since
2003, he has been the founding director of FAIRSYS Consulting. He has lead numerous projects in
the IT and organizational fields, in particular within PagesJaunes (in production IT change).
In 2012, he also became associate director within Ani-Tech, a company focusing on big data and data
processing. He is in charge of the business divisions (including requirements, support, steering and project
management) and the data visualization product.

124

Contributors

Alcatel-Lucent
Franck Jouenne Director, IT and Services Strategic Sourcing
Franck Jouenne has spent all of his career at Alcatel-Lucent in the IT sector. In 1993, he joined the
group's Research and Development Directorate (now Bell Labs) to oversee the development of
various innovations for network applications. Then in 2002, he joined the group product lines as
Director of Computing Systems and Industrialization and then in 2008, as Director, IT & Services Strategic
Sourcing, his current role.

Alcatel-Lucent Bell Labs


Frdrique Pain Experience Strategy Director
After a master's in research in industrial design and a master's in ergonomics, physiology and
cognitive psychology, Frdrique Pain started her career with three years of research into the design
process for interfaces at EADS in conjunction with CNAM and the University of Compigne. In 1997
she joined Alcatel's brand department, where she developed a design department for applications of all ranges of
mobile and land-line telephones in France and China. In 2009 she founded the Studio Application within Bell
Labs' research laboratories, where she develops and puts innovation into operation for practical use.

Aldecis
Claude-Henri Mldo Partner
Claude-Henri Mldo is the head of the consultancy company Aldecis and the Visual Decision
Association of French experts in data visualization. He has received several awards such as the silver
medal in the data visualization competition for the 2012 presidential election organized by Google
and the title of 2011 Chief Expert for Financial Graphics within the IIID World Organization for Information
Design. He has been producing dashboards for 20 years and custom visualizations for the largest groups in
France. He has conducted several studies comparing the graphics used in the annual reports of CAC-40 and Dow
Jones companies.

125

Contributors

BFM Business
Emmanuel Lechypre Editor, Head of the Analysis Center and Forecasting
An economist by training, Emmanuel Lechypre was the editor-in-chief of the General Economy
Department at l'Expansion, where he was also in charge of the group's forecasting center. He is
currently editor and head of the analysis and forecasting section at BFM Business. Together with
Franck Dedieu and Franois de Witt, he co-wrote 150 ides reues sur l'conomie (150 ideas on the economy),
published by L'Express in 2010. His latest book, with Batrice Mathieu and published by the same publisher, is
called La prsidentielle en 25 dbats (The Presidential Election in 25 debates).

Bleu-Cobalt
Bernard Lebelle Founding Partner
Bernard Lebelle, graduate from EDHEC, is the founder of Bleu-Cobalt, a consultancy, training and
facilitation firm. Trained with some of the largest consultants (PWC and Deloitte) and having honed
his experience with Air France, he has developed unique skill-set, the result of a synchronistic
approach aimed at obtaining the best from consulting and visual thinking.

Coliposte
Sandrine Noail Head of Analysis and Control
Sandrine Noail has more than 13 years experience in supporting sales. Having worked as head of
studies, project leader and project director, she is currently leading and developing Coliposte's
commercial eorts. By means of analysis, she supports strategic objectives set by the company.

126

Contributors

Decideo
Philippe Nieuwbourg Teaching journalist specializing in analytics, business intelligence and big
data
Since 1994, Philippe Nieuwbourg has been informing the French-speaking community about tools
to help analytics at www.decideo.fr, the go-to site for subjects such as big data, data warehouse,
reporting, analysis, data mining, graphic visualization and data quality. The author of several books on analytics,
he participates each year in numerous conferences in France and abroad. He also teaches a university course at
Quebec University in Montreal.

Fifty-five
Martin Daniel Data Visualization
Graduate of the ESCP Europe, Martin Daniel joined Fifty-five at its start after a notable spell at Google.
At Fifty-five, Martin is creating a team of data scientists for products based on data from Web
analyses to provide companies with more intelligence.

France Tlvisions
Antoine Allard Communications Manager and Social Media Marketing
Antoine Allard started his career at Ogilvy PR in digital- and editorial-influenced strategy. He joined
France Tlvisions in 2012 as online communication manager within the marketing directorate for
digital editions of the public audio-visual group. In particular, he is involved with community
management and social media strategy and community TV for France 5 and the group's Nouvelles critures et
Transmedia projects.

127

Contributors

Havas Media
Emmanuel Barbet Global Accounts Director
Emmanuel Barbet graduated from the IDRAC Lyon in Marketing and Management. Having worked
for 10 years in online marketing and three years as Digital Account Manager, he joined Omnicom
Media Group in order to coordinate and implement Renault EMEA's digital strategy. Having arrived at
Havas Media in 2012, he assists clients in developing their digital media strategy.

Havas Media
Irne Labus Director of Digital Analytics
After more than 15 years in business and marketing roles with Pure Players Internet, Irne Labus
joined Havas Digital in 2008 to develop the analytics section. In this role, she supports advertisers in
exploiting and analyzing their campaign data and helps them translate their business objectives into
strategies for harvesting and visualizing data. Irne is a graduate from Paris Dauphine (master's in
management science) and the IAE Paris (master's in marketing and commercial practice).

ID Logistics
Emmanuel Vexlard Director of IT Systems
Graduating with a master's in logistics from LSLI Bordeaux, Emmanuel Vexlard started his career
with Habitat as transport manager in 1996 before taking the role of transport and distribution
manager at Geodis Logistics in Taiwan in 1999. He then joined the logistics procurement directorate
at Rhodia. He joined ID Logistics in 2005 as operations director for La Flche Cavaillonnaise and group supply
chain director for the service provider. In 2011, he took charge of the new IT department for the logistics service
provider with the task of foreseeing developments in IT systems and steering IT projects in France and abroad.

128

Contributors

IFR Group
Fabrice Benaut Director of IT Systems
Holder of a master's degree in information technology (IAE University Sorbonne) and marketing
(HEC), Fabrice Benaut joined the GfK group in 1983. He has been in charge of IT services for 15 years
and is involved in a large number of projects to transform the group. Prior to taking the reins for
global IT systems at IFR in January 2011, Benaut was IT Director for GfK Retail and Technology France. Outside of
the company, he has earned several certifications: Microsoft (1997), Business Objects (SAP in 2003), Informatica
(2009) and QlikView (2012), and he coordinates the Informatica users club (5 years at CAB).

IMEC
Yves Daelmans ICT Project Manager
After a dozen years of experience in publishing, Yves Daelmans gradually drifted towards reporting,
data management and business intelligence. From 2006 to 2008 he was in charge of business
intelligence at Wolters Kluwer, the French subsidiary of the Dutch professional information
publisher. In 2008, he joined lInstitut des Mmoires de ldition Contemporaine (IMEC) as ICT Project Manager in
charge of the SAP program and all BI projects. He also coordinates IMEC's Virtual BI Skills Center (BICC).

Infopro Digital
Clment Delpirou Partner
Clment Delpirou is one of the partners of the Infopro Digital Group. He is in charge of the group's
online activity. Working exclusively in B2B, the 200 digital experts who work at Infopro control 60
sites around three business models: generation of leads, media and expert databases. The 3.5 million
monthly hits and the hundreds of thousands of activity leads each year make Infopro a key online player in B2B in
France. A graduated from HEC, Clment was Director General of Nurun France prior to joining Infopro.

129

Contributors

Ingenico
Dorina Ghiliotto-Young Head of Innovation
Dorina Ghiliotto-Young joined the Global Solutions, Sales and Marketing Directorate at Ingenico as
Head of Innovation. Previously she was in charge of business development at Mappy SA (France
Telecom), where she successfully launched GPS navigation with European mobile manufacturers
and operators. She also developed the e-commerce activity for the Amadeus group and QCNS (leading online
tour operator). Dorina gained a master's at the prestigious Italian university Bocconi-IULM and added to it in 2003
with a full-time MBA in marketing, innovation and high-tech at the EDHEC.

Keyrus
Vincent Lagorce Practice Practice Manager Data Recovery
Vincent Lagorce has more than 10 years experience in business intelligence. After having worked as
a Pre-Sales Consultant, Project Leader and Project Director, he is currently piloting and developing
the business unit data recovery activity. Based on strong expertise and knowledge of the market, his
team supports clients when selecting solutions and implementing data discovery projects.

LOral
Laurence Kerjean Global Digital Manager DPP
Laurence Kerjean started her career in 2002 at Louis Vuitton as head of CRM before leaving for
London as a Direct Marketing Manager at Selfridges, a leader in the luxury retail sector in Great
Britain. Since 2005, Laurence's career has been exclusively at LOral (in particular as Events and
Trade Marketing Project Leader for the Redken brand, Head of Group for LOral Professional and Operational
Marketing Director for the Krastase, Kraskin et Shu Uemura brands in Europe). In 2011 she joined the Business
Development group of the DPP and worked towards developing business education content for the group's eight
professional brands before being appointed DPP Global Digital Manager in September 2012.

130

Contributors

Lagardre Active
Pierre-Olivier Sicamois Head of Advertising
A trained financial engineer, Pierre-Olivier Sicamois started his career path at Lagardre Active
Publicit in 2004 as Project Leader. After having worked for some years on project management at
Lagardre Active, he was appointed head of IT for advertising in 2007. In this role, he leads Lagardre
Publicit's technology projects. Since 2011, he was also in charge of several cross-cutting projects for the
technology directorate.

Orange
Catherine Ramus Team Design Project Leader
Catherine Ramus is currently working at Orange in the integrated design section within the
Advanced Design team at the Technocentre, a team that created Orange's One Concept, awarded a
star last year by l'Observeur for design. Trained as an optoelectronics engineer, since 1996 she has
lead numerous projects in dierent innovation groups at France Tlcom. She has also spent two years in R&D in
San Francisco. Since 2005, she has been working with designers and recently undertook the Innovation by
Design training at ENSCI (cole Nationale Suprieure de Cration Industrielle).

Orange
Lamia Delenda Project Management Oce
A graduated from the cole Nationale Suprieure des Tlcommunications, Lamia Delenda has been
the Project Management Oce Director at the Orange Group's Management Network. She runs all of
the group's network architecture and design projects, allowing her to foresee trends and possible
developments for the group's business models when encountering new markets. Lamia previously worked in
network strategy roles and in the field of regulatory forecasting.

131

Contributors

Pagesjaunes.fr
ric Lajarige Head of Search Projects
A graduate from lcole centrale de Lille and HEC, Lajarige started his career in 2004 at Prosodie as
Product Manager. He then joined VSC Technologies (technology subsidiary of Voyages-sncf.com) as
Operational Consultant and participated in numerous online projects in the field of rail distribution
and travel. From 2008 to 2010, Eric oversaw the ticketless transport project at Voyages-sncf.com as well as the
launch of the first mobile applications. Since 2010, he has been the Head of Search Projects within the search and
data operations directorate at PagesJaunes.fr.

Pagesjaunes.fr
Richard Coffre Senior Business Analyst
Senior Business Analyst at the Directorate for Advertisers Production at PagesJaunes, Richard Core
was a pioneer in implementing data visualization tools at PagesJaunes. He supports the
implementation and control by reworking activity reports and implementing dashboards. A
proponent of flexible methods (in particular Scrum), he teaches project management at the European
Communication School of Paris. Convinced of the need to bring information closer to the user, he has opted for a
data visualization approach to KPIs. His guiding principle is Keep It Simple and Smart (KISS).

Philips
Menno Haijma Senior Global Search Manager
A graduate from the Dutch Rijksuniversiteit Groningen, Menno Hailima started an international
career in online marketing in 2005. After having studied the market in Denmark at Ventus
Publishing, followed by a post at NDC|VBK Publications in the Netherlands, he joined Google in
Ireland as a Senior Account Manager for the Benelux market. In 2008, he joined the management
team of Google Wroclaw in Poland as team strategist. Since 2011, he has been the Senior Global Search Manager at
Philips Consumer Lifestyle in the Netherlands.

132

Contributors

Prisma Media
Jrme Tharaud Head of Research and Development
After studying artificial intelligence, Jrme Tharaud started his career in 1992 as an engineer at an
IT consultancy. In 1994, he joined an IT consultancy specializing in BI, where he became Project
Manager/Consultant. He took the lead in implementing data warehousing and supported the
development of BI for various customers. In 2000, he joined the Prisma Media group as HR Project Manager,
where he led a reworking of HR IT until it was outsourced. Since 2006, he has been in charge of the group's
research and development.

Roambi
Thibaut de Lataillade Vice President for Southern Europe
Thibaut de Lataillade joined Roambi in 2011 as Vice President for Southern Europe. His prior
experience in management, sales, marketing and consulting on strategy puts him in an ideal
position to assist clients to understand the challenges of mobility when applied to business realities.
Thibaut has a diploma from the European Business School (EBS) in Paris, finance option.

Schlumberger
Olivier Brousseau Business Intelligence Program Manager
A graduate of the cole Suprieure d'lectricit de Paris, Olivier Brousseau worked for three years at
the Sema group, a consultancy specializing in providing IT integration and three years as a software
architect for Atos Origin before joining Schlumberger in 2005. Since 2012, after several years of
experience in the field of IT architecture, he has led the group's BI program.

133

Contributors

SFR
Franois Nguyen Director for Decision-Making IT and Relationship Marketing
A trained accountant, Franois Nguyen started his career in the auditing department at Deloitte
Touche Tohmatsu (1993-1997) and then joined SAP France as a consultant. Since 2001, he has held
various roles within the SFR group: Head of MOA DAF (2001-2003), Director MOA FinancesProcurement (2003-2006), Director of Finance-Procurement Logistics Solutions (2007-2008) and Director of
Internal IT (2008 -2011). He was appointed Director for Decision-Making IT and Relationship Marketing in 2011
and holds roles associated with client knowledge, the recommendation of products and reporting tools for
channels.

SM Marketing Convergence Inc.


Valente Corpuz Garin Assistant Vice President IT Information Technology Division
With a degree in computer science from La Salle University (Manila), Valente Corpuz Garin has more
than 20 years experience in IT. Since 2007, he has been the head of the IT Division at Marketing
Convergence Inc. SM Advantage (58 people, annual budget of 4.57 million euros). Before taking up
this role, Valente worked in project management and the implementation of solutions, in particular
at Oracle.

SNCF
Maguelonne Chandesris Head of the Statistics, Econometrics and Data Mining Team, Innovation
and Research Directorate
Graduate of the cole Nationale de la Statistique et de lAnalyse de lInformation (ENSAI) and Doctor
of Mathematics at the University of Paris VI (2005), Maguelonne Chandesris is in charge of a team
specializing in statistics, econometrics and data mining at the Innovation and Research Directorate at SNCF.

134

Contributors

SPF Economie
Caroline Denil Data Warehouse Program Manager
Prior to joining the public economy service in 2008, Caroline Denil worked for more than 10 years
for an insurance company. She gained many years of experience in the field of data warehousing
and business intelligence. Based on this expertise, she was awarded the SAS Best Business Analytics
prize in 2012 for excellence for the Ecozoom project.

STMicroelectronics
Guillaume Deschamps FMT / Central Functions YTI / EDA Manager
Guillaume Deschamps has spent all of his career at STMicroelectronics, first as a product engineer
before becoming involved in the group's optimization process. In 2007, he became the data
management project leader in charge of rolling out the company's new data analysis tool in
Singapore and then the creation of tools to support the transfer of five technologies from the United States to
Singapore. In 2009, he was appointed Head of Security and Client Interfaces at Wafer Foundries Outsourcing
before taking the lead for data analysis activities for the whole internal foundry part of STM (20,000 people).

Ubisoft
Charles du Rau Director of the Consumer and Market Knowledge Department
Charles du Rau joined the Ubisoft group in 2003 to form the consumer studies department. He
succeeded in introducing staggered study programs in key stages of the life cycle of new products.
Based on experience gained in London at Procter & Gamble, he introduced innovative
methodologies adapted to the world of games. In 2008, together with the Canadian oce, he formed a new team
dedicated to tracking in games, his introduction to the field of analytics. More recently, he reworked market
studies to assist production, finance and the general management steer towards digital sales.

135

Contributors

Voyages-sncf.com
Frdrique Ville Innovation Director
Frdrique Ville's objective is to develop innovation and new products within Voyages-sncf.com. A
graduate of the cole Polytechnique and of the cole Suprieure des Tlcommunications,
Frdrique started her career in 1994 within the TRT-Philips Communications group. She then joined
the National Networks Division at France Telecom before going on to Branche Enterprises Major Projects and
International Directorate as head of the technical department until 2001. Ville then joined the Orange group and
then Orange Experience at the Technocentre and finally became performance and process management director
within the Orange Business Services division.

X Prime Group
Pierre Pongi Project Director
Pierre Pongi obtained his master's degree in marketing and information and communication
technology at the IAE in Monpellier which led him to the post of project leader at BuzzParadise in
2007, an online advertising agency of the Vanksen group. After having developed his digital expertise
in numerous fields, he arrived at the X-Prime group in 2010. As a project director at this independent
online agency, he advises and supports his clients on a daily basis in various matters related to their business:
digital presence of brands, information structure and user experience, speaking on social media, e-commerce
optimization and recommendation.

136

SAS
1

SAS is a world leader in decision making and business


analytics solutions. Thanks to its innovative solutions, SAS
assists its clients at more than 70,000 sites around the world,
improving their performance and creating value by allowing
them to quickly make better decisions.
SAS is the only software company to propose an open and
integrated company decision-making platform covering
all functionalities of the decision-making process: data
governance, analysis, recovery (reporting/business
intelligence) and exploratory visualization of data. The
power of its solutions is based on a unique range of analysis
tools (standard, ad hoc, OLAP, alert, statistical, forecasting,
modeling, prediction and optimization) and strong
experience in the profession.

SAS solutions cover all business divisions finance,


commercial, marketing, procurement, risk, piloting IT
systems, human resources, etc., - and respond to the specific
features of multiple business sectors.
Founded in 1976, SAS is the most significant independent
developer in the decision-making market with revenue of
US$3.02 billion in 2013. Its solutions are used by 91 of the 100
top companies listed in Fortune 500.

sas France
Domaine de Grgy
Grgy-sur-Yerres
77257 Brie-Comte-Robert
www.sas.com/france
Contact:
Patricia Machet,
Head of Strategic Marketing
Patricia.Machet@sas.com
+33 1 60 62 12 38

137

SAS
1

Serge Boulet
Director of Marketing and Communication - SAS France
A trained engineer, Serge Boulet started his career as a manager in the finance sector. He joined SAS France in 1985
and participated in initiating activities for the French subsidiary. There he performed several commercial, business
development and marketing roles. Over the years he has actively contributed to the growth of the subsidiary in the
French market and has supported clients in developing innovative projects and rolling out SAS solutions in fields such as risk
control, multichannel marketing and all fields related to business analytics. His knowledge of business and IT requirements for
BI applications has enabled him to develop a particularly relevant approach and outlook to help organizations get the best value
from their data. Contact: Serge.Boulet@sas.com

Jrme Cornillet
Head of SAS Business Analytics - SAS France
A university-trained statistician and forecaster with an early career dedicated to analyzing therapeutic experiments
in the major pharmaceutical laboratories, Jrme Cornillet joined SAS France in 1990. With an in-depth knowledge
of the professional problems encountered by clients through consultancy and training assignments, and a passion
for new technology, he is now the head of SAS Business Analytics products bringing together the components of SAS' softwares
analytical infrastructure. Contact: Jerome.Cornillet@sas.com

Mouloud Dey
Director of Solutions and Emerging Markets - SAS France
A graduate in IT and mathematical economics from the University of Paris-Ouest, and in marketing from the
University Paris Dauphine, Mouloud Dey has devoted his career to the information and communications
technologies sector with consultancy and management assignments covering the main areas of decision-making
IT. His understanding of the strategic requirements of organizations allows him to become a trusted partner in the largest
organizations in various analytical fields. He regularly brings his expertise and personal outlook to innovative subjects
associated with the digital era: big data, social networks and media, digital marketing, client experience, etc. Contact:
Mouloud.Dey@sas.com

Behind these experts, there is a whole team at SAS that produced this book by way of
contributions, advice, vision and experience. We would like to thank Ariane LigerBelair Sioufi, Patricia Machet, Elisabeth Ziegler, Luca Garlanda, and all those whose
contributions enabled this book to be published.
138

ebg

EBG is France's largest professional community that brings together company decision

makers active in industry, services, the media, etc., and whose common goal is
innovation.
EBG's board:

EBGs activities

Stphane Richard
CEO of France Tlcom Orange

EBG's aim is to coordinate a network of skills and know-how


by organizing more than 120 round table discussions each
year that allow executives to share their experiences.

Didier Quillot
Chairman of Coyote System
Jean-Bernard Levy
Chairman of Vivendi until June 2012
Franois-Henri Pinault
Chairman of PPR
Wu Janmin
Chairman of the Beijing Diplomatic Institute, Chairman of
the International Bureau for Exhibitions
Steve Ballmer
CEO of Microsoft

EBG has 10 active communities: The CEOs/General


Managers, Procurement Directors, Heads of IT and Decision
Making, Finance Directors, Heads of Legal and HR,
Marketing and Online/e-commerce Marketing Directors,
Media and Mobility Directors and Heads of Communication,
etc.
Each community meets at least once a month to discuss and
focus on new practices. The sessions are broadcast live on
ebg.net.

Sir Martin Sorrell


Chairman of WPP
Philippe Rodriguez
Treasurer
Pierre Reboul
Secretary General

Pour plus dinformations


sur lebg :

Benjamin Glaesener
Managing Director
+33 1 45 23 05 12
benjamin.glaesener@ebg.net

Arthur Haimovici
Head of Studies
+33 1 48 01 65 61
arthur@ebg.net

139

Printing: Corlet Imprimeur Graphic design/model: Thierry Badin @ Hite design graphique
Copyright 2014 SAS - EBG-Elenbi
All rights reserved. This work may not be reproduced in any way, in part or in whole, in
whatever format or by mechanical or electronic means, including by data storage and its
transmission by electronic methods, without authority from the publishers, Elenbi and
SAS.
Brands are used without intending to advertise them. EBG and SAS are unable to accept
liability for accidental errors or omissions that may occur in this work despite their best
care and control.
Electronic Business Group/Elenbi and SAS.

Das könnte Ihnen auch gefallen