Professional Documents
Culture Documents
ORG
ALSO INSIDE:
JANUARY/ FEBRUARY 2013 DRIVING BETTER BUSINESS DECISIONS
Pick the right
analytics partner
Big data, analytics
and elections
Software survey:
decision analysis
WORKSOCIAL
Analyze This!
Vijay Mehrotra
seeks to make
sense of the
analytics world
New approach blends data,
process and collaboration for
better, faster decision-making.
BROUGHT TO YOU BY:
SPECIAL SUPPLEMENT: CERTIFIED ANALYTICS PROFESSIONAL CANDIDATE HANDBOOK
WWW. I NF OR MS . OR G 2 | A NA LY T I CS - MAGA Z I NE . OR G
Brave new mobile world
I NSI DE STORY
Pete Townshend of The Who was way
ahead of his time when he wrote Goin Mo-
bile more than 40 years ago, but the world
has fnally caught up. Today, seemingly ev-
eryone and their brother are goin mobile,
hooked on mobile communications.
In response to the growing number of
readers who prefer their reading material
served up via a mobile communication de-
vice, we begin the new year with a new look.
Instead of the landscape/horizontal format
weve employed for four years and which
worked well for desktop computer viewing,
weve switched to a portrait/vertical confgu-
ration more suitable for mobile apps, includ-
ing Apple iOS (iPad and iPhones) and the
Android (Google) platform, which includes
any Android device from phones to a variety
of tablets including Barnes & Nobles Nook,
Amazons Kindle Fire and Googles Nexus.
(Were investigating compatibility with the
new Windows phones.)
Keep in mind it takes a little time to
convert the finished issue and submit it
to Apple and Google for testing and ap-
proval, so give it a few weeks before you
try it. We look forward to your feedback.
The widespread use of mobile com-
munications, the social media it enables,
the enormous amount of data it produces
and the technology that drives it all has
given rise to a concept called worksocial,
the topic of this months cover story by
Samir Gulati. As Gulati explains, work-
social starts with the processes and data
that drive the business, and overlays the
innovations of social collaboration and
mobile access directly over those busi-
ness engines. The goal: better, faster
organizational decision-making.
Along with the usual variety of articles
designed to inform, enlighten and enhance
your analytical career (see, for example,
Ten things to consider when evaluating
analytics and decision sciences partners
by David Zakkam and Deepinder Singh
Dhingra or Modeling experience yields key
insights by Bruce W. Patty), this issue fea-
tures an article and a 20-plus page special
supplement on a new program that can ad-
vance your career: Certifed Analytics Pro-
fessional (CAP).
An important initiative of the Institute for
Operations Research and the Management
Sciences (INFORMS), CAP represents a
gold seal of approval for analytics profes-
sionals from the premier organization for
advanced analytics in the world. To fnd out
more about this ground-breaking program
and what it could mean for your career, visit
the CAP website.
PETER HORNER, EDITOR
peter.horner
@
mail.informs.org
What are customers saying
about AIMMS?
Kepler embraces optimization by using the AIMMS platform to generate solutions that provide significant business
value to our clients. Kepler is currently developing a series of large-scale, complex solutions for clients using our
Resource Allocation Model (RAM). A recent success was the development of a model that optimally assigned
resources to fulfill critical mission requirements for a Department of Defense Service. The model dealt with close to
a million variables and constraints; dramatically reducing the time to schedule and deploy a large workforce.
Dan Markowitz, Vice President Advance Programs at Kepler Research
www.aimms.com/kepler-research
KeplerResearch
Kepler Research provides government leaders with innovative
solutions in acquisition services, information technology and
advanced analytics.
Technology from AIMMS is a strategic component of Viridity Energys unique software solution, VPower.
VPower works within the existing operations of large energy users to give them the optimal demand
management strategy to reduce their overall energy spend. AIMMS provides the computing engine associated
with linear and non-linear computations that are a key component of Viridity Energys optimization process.
With support from AIMMS, VPower can give our customers customized decision-making tools to align their
operations with load management strategies, turning sustainability into a smart economic choice.
Audrey Zibelman, CEO & Founder of Viridity Energy
www.aimms.com/viridity-energy
Viridity Energy works with large energy users to create
a customized demand-side management solution that is based
upon the unique operations of our clients.
Our mission is to bring the benefits of optimization to
society. Our track record, customer base, and award
winning technology* prove that we are capable of doing
just that. We are confident that we can help you too.
Contact us and let us show you how AIMMS can improve
your business.
*2011 Franz Edelman Award
Bellevue, WA, USA Haarlem, the Netherlands Singapore Shanghai, China
+1 425 458 4024 +31 23 5 511 512 +65 6521 2827 +86 21 51160733
www.aimms.com ino@aimms.com
AIMMS is a registered trademark of Paragon Decision Technology B.V. Other brands and their products are trademarks of their respective holders
Contact us:
425-458-4024
info@aimms.com
ParagonDecision_Dec2012_ Shell 11/30/12 1:08 PM Page 1
WWW. I NF OR MS . OR G 4 | A NA LY T I CS - MAGA Z I NE . OR G
DRIVING BETTER BUSINESS DECISIONS
C O N T E N T S
FEATURES
WELCOME TO WORKSOCIAL WORLD
By Samir Gulati
New approach, technology blends data, process and
collaboration for better, faster decision-making.
HOW TO PICK A BUSINESS PARTNER
By David Zakkam and Deepinder Singh Dhingra
Ten things to consider when evaluating analytics and decision
sciences partners.
BIG DATA, ANALYTICS AND ELECTIONS
By George Shen
Obama teams all-front campaign leveraged Web, mobile, TV,
social media and analytics.
MODELING EXPERIENCE YIELDS INSIGHTS
By Bruce W. Patty
Lessons learned while working with the largest domestic
container eet in North America.
SOFTWARE SURVEY: DECISION ANALYSIS
By William M. Patchak
Relationship between technology, thoughtful analysis remains
essential to success of any software tool.
SPECIAL SUPPLEMENT: CAP CANDIDATE HANDBOOK
24
30
40
46
52
52
40
24
J ANUARY/ FEBRUARY 2013
Brought to you by
Tel 775 831 0300 Fax 775 831 0314 info@solver.com
ANALYTIC SOLVER PLATFORM
From Solver to Full-Power Business Analytics in Excel
The Excel Solvers Big Brother Has Everything You
Need for Predictve and Prescriptve Analytcs
From the developers of the Excel Solver, Analytc Solver
Platorm makes the worlds best optmizaton sofware
accessible in Excel. Solve your existng models faster,
scale up to large size, and solve new kinds of problems.
From Linear Programming to Stochastc Optmizaton
Fast linear, quadratc and mixed-integer programming is
just the startng point in Analytc Solver Platorm. Conic,
nonlinear, non-smooth and global optmizaton are just
the next step. Easily incorporate uncertainty and solve
with simulaton optmizaton, stochastc programming,
and robust optmizaton all at your fngertps.
Ultra-Fast Monte Carlo Simulaton and Decision Trees
Analytc Solver Platorm is also a full-power tool for
Monte Carlo simulaton and decision analysis, with a
Distributon Wizard, 50 distributons, 30 statstcs and
risk measures, and a wide array of charts and graphs.
Comprehensive Forecastng and Data Mining
Analytc Solver Platorm samples data from Excel,
PowerPivot, and SQL databases for forecastng and data
mining, from tme series methods to classifcaton and
regression trees, neural networks and associaton rules.
And you can use visual data exploraton, cluster analysis
and mining on your Monte Carlo simulaton results.
Find Out More, Download Your Free Trial Now
Analytc Solver Platorm comes with Wizards, Help, User
Guides, 90 examples, and unique Actve Support that
brings live assistance to you right inside Microsof Excel.
Visit www.solver.com to learn more, register and
download a free trial or email or call us today.
6 |
DRIVING BETTER BUSINESS DECISIONS
REGISTER FOR A FREE SUBSCRIPTION:
http://analytics.informs.org
INFORMS BOARD OF DIRECTORS
President Anne G. Robinson, Verizon Wireless
President-Elect Stephen M. Robinson, University of
Wisconsin-Madison
Past President Terry Harrison, Penn State University
Secretary Brian Denton,
University of Michigan
Treasurer Nicholas G. Hall, Ohio State University
Vice President-Meetings William Bill Klimack, Chevron
Vice President-Publications Eric Johnson, Dartmouth College
Vice President-
Sections and Societies Paul Messinger, University of Alberta
Vice President-
Information Technology Bjarni Kristjansson, Maximal Software
Vice President-Practice Activities Jack Levis, UPS
Vice President-International Activities Jionghua Judy Jin, Univ. of Michigan
Vice President-Membership
and Professional Recognition Ozlem Ergun, Georgia Tech
Vice President-Education Joel Sokol, Georgia Tech
Vice President-Marketing,
Communications and Outreach E. Andrew Andy Boyd,
University of Houston
Vice President-Chapters/Fora Olga Raskina, Con-way Freight
INFORMS OFFICES
www.informs.org Tel: 1-800-4INFORMS
Executive Director Melissa Moore
Meetings Director Teresa V. Cryan
Marketing Director Gary Bennett
Communications Director Barry List
Headquarters INFORMS (Maryland)
7240 Parkway Drive, Suite 300
Hanover, MD 21076 USA
Tel.: 443.757.3500
E-mail: informs@informs.org
ANALYTICS EDITORIAL AND ADVERTISING
Lionheart Publishing Inc., 506 Roswell Street, Suite 220, Marietta, GA 30060 USA
Tel.: 770.431.0867 Fax: 770.432.6969
President & Advertising Sales John Llewellyn
john.llewellyn@mail.informs.org
Tel.: 770.431.0867, ext.209
Editor Peter R. Horner
peter.horner@mail.informs.org
Tel.: 770.587.3172
Art Director Lindsay Sport
lindsay.sport@mail.informs.org
Tel.: 770.431.0867, ext.223
Advertising Sales (A-L) Peter Fenlon
pete.fenlon@mail.informs.org
770.578.8950
Advertising Sales (M-Z) Sharon Baker
sharon.baker@mail.informs.org
Tel.: 813.852.9942
Analytics (ISSN 1938-1697) is published six times a year by
the Institute for Operations Research and the Management
Sciences (INFORMS). For a free subscription, register at
http://analytics.informs.org. Address other correspondence to
the editor, Peter Horner, peter.horner@mail.informs.org. The
opinions expressed in Analytics are those of the authors, and
do not necessarily refect the opinions of INFORMS, its offcers,
Lionheart Publishing Inc. or the editorial staff of Analytics.
Analytics copyright 2013 by the Institute for Operations
Research and the Management Sciences. All rights reserved.
DEPARTMENTS
Inside Story
Prot Center
Analyze This!
Viewpoint
Certied Analytics Professional
Conference Preview
Industry Notes
Five-Minute Analyst
Thinking Analytically
2
8
12
16
22
60
64
66
70
70
Tel 775 831 0300 Fax 775 831 0314 info@solver.com
ANALYTIC SOLVER PLATFORM
Easy to Use Predictive and Prescriptive Analytics in Excel
How can you get results quickly for business decisions,
without a huge budget for enterprise analytcs
sofware, and months of learning tme? Heres how:
Analytc Solver Platorm does it all in Microsof Excel,
accessing data from PowerPivot and SQL databases.
Sophistcated Data Mining and Predictve Analytcs
Go far beyond other statstcs and forecastng add-ins
for Excel. Use classical multple regression, exponental
smoothing, and ARIMA models, then go further with
regression trees, k-nearest neighbors, and neural
networks for predicton, discriminant analysis, logistc
regression, k-nearest neighbors, classifcaton trees,
nave Bayes and neural nets for classifcaton, and
associaton rules for afnity (market basket) analysis.
Use principal components, k-means clustering, and
hierarchical clustering to simplify and cluster your data.
Simulaton, Optmizaton and Prescriptve Analytcs
Analytc Solver Platorm also includes decision trees,
Monte Carlo simulaton, and powerful conventonal and
stochastc optmizaton for prescriptve analytcs.
Help and Support to Get You Started
Analytc Solver Platorm can help you learn while
getng results in business analytcs, with its Guided
Mode and Constraint Wizard for optmizaton, and
Distributon Wizard for simulaton. Youll beneft from
User Guides, Help, 30 datasets, 90 sample models, and
new textbooks supportng Analytc Solver Platorm.
Analytc Solver Platorm goes further than any other
sofware with Actve Support that alerts us when youre
having a problem, and brings live assistance to you right
where you need it inside Microsof Excel.
Find Out More, Download Your Free Trial Now
Visit www.solver.com to learn more, register and
download a free trial or email or call us today.
J A NUAR Y / F E BR UAR Y 2013 | 9 J A NUAR Y / F E BR UAR Y 2013 | 9 WWW. I NF OR MS . OR G 8 | A NA LY T I CS - MAGA Z I NE . OR G
What does a typical analytics professional do?
What is his or her background?
Interesting questions questions that would ben-
eft from the application of analytics. Thats what the
company Talent Analytics has done in cooperation
with the International Institute for Analytics.
Given that many people who practice analytics dont
have analytics in their title, fnding an appropriate sam-
ple of practitioners wasnt an easy task. Talent Analytics
put together its sample through business contacts and
through names gathered at analytics-oriented conferenc-
es. Individuals who said their only involvement with ana-
lytics was the use of spreadsheets were struck from the
list of names, leaving a sample of just over 300 people.
A dyed-in-the-wool statistician wouldnt be satis-
fed with this approach, and any conclusions based
on the sample need to be placed in the context of
how it was obtained. Nonetheless, it provides the only
coordinated effort of which Im aware to fnd out who
analytics professionals are and what they do.
Talent Analytics examined the data in a variety of
ways. I wont seek to cover everything they shared
with me it was quite extensive but a number of
fndings did stand out.
The analytics professional
BY E. ANDREW BOYD
Professionals employed in
the field of analytics are
young and mobile. Of
those who responded, 45
percent had been in the
workforce for less than 10
years.
PROFI T CENTER
J A NUAR Y / F E BR UAR Y 2013 | 9 A NA L Y T I C S
Not surprisingly, analytics profession-
als are well trained. Roughly half indicat-
ed their highest level of education was a
masters degree, 16 percent held doctor-
ates and virtually everyone had complet-
ed college.
The backgrounds, however, were
quite varied. Respondents were allowed
to check multiple boxes describing their
education. So, for example, an individual
who held an undergraduate degree in
liberal arts and then completed gradu-
ate work in business could check both
liberal arts and business. Not surpris-
ingly, mathematically oriented disciplines
were well represented, with 120 respon-
dents indicating they held degrees in the
category mathematics/statistics and an-
other 60 holding degrees in operations
research/engineering. Other degrees
made strong showings as well. Fully 108
respondents held degrees in business,
though degrees in fnance and econom-
ics were scant at 11. Interestingly, liberal
arts degrees edged out those in comput-
er science 71 to 69.
YOUNG AND MOBILE
As is the case in many technical dis-
ciplines, professionals employed in the
feld of analytics are young and mobile.
Of those who responded, 45 percent
had been in the workforce for less than
10 years, while only 9 percent had been
working for at least 30 years. A little more
than half had been with their current
employer for less than three years com-
pared with 7 percent for at least 10 years.
The data also supported the increasing
recognition of analytics as its own dis-
cipline, since the time respondents had
been employed as an analytics profes-
sional was, on average, far shorter than
the time theyd been employed.
To determine where analytics profes-
sionals devote their time, respondents
were asked how long they spent on
various activities, from analyzing data to
managing people. Based on the respons-
es and, of course, employing analytics,
Talent Analytics grouped respondents
into one of four functional clusters:
1. Data Preparation. Time largely spent
acquiring data and preparing it for
analysis.
2. Programmer. Time largely spent
developing software to perform data
analysis.
3. Manager. Time spent performing
general management and
administration, designing analyses,
interpreting results and presenting
conclusions.
4. Generalist. Time spent doing a little
bit of everything.
The clusters arent entirely surprising.
Anyone whos worked with analytics knows
J A NUAR Y / F E BR UAR Y 2013 | 11
WWW. I NF OR MS . OR G 10 | A NA LY T I CS - MAGA Z I NE . OR G
that programming and data acquisition/
preparation are as vital as knowing the in-
tricacies of various mathematical tools. And
anyone whos worked in a small frm knows
you get to do a little bit of everything.
EDUCATIONAL BACKGROUND
The educational background of peo-
ple in the different clusters provided
further insight into who analytics pro-
fessionals are. Individuals in the data
preparation cluster had more degrees in
business than anything else (33) followed
by mathematics/statistics (27), computer
science (17) and liberal arts (16). Gener-
alists were also led by business degrees
(57) with mathematics/statistics coming
in second (47) and liberal arts rounding
out the top three (35). The fact that busi-
ness degrees lead both of these clusters
helps emphasize that data acquisition/
preparation isnt just about putting num-
bers in a database. It requires navigating
organizations to discover data availabil-
ity, uncovering the actual meaning of the
data (is a sale an order or receipt of pay-
ment?) and arranging for the data to be
prepared for analysis.
Interestingly, the No. 1 degree held
by managers is mathematics/statistics
(17) followed by business (10), with
operations research/engineering and
liberal arts tied for third (nine). Pro-
grammers are led by mathematics/
statistics (29), with computer science
taking second position (22) and op-
erations research/engineering and lib-
eral arts again tied for third (11). The
strong showing by liberal arts degrees
in all clusters begs the question of how
these individuals find their way into the
analytics profession.
The work done by Talent Analytics
has given us a glimpse into the make-
up of analytics professionals. And while
there are some surprises, the overall
picture that emerges isnt a surprise at
all. Analytics professionals have a mix
of talents that span the technical to the
interpersonal. Theyre not afraid of num-
bers, and theyre not defned by a unique
educational background. All in all, the
medley of skills enjoyed by analytics pro-
fessionals is a good thing. Its just what
you want from a group of creative prob-
lems solvers.
Andrew Boyd, senior INFORMS member and
INFORMS VP of Marketing, Communications
and Outreach, has been an executive and chief
scientist at an analytics frm for many years. He
can be reached at e.a.boyd@earthlink.net. The
author thanks the senior management team at
Talent Analytics for taking the time to share the
results of the study.
PROFI T CENTER
Subscribe to Analytics
Its fast, its easy and its FREE!
Just visit: http://analytics.informs.org/
J A NUAR Y / F E BR UAR Y 2013 | 11
The Optmizaton / Simulaton Library that Speaks the
Languages of Desktop and Web Computng.
No other callable library matches the fexibility of
Solver SDK Platorm from solving Excel models on
your server, to solving models created in your browser.
Build Your Model Faster in Excel or Visual Studio.
Use Risk Solver Platorm or its subsets to build your
model in Excel. Use Microsof Visual Studio to create
your model in C++, C# or VB.NET. Use C, R, MATLAB,
Java, PHP or JavaScript its your choice.
Use Optmizaton and Simulaton as a Web Service.
Run the Solver Server that comes with Solver SDK on
any PC or server, then solve your models over the
wire using our standard Web Service, SOAP and WSDL
protocols. Get real-tme progress, or just disconnect
and retrieve the soluton later its your choice.
Solve Your Most Challenging Linear Mixed-Integer,
Nonlinear and Stochastc Optmizaton Models.
Solver SDKs wide choice of built-in and plug-in Solver
Engines handle every type and size of optmizaton
problem, from Gurobi and Xpress-MP to Mosek and
KNITRO. Use the SDKs built-in simulaton optmizaton
and stochastc programming algorithms, and support
for mult-core CPUs.
Download a Free Trial at www.solver.com.
Learn more and download a FREE trial version of Risk
Solver Platorm, Solver SDK Platorm, and our eight
plug-in Solver Engines from Solver.com
THE MOST FLEXIBLE WAY TO DEPLOY YOUR MODEL
SOLVER SDK PLATFORM
Tel 775 831 0300 Fax 775 831 0314 info@solver.com
J A NUAR Y / F E BR UAR Y 2013 | 13 WWW. I NF OR MS . OR G 12 | A NA LY T I CS - MAGA Z I NE . OR G
My wife and I recently hosted our annual holiday
party, in which a large and boisterous crowd of friends
and family members descended upon our house for
the evening. One part of this annual ritual is dropping
off fyers at nearby houses. The fyers serve both as
an invitation to join in the festivities and as a warning
to: (a) expect a lot of unfamiliar cars to appear, and
(b) anticipate a lot of noise coming out of our house
that night.
This year, a neighbor we had never met before
arrived at our house on the night of our party (thank
you for the invitation) and introduced himself to us
(Im Bill I live a few houses up the street). It turns
out that Bill is 65 years old and has lived his entire life
in his house on our (that is, his) street.
I took piano lessons in this house as a boy, he
explained, but it has been a long time since I have
been inside. And for the next hour, as we listened in
awe, Bill took us on a historical tour of our own house.
This was originally Poseys house, he began. We
were dumbfounded to discover that apparently the
original owner of our house was George A. Posey [1],
the chief engineer for the tube connecting Oakland,
Calif. (where we live) with the neighboring island-city
Survey seeks to make sense
of analytics world
BY VIJAY MEHROTRA
What has perhaps been
most surprising about the
interviews is the sheer
variety of things that
people in the field of
analytics are concerned
about.
ANALYZE THI S!
J A NUAR Y / F E BR UAR Y 2013 | 13 A NA L Y T I C S
of Alameda, only the second underwater
tunnel ever built in the United States.
Later on, Bill told us, the house was
owned by a tool and die maker of some
local renown, whose machine shop was
on the second foor of the detached ga-
rage (this is now my home offce, the very
room in which I now sit writing this col-
umn). As we walked through the house,
he pointed out where additions had
been made, where moldings had been
replaced, and where walls had been
knocked down. When he left, we had a
much better sense of the place in which
we live.
WE NEED YOU TO TAKE A
SURVEY
As it happens, I am also trying to get
a better sense of the professional world
that I live in, and I need you to help fll
out the picture. Under the umbrella of Ac-
centures Institute for High Performance,
I am partnering with Jeanne Harris (co-
author of Competing on Analytics and
Analytics at Work) on a detailed study
to examine the world of analytics today.
To date, we have done several one-
on-one interviews with analytics profes-
sionals from several different vertical
industries. What has perhaps been most
surprising about these interviews is the
sheer variety of things that people in the
feld of analytics are concerned about,
including such diverse topics as choos-
ing hardware platforms for data and
analysis, selecting software tools, clean-
ing data, integrating data from multiple
sources, setting up smart organizational
structures to support analytics, properly
prioritizing projects, collaborating effec-
tively with the IT department, impressing
business customers with insightful re-
sults, and dealing with the challenges of
employee retention amidst a perceived
shortage of analytic talent.
Based on these frank and thoughtful
conversations, we have put together a
survey that we are in the midst of rolling
out right now. We would like to invite you
to take our survey and tell us what your
view into the world of analytics looks like.
To take this survey, click here and follow
the instructions. Please forward this link
to everyone in your network who works
in the analytics feld. (You all know how
important getting good data is.)
Thanks in advance for your participa-
tion and for spreading the word about our
work. Jeanne and I look forward to gath-
ering the data, conducting our analysis
and sharing the results with the analytics
community in the near future.
BIG DATA, ADVANCED ANALYTICS
Along these same lines, the focus
of the October 2012 edition of the Har-
vard Business Review is big data and
J A NUAR Y / F E BR UAR Y 2013 | 15
WWW. I NF OR MS . OR G 14 | A NA LY T I CS - MAGA Z I NE . OR G
advanced analytics. Its cover reads Get-
ting Control of Big Data: How Vast New
Streams of Information are Changing the
Art of Management, and in a series of
three Spotlight articles by academic
and industry leaders, the journal makes
an admirable attempt at describing the
professional world that we live in for an
executive audience.
In Big Data: The Management Revo-
lution, MITs Andrew McAfee and Erik
Brynjolfsson point out that data and
analysis alone are not enough, asserting
that, in order to capitalize on the potential
of advanced analytics, companies must
also change the way in which the orga-
nization digests information and makes
decisions. The authors provide a descrip-
tion of many of the challenges of making
such changes, but also offer up a tanta-
lizing prize, citing some of their recent re-
search in which they show that the more
companies characterized themselves as
data-driven, the better they performed on
objective measures of fnancial and op-
erational results.
In Data Scientist: The Sexiest Job
of the 21st Century, Tom Davenport
(Harris co-author on Competing on
Analytics and Analytics at Work, and
also currently a visiting professor at
the Harvard Business School) and D.J.
Patil (data scientist in residence at ven-
ture firm Greylock Partners) note that
amidst all of the big data hype, there
is also significant confusion about the
people that turn all that data into value.
There is little consensus on where the
role fits in an organization, how data
scientists can add the most value, and
how their performance should be mea-
sured, they write. From here, they
go on to describe their prototype of a
successful data scientist (a hybrid of
a data hacker, analyst, communica-
tor and trusted advisor) and provide
high-level guidance on how to find, re-
cruit and manage data scientists.
Finally, McKinseys Dominic Bar-
ton and David Court open their article
Making Advanced Analytics Work For
You by asserting that Big data and
analytics have rocketed to the top of
the corporate agenda, and then pro-
vide a series of tips for executives anx-
ious to avoid making the same types
ANALYZE THI S!
Request a no-obligation INFORMS Member Benets Packet
For more information, visit: http://www.informs.org/Membership
J A NUAR Y / F E BR UAR Y 2013 | 15
A NA L Y T I C S
of mistakes made in the adoption of
previously hot technologies such as
CRM.
All interesting articles and all worth
reading. Perhaps more notable, I believe,
is that HBRs Editorial Board chose to
focus on this topic. For those who have
been working on this stuff for a long time,
it seems that (to quote an old cigarette
ad), youve come a long way, baby [2].
Looking forward to seeing whats next
in 2013.
Vijay Mehrotra (vmehrotra@usfca.edu) is
an associate professor in the Department of
Analytics and Technology at the University of San
Franciscos School of Management. He is also an
experienced analytics consultant and entrepreneur,
an angel investor in several successful analytics
companies and a longtime member of INFORMS.
http://jps.informs.org
Attention Companies Who Hire Analytics Professionals
Reserve your space now
for the industrys premier,
professional job fair!
Brought to you by Analytics Magazine
Find the seasoned professionals you need over 700 analytics professionals expected
Provide your recruitment materials in a casual setting
Arrange discreet on-site meetings in private booths
Enjoy discounted combination pricing with the fall Annual Meeting Job Fair
Enhance your visibility with an ad in Analytics or ORMS Today
jps@mail.informs.org or call (800) 4-INFORMs
Questions?
Analytics Connect
April 7-9, 2013 | Grand Hyatt San Antonio | San Antonio, Texas
Applying Science to the Art of Business
USINESS ANALYTICS &
PERATIONS RESEARCH
INFORMS CONFERENCE ON
REFERENCES
1. For more on George A. Posey and the Posey
tube, see www.alamedainfo.com/posey_tube.htm.
2. http://en.wikipedia.org/wiki/Virginia_Slims.
J A NUAR Y / F E BR UAR Y 2013 | 17 J A NUAR Y / F E BR UAR Y 2013 | 17 WWW. I NF OR MS . OR G 16 | A NA LY T I CS - MAGA Z I NE . OR G
In making election forecasts for the FiveThir-
tyEight blog (538) at the New York Times, Nate Silver
uses a statistical model that is subtle, sophisticated
and comprehensive. Real Clear Politics uses a shal-
low approach to forecasting that could have been
devised by a statistical Forrest Gump. But which fore-
caster better predicted the results in the 2012 presi-
dential election? Did the intellectual tortoise hold its
own against the hare?
From a conceptual standpoint, it should have been
no contest. In an approach that would make statisti-
cians shudder, Real Clear Politics (RCP) estimated
the Obama/Romney difference in a given state by the
simple average of differences in recent polls. Differ-
ences in sample sizes were ignored, the word recent
was defned differently in different states, undecided
voters were simply excluded, and evidence that some
polls skew toward Republicans and others toward
Democrats got no weight. The 538 model, in contrast,
avoided all these limitations, and took account of cor-
relations among outcomes in similar states and the
demographic makeup of each.
FROM THEORY TO PRACTICE
But how did the final state-by-state predictions
under the two approaches compare in accuracy?
Did Nate Silver beat the
tortoise?
BY ARNOLD BARNETT
The candidates and
everyone else recognized
that the outcome would be
determined by what
happened in about a
dozen swing states.
VI EWPOI NT
J A NUAR Y / F E BR UAR Y 2013 | 17 A NA L Y T I C S
RCP only made forecasts in 30 of the
51 states (including the District of Co-
lumbia), but these included all swing
states and all large states. And, at first
blush, it might appear that the race be-
tween the two methodologies in the 30
states was (in the familiar phrase) too
close to call.
The most obvious dimension for
comparison is the bottom line: Did the
forecast in a given state correctly iden-
tify the winner there? By that standard,
both methods did very well: In 29 of the
30 states, they agreed who the winner
would be and that candidate actually
won. (Complete data tables will ap-
pear in a longer version of this article
in the February 2013 issue of OR/MS
Today.) In Florida, neither forecaster
made a correct forecast: RCP errone-
ously projected a narrow Romney vic-
tory (1.5 percentage points), while 538
projected an exact tie (and thus ab-
stained from forecasting). Obama car-
ried Florida by 0.9 percentage points.
We can say, therefore, that 538 scored
a partial victory over RCP in one of
30 states, but that is hardly a decisive
advantage.
As for the absolute forecast errors
in the various states, the results were
once again similar. The mean absolute
error over the 30 states was 2.87 per-
centage points for RCP and 2.25 for
538. However, there is a blue state
bias among the 30 states: Romney
carried only 27 percent of them (eight
out of 30), while he captured 47 per-
cent (24 out of 51) in the entire nation.
When an adjustment is made for this
bias, the mean absolute error becomes
2.57 points for RCP and 2.33 for 538.
This revised difference of one-quar-
ter of one percentage point is hardly
decisive.
ON THE OTHER HAND
Yet this aggregate analysis is
oblivious to the central dynamic of the
2012 election. Given the realities of
the Electoral College, the candidates
and everyone else recognized that the
outcome would be determined by what
happened in about a dozen swing
states that either candidate could
plausibly win. In the other states, the
winner was a foregone conclusion so
there was little campaigning and little
interest in polling results.
Under the circumstances, a com-
parison between RCP and 538 should
focus primarily if not exclusively on their
Help Promote Analytics
Its fast and its easy! Visit:
http://analytics.informs.org/button.html
J A NUAR Y / F E BR UAR Y 2013 | 19
WWW. I NF OR MS . OR G 18 | A NA LY T I CS - MAGA Z I NE . OR G
accuracy in swing states. RCP identi-
fed 11 states as toss up just before the
election: Colorado, Florida, Iowa, Michi-
gan, New Hampshire, Nevada, North
Carolina, Ohio, Pennsylvania, Virginia
and Wisconsin.
Within these states, the two ap-
proaches differed markedly in per-
formance. 538 outperformed RCP in
absolute forecast accuracy in all but
one of the 11 swing states (Ohio).
Both forecasters were on average
more favorable to Romney than the
actual voters, but the net bias was
only 0.76 percentage points for 538
over the 11 states as opposed to 2.44
points for RCP. That difference of 1.68
(2.44-0.76) points is especially note-
worthy because regression analysis
makes clear that the 538s estimates
about Obamas performance were
consistently about 1.5 points higher
than those of RCP. Again and again,
this adjustment was vindicated by the
swing-state results: RCP underesti-
mated Obamas actual vote share,
while 538 eliminated roughly 75 per-
cent of the underestimation.
In the 19 states out of the 30 origi-
nally compared that were not swing
states, 538 and RCP performed about
equally well, which is why statistics
based on all 30 states yielded less
disparity between the two approaches
than the swing states alone. It could be
that Obama outperformed the swing-
state polls upon which RCP relied be-
cause of the major voter-turnout drives
that his campaign undertook in those
states, which brought many people
to the voting booths whom pollsters
had not included in tabulations about
likely voters. In the other states, the
Obama campaign may not have waged
such efforts, so no comparable surge
occurred.
Nate Silver would be the frst to agree
that his state-by-state forecasts were cor-
related, and that circumstance stymies
assessments of whether his swing-state
victory over RCP was statistically signif-
cant. In effect, he made an all-or-nothing
bet on the premise that the polls under-
estimated Obamas strength in swing
states: Had this premise been wrong,
his 11-1 victory over RCP could easily
VI EWPOI NT
Join the Analytics Section of INFORMS
For more information, visit:
http://www.informs.org/Community/Analytics/Membership
J A NUAR Y / F E BR UAR Y 2013 | 19
APRIL 7, JUNE 23, JULY 13, & OCTOBER 5, 2013
Conducted BY INFORMS, the leading professional society in advanced analytics
INAUGURAL ANALYTICS CERTIFICATION EXAMS
Be among the first to become a Certified Analytics Professional (CAP).
Make plans now to take the profession's first analytics certification exam.
Candidate Handbook now available at
www.informs.org/Build-Your-Career/Analytics-Certification.
BENEFITS OF CERTIFICATION
Advances your career potential by setting you apart from the competition
Drives personal satisfaction of accomplishing a key career milestone
Helps improve your overall job performance by stressing continuing
professional development
Recognizes that you have invested in your analytics career by pursuing this rigorous credential
Boosts your salary potential by being viewed as experienced analytics professional
Shows competence in the principles and practices of analytics
ELIGIBILITY
BA/BS or MA/MS degree or higher
At least five years of analytics work-related experience for BA/BS holder in a related area
At least three years of analytics work-related experience for MA/MS (or higher) holder in
a related area
At least seven years of analytics work-related experience
for BA/BS (or higher) holder in an unrelated area
APPLICATIONS
Open in January, 2013
Prepare to apply by reviewing Candidate Handbook now
Arrange now to secure academic transcript and
confirmation of soft skills from employer
to send to INFORMS
COST
$495 INFORMS Members
$695 Non-Members
Bundled rates with meetings available
QUESTIONS
Email certification@mail.informs.org
2013 CAP EXAM SCHEDULE
APRIL 7, 2013
INFORMS Conference on
Business Analytics & O.R.,
San Antonio, TX
DOMAINS OF ANALYTICS PRACTICE
Domain Description Weight*
Business Problem (Question) Framing
Analytics Problem Framing
Data
Methodology (Approach) Selection
Model Building
Deployment
Life Cycle Management
*Percentage of questions in exam
I
II
III
IV
V
VI
VII
15%
17%
22%
15%
16%
9%
6%
100%
OCTOBER 5, 2013
INFORMS Annual
Meeting,
Minneapolis, MN
JUNE 23, 2013
INFORMS Healthcare
Conference,
Chicago, IL
July 13, 2013
Booz Allen Hamilton
Offices,
McLean, VA
J A NUAR Y / F E BR UAR Y 2013 | 21 WWW. I NF OR MS . OR G 20 | A NA LY T I CS - MAGA Z I NE . OR G
have been a 12-0 defeat. Yet uncertainties about how
to defne statistical signifcance cannot obscure the
fundamental point: 538 did extremely well in 2012 in
those states where accuracy was most important.
FINAL REMARKS
So how does it all add up? Under the Occams Razor
principle, there is a clear starting preference for simple
models over more complicated formulations. A complex
model must justify its intricacy by offering more accurate
information than a simpler counterpart; moreover, this
added information should arise in places where it is most
needed. In the present setting, the question is whether
Nate Silvers 538 model outperformed the straightfor-
ward RCP method to an extent that makes 538 the wiser
choice, even if the less transparent one.
Readers can reach their own judgments, but be-
cause of the results in the swing states, the author
believes that 538 met the test for superiority just
posed. While the tortoise catches up with the hare in
the nursery stories, it seems here that the hare won
hands down. But the outcome does not contradict Ae-
sops fable because, far from being lazy, the 538 hare
ran the race as hard as it could. And, if the evidence
is any guide, it is very much a world-class runner.
Arnold Barnett (abarnett@mit.edu) is the George Eastman
Professor of Management Science at the MIT Sloan School of
Management. His research specialty is applied mathematical
modeling with a focus on problems of health and safety. Barnett is a
senior member of INFORMS.
A complex model must
justify its intricacy by
offering more accurate
information than a simpler
counterpart; moreover, this
added information should
arise in places where it is
most needed.
VI EWPOI NT
Request a no-obligation INFORMS Member Benets Packet
For more information, visit: http://www.informs.org/Membership
J A NUAR Y / F E BR UAR Y 2013 | 21
A leading marketing companys scoring models used to take
4.5 hours to process. Now, with high-performance analytics
from SAS,
and R
GDXMRW (for MATLAB) and GDXRRW (for R) are utilities to exchange data
between GAMS and MATLAB or R and to call GAMS models from MATLAB or R:
Give MATLAB or R users access to all the optimization capabilities of GAMS
Allow visualization and analysis of GAMS data directly within MATLAB or R
Extend the existing GDX data utilities
GDXMRW is included with GAMS distributions
GDXRRW is freely downloadable
High-Level Modeling
The General Algebraic Modeling System (GAMS)
is a high-level modeling system for mathemati-
cal programming problems. GAMS is tailored for
complex, large-scale modeling applications, and
allows you to build large maintainable models that can
be adapted quickly to new situations. Models are fully
portable from one computer platform to another.
Wide Range of Model Types
GAMS allows the formulation of models in
many different problem classes, including
Linear (LP) and Mixed Integer Linear (MIP)
Quadratic Programming (QCP) and
Mixed Integer QCP (MIQCP)
Nonlinear (NLP) and Mixed Integer NLP (MINLP)
Constrained Nonlinear Systems (CNS)
Mixed Complementary (MCP)
Programs with Equilibrium Constraints (MPEC)
Conic Programming Problems
Stochastic Linear Problems
State-of-the-Art Solvers
GAMS incorporates all major commercial and
academic state-of-the-art solution techno-
logies for a broad range of problem types,
including global nonlinear optimization solvers.
For more information please visit: http://support.gams.com/doku.php?id=gdxrrw:interfacing_gams_and_r
MATLAB is a product of MathWorks, R is a free software environment for statistical computing and graphics.