Download Analytics Magazine PDF

TitleAnalytics Magazine
File Size6.0 MB
Total Pages118
Document Text Contents
Page 1

H T T P : / / W W W. A N A LY T I C S - M A G A Z I N E . O R G




Executive Edge
Verisk Innovative
Analytics President
Marty Ellingsworth
on the future, big data
and bigger analytics

• Eighteen things nobody tells you
about solo practice

• Certification: What it means
for employers, practitioners

• Analytics-driven culture:
Why it’s a corporate

• Predictive analytics in the cloud

• Analytics & health management

• Dealing with missing values in data

Special Supplement:
CAP Candidate


Page 2

W W W. I N F O R M S . O R G2 | A N A LY T I C S - M AGA Z I N E . O R G

Big dreams, small data

Like everyone else involved in the
analytics space, we’ve been yapping end-
lessly about “big data” in this column. You
all know the story – unfathomable amounts
of data coming in from multiple sources at
incredible speed have analysts everywhere
scrambling to make sense of it all. Let’s face
it, big data is the elephant in the room in any
discussion of analytics, and the elephant is
only going to get bigger (think hybrid data, in-
cluding video, images, sound, text, etc. from
countless sources and sensors).

But wait, there’s more; there’s a “small”
angle to the “big data” story. Even in the Big
Data Era, many companies do not have the
data they need to make data-based deci-
sions. A start-up, for example, almost cer-
tainly does not have the historic data that
an established firm has collected. Even
well-established companies probably lack
the data they need when considering intro-
ducing a new product or service or entering
a new market.

With that in mind, Analytics magazine
will launch a new column by Brian Lewis in
the March/April issue that will address the
issue of insufficient data and how to over-
come it. The name of the column: “Big Data
Dreams, Small Data Reality.” Chew on that
concept for a minute.

Lewis, chief data scientist and co-found-
er of Fractal Sciences, provides more de-
tails in an introductory column in this issue.

Of course, big data remains the big
fish in the analytics pond, so we’ll continue
to cover it and all of its ramifications. For
example, in this issue’s Executive Edge
column, Marty Ellingsworth, president of
Verisk Innovative Analytics, discusses the
“promise of big data and bigger analytics”
that “will drive the future” as the corporate
world shifts from a company-centric to a
customer-centric culture.

Meanwhile, INFORMS, publishers of
Analytics magazine and the world’s lead-
ing organization for high-end analytics, will
present its inaugural INFORMS Confer-
ence on Big Data in San Jose, Calif., June
22-24. The conference will focus on the
business of big data and making the jour-
ney from data-rich to decision-smart. For a
preview of the conference, click here.

The issue also includes a couple of “ca-
reer-builder” feature articles that should pique
the interest of any analytics professional
looking to get an edge in a competitive en-
vironment. Veteran analyst Doug Samuelson
outlines some of the consulting lessons he’s
learned the hard way, while Polly Mitchell-
Guthrie and Scott Nestler give an update on
INFORMS’ Certified Analytics Professional
program and how it can help employers and
clients of analytics professionals, as well as
analytics professionals themselves.

[email protected]

Page 59

J A N U A R Y / F E B R U A R Y 2 014 | 5 9A N A L Y T I C S

decide whether a sufficient number of re-
cords remain for the analysis to produce
meaningful results. The following example
illustrates how the true distribution can be-
come distorted when the source of missing
values is not identified properly.


(Or missing values in the “age variable”
in the customer database of a telephone

My Aunt Susanne purchased her
phone in the mid-60s. Her date of birth

was not collected at that time as the phi-
losophy of “know your customers” and
the need for customer data was nowhere
near as vital then as it is today. Things
changed in the 1990s with the deregula-
tion of the telecommunications market;
suddenly, the analysis of customer be-
havior became important. Since then, it
has become mandatory for customers to
provide their date of birth on a new con-
tract or with a contract change. My aunt,
however, never changed her contract type
or answered any customer questionnaire.

Topics areas:

» Accessing Data

» Understanding Raw Data

» Cleaning & Transforming Data

» Exploring & Visualizing Data

» Dimension Reduction

This course will be held
Boston, MA – March 28-29, 2014
San Jose, CA – June 25-26, 2014

Topic areas:

» Problem Framing

» Developing the Work Plan

» Testing Recommendations

» Presenting Results

» Impact Assessment

This course will be held
Atlanta, GA – February 20-21, 2014
Boston, MA – March 28-29, 2014
San Jose, CA – June 20-21, 2014



continuing Learn more about these courses at:

Next time I get a visionary assignment
that needs some clarity, I’ll be using what
I learned in this course to work towards
a great solution!
- Caroline Alexander,
Fed Ex Corporation

INFORMS Continuing Education
program offers intensive, two-day
in-person courses providing analytical
professionals with key skills, tools, and
methods that can be implemented
immediately in their work environment.

The datasets on which the
exercises are based are taken
from real-life scenarios, are
fun to work with and very
challenging. The course
provides a general framework
for tackling data analysis and
the instructors highlight the
pitfalls one can made along
the process.
- Ivan Hernandez,
Stevens Institute
of Technology

Page 60


W W W. I N F O R M S . O R G6 0 | A N A LY T I C S - M AGA Z I N E . O R G

Thus, the field “date of birth” is missing in
the customer database of her phone pro-
vider, and we can assume she is not the
only customer with a missing value.

If an analyst now looks at the distri-
bution of variable “age” in this customer
database, he might get a histogram as
shown in Figure 1. Additionally, he will
see that he has 9.1 percent missing val-
ues. The question is how to treat these
missing values.
• Shall the mean be used as imputation

• Shall different imputation values be

sampled from the actual distribution?
In our case, we can assume that the

true age value for Aunt Susanne and her
friends is not distributed over the whole

range of values. After a certain year it was
mandatory to provide the date of birth with
new contracts. So the missing values will
mostly occur for a certain age segment
(the older customers) and probably also for
a certain behavior segment (those who did
not change their contract type).

In the Figure 2 histogram, the true distri-
bution of the unknown age values is shown
in red. We realize that we would make a
wrong assumption when we treat the miss-
ing values as random, as we found out that
there is a systematic pattern behind them.
In order to qualify such a situation correctly,
business and process knowledge is need-
ed. This know-how is also important to
formulate an adequate imputation rule as
the imputation values should be from the

Figure 1:Distribution of variable “age”
in a customer database.

Page 117

NOTE: if you are an accredited public institution,
INFORMS will list your courses free of charge.
We are a public institution and hold accreditation
from _______________________________________

member colleague website contact from INFORMS other_______________________________________

Name of Organization

Address of Organization

Recognized Analytics Continuing Education Provider Form
To be recognized by INFORMS for training/education in the analytics profession,
complete the application below and return it with appropriate fee(s) and items listed
below. The completed form must be returned to INFORMS by the contact in one of
the following methods:

1. The preferred method is to scan the completed form and send it as PDF to the
INFORMS Certification Manager at [email protected]
2. Return the completed signed form to INFORMS by postal mail. Please form and
attachments to:
Certification Manager
5521 Research Park Drive, Suite 200
Catonsville, Maryland 21228 USA
3. Or you may fax form to Louise Wehrle at 443.757.3515.

As part of the application, provider must provide examples of:
Certificate granted to student
Brochure/website copy/advertising of courses
Registration requirements
Refund policy
Instructor/author qualifications
Facilities management policy
Org chart of organization

For further information on the
CAP™ credential, please visit

Point of Contact Name Point of Contact Phone number

Please PRINT the following information.


Point of Contact Email

Years of Operation Number of Courses

Names of Courses

Description of Courses

Instructor(s) Name & Qualifications

How did you hear about listing/CAP™/INFORMS?

Fees: $300 non-refundable application fee per organization
$200 first-year approval fee per organization

Payment method:
_____ Check Enclosed (Make payable to INFORMS and must be drawn on U.S. bank in U.S. dollars. )
_____ MasterCard _____ Visa _____ American Express _____ Discover

Card # (15 digits) Expiration date

Name on Card Signature

Page 118


5521 Research Park Drive, Suite 200
Catonsville, Maryland 21228

Similer Documents