THE

FIELD GUIDE
to DATA SCIENCE

SECOND
EDITION

© COPYRIGHT 2015 BOOZ ALLEN HAMILTON INC. ALL RIGHTS RESERVED.

›› F O R E WOR D
Data Science touches every aspect of our lives on a
daily basis. When we visit the doctor, drive our cars,
get on an airplane, or shop for services, Data Science
is changing the way we interact with and explore
our world.
Our world is now measured, As we move into this new
mapped, and recorded in digital future, it is clearer than ever, that
bits. Entire lives, from birth to businesses must adjust to these
death, are now catalogued in changes or risk being left behind.
the digital realm. These data, From influencing retail markets,
originating from such diverse to setting public health and safety
sources as connected vehicles, policies, or to addressing social
underwater microscopic cameras, unrest, organizations of all types
and photos we post to social are generating value through
media, have propelled us into Data Science. Data is our new
the greatest age of discovery currency and Data Science is
humanity has ever known. It is the mechanism by which we tap
through Data Science that we into it.
are unlocking the secrets hidden
within these data. We are making Data Science is an auspicious and
discoveries that will forever profound way of applying our
change how we live and interact curiosity and technical tradecraft
with the world around us. to solve humanity’s toughest
challenges. The growing power,
The impact of these changes importance, and responsibility
is having a profound effect on of applying Data Science
humanity. We have propelled methodologies to these challenges
ourselves into this age of is unimaginable. Our own
discovery through our incremental biases and assumptions can have
technological improvements. profound outcomes on business,
Data Science has become the national security, and our daily
catalyzing force behind our next lives. A new class of practitioners
evolutionary leap. Our own and leaders are needed to navigate
evolution is now inextricably this new future. Data Scientists
linked to that of computers. The are our guides on this journey as
way we live our lives and the skills they are creating radical new ways
that are important to our very of thinking about data and the
existence are directly dependent world around us.
upon the functions Data Science
can achieve on our behalf.

We want to share our passion for Data Science and start a

conversation with you. This is a journey worth taking.
Everyone you
will ever meet
knows something
you don’t.
[1]

›› T H E Sof TO RY
T
HE FIELD
GUIDE
Several years ago we created The Field Guide to Data Science because
we wanted to help organizations of all types and sizes. There were
countless industry and academic publications describing what Data
Science is and why we should care, but very little information was
available to explain how to make use of data as a resource. We find
that situation to be just as true today as we did two years ago, when
we created the first edition of the field guide.

At Booz Allen Hamilton, we built an industry-leading team of

Data Scientists. Over the course of hundreds of analytic challenges
for countless clients, we’ve unraveled the DNA of Data Science.
Many people have put forth their thoughts on single aspects of
Data Science. We believe we can offer a broad perspective on the
conceptual models, tradecraft, processes and culture of Data Science
– the what, the why, the who and the how. Companies with strong
Data Science teams often focus on a single class of problems – graph
algorithms for social network analysis, and recommender models for
online shopping are two notable examples. Booz Allen is different.
In our role as consultants, we support a diverse set of government
and commercial clients across a variety of domains. This allows us to
uniquely understand the DNA of Data Science.

Our goal in creating The Field Guide to Data Science was to

capture what we have learned and to share it broadly. The field
of Data Science has continued to advance since we first released
the field guide. As a result, we decided to release this second edition,
incorporating a few new and important concepts. We also added
technical depth and richness that we believe practitioners will
find useful.

We want this effort to continue driving forward the science and

art of Data Science.

This field guide came from the passion our team feels for its
work. It is not a textbook nor is it a superficial treatment.
Senior leaders will walk away with a deeper understanding of
the concepts at the heart of Data Science. Practitioners will
add to their toolbox. We hope everyone will enjoy the journey.
›› W E A R E A L L
AUTHORS of T HIS
STORY
We recognize that Data Science is a team sport. The Field Guide
to Data Science provides Booz Allen Hamilton’s perspective on the
complex and sometimes mysterious field of Data Science. We cannot
capture all that is Data Science. Nor can we keep up - the pace at
which this field progresses outdates work as fast as it is produced.
As a result, we opened this field guide to the world as a living
document to bend and grow with technology, expertise, and
evolving techniques.
Thank you to all the people that have emailed us your ideas as
well as the 100+ people who have watched, starred, or forked our
GitHub repository. We truly value the input of the community, as
we work together to advance the science and art of Data Science.
This is why we have included authors from outside Booz Allen
Hamilton on this second edition of The Field Guide to Data Science.
If you find the guide to be useful, neat, or even lacking, then
we encourage you to add your expertise, including:

›› Case studies from which you have learned

›› Citations from journal articles or papers that inspire you
›› Algorithms and techniques that you love
›› Your thoughts and comments on other people’s additions

Email us your ideas and perspectives at [email protected]

or submit them via a pull request on the GitHub repository.
Join our conversation and take the journey with us. Tell us and
the world what you know. Become an author of this story.
›› ACKNOWLEDGEMENTS
We would like to express our sincerest gratitude to
all those who have made The Field Guide to Data
Science such a success.
Thank you to the nearly 15,000 Thank you to the educators and
people who have downloaded academics who have incorporated
the digital copy from our website The Field Guide into your course
and the 100+ people who have work. We appreciate your trusting
connected with The Field Guide this guide as a way to introduce
on our GitHub page. We have your students to Data Science.
been overwhelmed by the It is an honor to know that we are
popularity of the work within the shaping the next generation of
Data Science community. Data Scientists.
Thank you to all of the Thank you to the organizational
practitioners who are using The leaders who have shared your
Field Guide as a resource. We are feedback, encouragement, and
excited to know that the work has success stories. We are thrilled
had such a strong influence, from to know that The Field Guide
shaping technical approaches to has helped so many organizations,
serving as the foundation for the from energy, to life sciences,
very definition and role of Data to retail, to begin their Data
Science within major government Science journeys.
and commercial organizations.

We hope you will all continue to find value from The Field
Guide to Data Science and to share in our excitement around the
release of this second edition. Please continue to be part of the
conversation and take this journey with us.
 ›› T H E O UTLINE
of OUR STORY
12 ›› Meet Your Guides

17 ›› The Short Version – The Core Concepts of Data Science

18 ›› Start Here for the Basics – An Introduction to Data Science

What Do We Mean by Data Science?
How Does Data Science Actually Work?
What Does It Take to Create a Data Science Capability?

46 ›› Take off the Training Wheels – The Practitioner’s Guide to Data Science
Guiding Principles
The Importance of Reason
Component Parts of Data Science
Fractal Analytic Model
The Analytic Selection Process
Guide to Analytic Selection
Detailed Table of Analytics

84 ›› Life in the Trenches – Navigating Neck Deep in Data

Going Deep into Machine Learning
Feature Engineering
Feature Selection
Ensemble Models
Data Veracity
Application of Domain Knowledge
The Curse of Dimensionality
Model Validation

102 ›› Putting it all Together – Our Case Studies

Streamlining Medication Review
Reducing Flight Delays
Making Vaccines Safer
Forecasting the Relative Risk for the Onset of
Mass Killings to Help Prevent Future Atrocities
Predicting Customer Response

114 ›› Closing Time

The Future of Data Science
Parting Thoughts
References
About Booz Allen Hamilton
›› M EE T your G U I D E S

Fred Blackburn Josh Sullivan Peter Guerra

(@boozallen) (@joshdsullivan) (@petrguerra)

Data Science is a field that is Leading our Data Science team Data Science is the most fascinating
evolving at a very rapid pace…be shows me every day the incredible blend of art and math and code
part of the journey. power of discovery and human and sweat and tears. It can take
curiosity. Don’t be afraid to blend you to the highest heights and the
art and science to advance your lowest depths in an instant, but it
own view of data analytics – it is the only way we will be able to
can be a powerful mixture. understand and describe the why.

Angela Zutavern Steve Escaravage Ezmeralda Khalil

(@angelazutavern) (@sescarav) (@ezmeraldakhalil)

Data Science is about asking bigger Invest your time and energy The power of data science
questions, seeing future possibilities, in data that is difficult to lies in the execution.
and creating outcomes you desire. assemble. If it doesn’t exist,
find a way to make it exist.

T H E F I ELD G U I D E to D A T A S C I E N C E
 Steven Mills Alex Cosmas Brian Keller
(@stevndmills) (@boozallen) (@boozallen)

Data Science truly can Data scientists should be truth- Grit will get you farther than talent.
change the world. seekers, not fact-seekers.

Stephanie Beben Kirk Borne Drew Farris

(@boozallen) (@KirkDBorne) (@drewfarris)

Begin every new data challenge Focus on value, not volume. Don’t forget to play. Play with
with deep curiosity along with tools, play with data, and play with
a healthy dose of skepticism. algorithms. You just might discover
something that will help you solve
that next nagging problem.

Meet Your Guides 13

 Paul Yacci Charles Glover Michael Kim
(@paulyacci) (@MindAfterMath) (@boozallen)

In the jungle of data, don't The beauty of data science lies Data science is both an art
miss the forest for the trees, in satisfying curiosities about and science.
or the trees for the forest. important problems by playing
with data and algorithms.

Stephanie Rivera Aaron Sander

(@boozallen) (@ajsander)

I treat Data Science like I do rock Data science is changing corporate

climbing: awesome dedication culture to be more like the open
leads to incremental improvement. source environment. More open,
Persistence leads to the top. more collaborative, and faster paced.

We would like to thank the following people for their

contributions and edits:
Tim Andrews, Mike Delurey, Greg Dupier, Jason Escaravage,
Christine Fantaskey, Juergen Klenk, Dan Liebermann, Mark
Rockley and Katie Wilks.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› CO M M U N I T Y CONTRIBUTORS

Will Cukierski, kaggle Mark Herman Ed Kohlwey

(@kaggle) (@cloudEBITDA) (@ekohlwey)

Two roads diverged in a wood, and I— End every analysis with… Data Science is about formally
I took the one in the direction ‘and therefore.’ analyzing everything around you
of the negative gradient, and becoming data driven.
And that has made all the difference.

Armen Kherlopian
(@akherlopian)

A Data Scientist must

continuously seek truth in spite
of ambiguity; therein rests the
basis of rigor and insight.

Meet Your Guides 15

›› The S H O RT
V E R S I O N
›› Data Science is the art of turning data into actions.
It’s all about the tradecraft. Tradecraft is the process, tools and
technologies for humans and computers to work together to
transform data into insights.

›› Data Science tradecraft creates data products.

Data products provide actionable information without exposing
decision makers to the underlying data or analytics (e.g., buy/sell
strategies for financial instruments, a set of actions to improve
product yield, or steps to improve product marketing).

›› Data Science supports and encourages shifting between

deductive (hypothesis-based) and inductive (pattern-
based) reasoning.
This is a fundamental change from traditional analysis approaches.
Inductive reasoning and exploratory data analysis provide a means
to form or refine hypotheses and discover new analytic paths.
Models of reality no longer need to be static. They are constantly
tested, updated and improved until better models are found.

›› Data Science is necessary for companies to stay with the

pack and compete in the future.
Organizations are constantly making decisions based on gut
instinct, loudest voice and best argument – sometimes they are
even informed by real information. The winners and the losers in
the emerging data economy are going to be determined by their
Data Science teams.

›› Data Science capabilities can be built over time.

Organizations mature through a series of stages – Collect,
Describe, Discover, Predict, Advise – as they move from data
deluge to full Data Science maturity. At each stage, they can
tackle increasingly complex analytic goals with a wider breadth
of analytic capabilities. However, organizations need not reach
maximum Data Science maturity to achieve success. Significant
gains can be found in every stage.

›› Data Science is a different kind of team sport.

Data Science teams need a broad view of the organization. Leaders
must be key advocates who meet with stakeholders to ferret out
the hardest challenges, locate the data, connect disparate parts of
the business, and gain widespread buy-in.

The Short Version 17

S TA RT H ERE for T H E B A S I C S
AN INTRODUCTION TO DATA SCIENCE
If you haven’t heard of Data Science, you’re behind the
times. Just renaming your Business Intelligence group
the Data Science group is not the solution.
 What do We Mean by
Data Science?

Describing Data Science is like trying to describe a sunset – it

should be easy, but somehow capturing the words is impossible.

T H E F I ELD G U I D E to D A T A S C I E N C E
Data Science Defined
Data Science is the art of turning data into actions. This is
accomplished through the creation of data products, which provide
actionable information without exposing decision makers to the
underlying data or analytics (e.g., buy/sell strategies for financial
instruments, a set of actions to improve product yield, or steps to
improve product marketing).

Performing Data Science requires the extraction of timely, actionable »»Data Product
information from diverse data sources to drive data products.
Examples of data products include answers to questions such as: A data product provides actionable
“Which of my products should I advertise more heavily to increase information without exposing
profit? How can I improve my compliance program, while reducing decision makers to the underlying
costs? What manufacturing process change will allow me to build a data or analytics. Examples include:
better product?” The key to answering these questions is: understand • Movie Recommendations
the data you have and what the data inductively tells you.
• Weather Forecasts
• Stock Market Predictions
• Production Process
Read this for additional background: Improvements
• Health Diagnosis
The term Data Science appeared infrastructure development
in the computer science literature work. We saw the need for a • Flu Trend Predictions
throughout the 1960s-1980s. new approach to distill value • Targeted Advertising
It was not until the late 1990s from our clients’ data. We
however, that the field as we approached the problem
describe it here, began to with a multidisciplinary
emerge from the statistics and team of computer scientists,
data mining communities mathematicians and domain
(e.g., [2] and [3]). Data Science experts. They immediately
was first introduced as an produced new insights and
independent discipline in 2001.[4] analysis paths, solidifying the
Since that time, there have been validity of the approach. Since
countless articles advancing the that time, our Data Science
discipline, culminating with team has grown to 250 staff
Data Scientist being declared the supporting dozens of clients
sexiest job of the 21st century.[5] across a variety of domains.
This breadth of experience
We established our first Data provides a unique perspective
Science team at Booz Allen on the conceptual models,
in 2010. It began as a natural tradecraft, processes and
extension of our Business culture of Data Science.
Intelligence and cloud

Start Here for the Basics 21

 What makes Data Science Different?
Data Science supports and encourages shifting between deductive
(hypothesis-based) and inductive (pattern-based) reasoning. This is
a fundamental change from traditional analytic approaches. Inductive
reasoning and exploratory data analysis provide a means to form or
refine hypotheses and discover new analytic paths. In fact, to do the
discovery of significant insights that are the hallmark of Data Science,
you must have the tradecraft and the interplay between inductive
and deductive reasoning. By actively combining the ability to reason
deductively and inductively, Data Science creates an environment
where models of reality no longer need to be static and empirically
based. Instead, they are constantly tested, updated and improved until
better models are found. These concepts are summarized in the figure,
The Types of Reason and Their Role in Data Science Tradecraft.

THE TYPES OF REASON…

DEDUCTIVE REASONING: INDUCTIVE REASONING

›› Commonly associated ›› Commonly known as “informal

with “formal logic.” logic,” or “everyday argument.”
›› Involves reasoning from known ›› Involves drawing uncertain
premises, or premises presumed inferences, based on
to be true, to a certain conclusion. probabilistic reasoning.
›› The conclusions reached are ›› The conclusions reached
certain, inevitable, inescapable. are probable, reasonable,
plausible, believable.

…AND THEIR ROLE IN DATA SCIENCE TRADECRAFT.

DEDUCTIVE REASONING: INDUCTIVE REASONING

›› Formulate hypotheses about ›› Exploratory data analysis to

relationships and underlying models. discover or refine hypotheses.
›› Carry out experiments with the data ›› Discover new relationships, insights
to test hypotheses and models. and analytic paths from the data.

Source: Booz Allen Hamilton

The Types of Reason and Their Role in Data Science Tradecraft

T H E F I ELD G U I D E to D A T A S C I E N C E
The differences between Data Science and traditional analytic
approaches do not end at seamless shifting between deductive
and inductive reasoning. Data Science offers a distinctly different
perspective than capabilities such as Business Intelligence. Data
Science should not replace Business Intelligence functions within
an organization, however. The two capabilities are additive and
complementary, each offering a necessary view of business operations
and the operating environment. The figure, Business Intelligence and
Data Science – A Comparison, highlights the differences between the
two capabilities. Key contrasts include:
›› Discovery vs. Pre-canned Questions: Data Science actually
works on discovering the question to ask as opposed to just
asking it.
›› Power of Many vs. Ability of One: An entire team provides
a common forum for pulling together computer science,
mathematics and domain expertise.
›› Prospective vs. Retrospective: Data Science is focused on
obtaining actionable information from data as opposed to
reporting historical facts.

LOOKING BACKWARD AND FORWARD

FIRST THERE WAS NOW WE'VE ADDED

BUSINESS INTELLIGENCE DATA SCIENCE

Deductive Reasoning Inductive and Deductive Reasoning

Backward Looking Forward Looking
Slice and Dice Data Interact with Data
Warehoused and Siloed Data Distributed, Real Time Data
Analyze the Past, Guess the Future Predict and Advise
Creates Reports Creates Data Products
Analytic Output Answer Questions and Create New Ones
Actionable Answer

Source: Booz Allen Hamilton

Business Intelligence and Data Science - A Comparison (adapted in part from [6])

Start Here for the Basics 23

 What is the Impact of Data Science?
As we move into the data economy, Data Science is the competitive
advantage for organizations interested in winning – in whatever way
winning is defined. The manner in which the advantage is defined
is through improved decision-making. A former colleague liked to
describe data-informed decision making like this: If you have perfect
information or zero information then your task is easy – it is in between
those two extremes that the trouble begins. What he was highlighting is
the stark reality that whether or not information is available, decisions
must be made.

The way organizations make decisions has been evolving for half a
century. Before the introduction of Business Intelligence, the only
options were gut instinct, loudest voice, and best argument. Sadly, this
method still exists today, and in some pockets it is the predominant
means by which the organization acts. Take our advice and never, ever
work for such a company!

Fortunately for our economy, most organizations began to inform

their decisions with real information through the application of
simple statistics. Those that did it well were rewarded; those that did
not failed. We are outgrowing the ability of simple stats to keep pace
with market demands, however. The rapid expansion of available data
and the tools to access and make use of the data at scale are enabling
fundamental changes to the way organizations make decisions.

Data Science is required to maintain competitiveness in the

increasingly data-rich environment. Much like the application of
simple statistics, organizations that embrace Data Science will be
rewarded while those that do not will be challenged to keep pace. As
more complex, disparate datasets become available, the chasm between
these groups will only continue to widen. The figure, The Business
Impacts of Data Science, highlights the value awaiting organizations
that embrace Data Science.

T H E F I ELD G U I D E to D A T A S C I E N C E
 DATA SCIENCE IS NECESSARY...

17-49% increase in productivity when organizations increase data

usability by 10%

11-42% return on assets (ROA) when organizations increase data

access by 10%

241% increase in ROI when organizations use big data to

improve competitiveness

increase in ROI when deploying analytics across most of

1000% the organization, aligning daily operations with senior
management's goals, and incorporating big data

5-6% performance improvement for organizations making

data-driven decisions.

...TO COMPETE IN THE FUTURE

Source: Booz Allen Hamilton

The Business Impacts of Data Science (adapted from [7], [8] and [9])

Start Here for the Basics 25

What is Different Now?
For 20 years IT systems were built the same way. We separated
the people who ran the business from the people who managed the
infrastructure (and therefore saw data as simply another thing they
had to manage). With the advent of new technologies and analytic
techniques, this artificial – and highly ineffective – separation of
critical skills is no longer necessary. For the first time, organizations
can directly connect business decision makers to the data. This simple
step transforms data from being ‘something to be managed’ into
‘something to be valued.’

In the wake of the transformation, organizations face a stark choice:

you can continue to build data silos and piece together disparate
information or you can consolidate your data and distill answers.

From the Data Science perspective, this is a false choice: The siloed
approach is untenable when you consider the (a) the opportunity
cost of not making maximum use of all available data to help
an organization succeed, and (b) the resource and time costs of
continuing down the same path with outdated processes. The tangible
benefits of data products include:

›› Opportunity Costs: Because Data Science is an emerging field,

opportunity costs arise when a competitor implements and
generates value from data before you. Failure to learn and account
for changing customer demands will inevitably drive customers
away from your current offerings. When competitors are able
to successfully leverage Data Science to gain insights, they can
drive differentiated customer value propositions and lead their
industries as a result.
›› Enhanced Processes: As a result of the increasingly interconnected
world, huge amounts of data are being generated and stored
every instant. Data Science can be used to transform data into
insights that help improve existing processes. Operating costs
can be driven down dramatically by effectively incorporating the
complex interrelationships in data like never before. This results
in better quality assurance, higher product yield and more
effective operations.

Start Here for the Basics 27

 How does Data Science
Actually Work?

It’s not rocket science… it’s something better - Data Science

Let’s not kid ourselves - Data Science is a complex field. It is difficult,

intellectually taxing work, which requires the sophisticated integration
of talent, tools and techniques. But as a field guide, we need to cut
through the complexity and provide a clear, yet effective way to
understand this new world.

To do this, we will transform the field of Data Science into a set of

simplified activities as shown in the figure, The Four Key Activities of a
Data Science Endeavor. Data Science purists will likely disagree with
this approach, but then again, they probably don’t need a field guide,
sitting as they do in their ivory towers! In the real world, we need
clear and simple operating models to help drive us forward.

T H E F I ELD G U I D E to D A T A S C I E N C E
 High

Degree
of
Effort
Setup Try Do

Evaluate

Evaluate

1 2 3 4
Low

Acquire Prepare Analyze Act

Data Science Activities

Activity 1: Acquire Activity 2: Prepare Activity 3: Analyze Activity 4: Act

This activity focuses Great outcomes This is the activity Every effective
on obtaining the don’t just happen that consumes the Data Science team
data you need. by themselves. lion’s share of the analyzes its data
Given the nature of A lot depends on team’s attention. with a purpose
data, the details of preparation, and It is also the most – that is, to turn
this activity depend in Data Science, challenging and data into actions.
heavily on who you that means exciting (you will Actionable and
are and what you manipulating the see a lot of ‘aha impactful insights
do. As a result, we data to fit your moments’ occur in are the holy grail
will not spend a analytic needs. this space). As the of Data Science.
lot of time on this This stage can most challenging Converting insights
activity other than consume a great and vexing of the into action can be a
to emphasize its deal of time, but four activities, politically charged
importance and it is an excellent this field guide activity, however.
to encourage an investment. The focuses on helping This activity
expansive view on benefits are you do this better depends heavily
which data can and immediate and and faster. on the culture and
should be used. long term. character of your
organization, so
we will leave you
to figure out those
details for yourself.

Source: Booz Allen Hamilton

The Four Key Activities of a Data Science Endeavor

Start Here for the Basics 29

 Acquire
All analysis starts with access to data, and for the Data Scientist
this axiom holds true. But there are some significant differences –
particularly with respect to the question of who stores, maintains and
owns the data in an organization.

But before we go there, lets look at what is changing. Traditionally,

rigid data silos artificially define the data to be acquired. Stated
another way, the silos create a filter that lets in a very small amount of
data and ignores the rest. These filtered processes give us an artificial
view of the world based on the ‘surviving data,’ rather than one that
shows full reality and meaning. Without a broad and expansive
dataset, we can never immerse ourselves in the diversity of the
data. We instead make decisions based on limited and constrained
information.

Eliminating the need for silos gives us access to all the data at once –
including data from multiple outside sources. It embraces the reality
that diversity is good and complexity is okay. This mindset creates a
completely different way of thinking about data in an organization by
giving it a new and differentiated role. Data represents a significant
new profit and mission-enhancement opportunity for organizations.

But as mentioned earlier, this first activity is heavily dependent upon

the situation and circumstances. We can’t leave you with anything
more than general guidance to help ensure maximum value:

›› Look inside first: What data do you have current access

to that you are not using? This is in large part the data
being left behind by the filtering process, and may be
incredibly valuable.
»»Not All Data Is Created Equal ›› Remove the format constraints: Stop limiting your data
acquisition mindset to the realm of structured databases.
As you begin to aggregate data, Instead, think about unstructured and semi-structured data
remember that not all data is as viable sources.
created equally. Organizations have
a tendency to collect any data that ›› Figure out what’s missing: Ask yourself what data would
is available. Data that is nearby make a big difference to your processes if you had access to it.
(readily accessible and easily Then go find it!
obtained) may be cheap to collect,
›› Embrace diversity: Try to engage and connect to publicly
but there is no guarantee it is the
available sources of data that may have relevance to your
right data to collect. Focus on the
domain area.
data with the highest ROI for your
organization. Your Data Science
team can help identify that data.
Also remember that you need to
strike a balance between the data
that you need and the data that you
have. Collecting huge volumes of
data is useless and costly if it is not
the data that you need.

T H E F I ELD G U I D E to D A T A S C I E N C E
 Prepare
Once you have the data, you need to prepare it for analysis.
Organizations often make decisions based on inexact data. Data
stovepipes mean that organizations may have blind spots. They are
not able to see the whole picture and fail to look at their data and
challenges holistically. The end result is that valuable information is
withheld from decision makers. Research has shown almost 33% of
decisions are made without good data or information. [10]

When Data Scientists are able to explore and analyze all the data, new
opportunities arise for analysis and data-driven decision making. The
insights gained from these new opportunities will significantly change
the course of action and decisions within an organization. Gaining
access to an organization’s complete repository of data, however,
requires preparation.

Our experience shows time and time again that the best tool for
Data Scientists to prepare for analysis is a lake – specifically, the Data
Lake.[11] This is a new approach to collecting, storing and integrating
data that helps organizations maximize the utility of their data.
Instead of storing information in discrete data structures, the Data
Lake consolidates an organization’s complete repository of data in
a single, large view. It eliminates the expensive and cumbersome
data-preparation process, known as Extract/Transform/Load (ETL),
necessary with data silos. The entire body of information in the Data
Lake is available for every inquiry – and all at once.

Start Here for the Basics 31

 Analyze
We have acquired the data… we have prepared it… now it is time to
analyze it.

The Analyze activity requires the greatest effort of all the activities
in a Data Science endeavor. The Data Scientist actually builds the
analytics that create value from data. Analytics in this context is
an iterative application of specialized and scalable computational
resources and tools to provide relevant insights from exponentially
growing data. This type of analysis enables real-time understanding
of risks and opportunities by evaluating situational, operational and
behavioral data.

With the totality of data fully accessible in the Data Lake,

organizations can use analytics to find the kinds of connections and
patterns that point to promising opportunities. This high-speed
analytic connection is done within the Data Lake, as opposed to
older style sampling methods that could only make use of a narrow
slice of the data. In order to understand what was in the lake, you had
to bring the data out and study it. Now you can dive into the lake,
bringing your analytics to the data. The figure, Analytic Connection in
the Data Lake, highlights the concept of diving into the Data Lake to
discover new connections and patterns.

Source: Booz Allen Hamilton

Analytic Connection in the Data Lake

T H E F I ELD G U I D E to D A T A S C I E N C E
Data Scientists work across the spectrum of analytic goals – Describe,
Discover, Predict and Advise. The maturity of an analytic capability
determines the analytic goals encompassed. Many variables play key
roles in determining the difficulty and suitability of each goal for an
organization. Some of these variables are the size and budget of an
organization and the type of data products needed by the decision
makers. A detailed discussion on analytic maturity can be found in
Data Science Maturity within an Organization.

In addition to consuming the greatest effort, the Analyze activity

is by far the most complex. The tradecraft of Data Science is an
art. While we cannot teach you how to be an artist, we can share
foundational tools and techniques that can help you be successful.
The entirety of Take Off the Training Wheels is dedicated to sharing
insights we have learned over time while serving countless clients.
This includes descriptions of a Data Science product lifecycle and
the Fractal Analytic Model (FAM). The Analytic Selection Process and
accompanying Guide to Analytic Selection provide key insights into one
of the most challenging tasks in all of Data Science – selecting the
right technique for the job.

Act
Now that we have analyzed the data, it’s time to take action.

The ability to make use of the analysis is critical. It is also very

situational. Like the Acquire activity, the best we can hope for is to
provide some guiding principles to help you frame the output for
maximum impact. Here are some key points to keep in mind when
presenting your results:

1. The finding must make sense with relatively little up-front

training or preparation on the part of the decision maker.
2. The finding must make the most meaningful patterns, trends
and exceptions easy to see and interpret.
3. Every effort must be made to encode quantitative data
accurately so the decision maker can accurately interpret and
compare the data.
4. The logic used to arrive at the finding must be clear and
compelling as well as traceable back through the data.
5. The findings must answer real business questions.

Start Here for the Basics 33

Data Science Maturity within
an Organization
The four activities discussed thus far provide a simplified view of Data
Science. Organizations will repeat these activities with each new Data
Science endeavor. Over time, however, the level of effort necessary
for each activity will change. As more data is Acquired and Prepared
in the Data Lake, for example, significantly less effort will need to
be expended on these activities. This is indicative of a maturing Data
Science capability.

Assessing the maturity of your Data Science capability calls for a

slightly different view. We use The Data Science Maturity Model as
a common framework for describing the maturity progression and
components that make up a Data Science capability. This framework
can be applied to an organization’s Data Science capability or even
to the maturity of a specific solution, namely a data product. At each
stage of maturity, powerful insight can be gained.

Maturity

Advise
Data Silos

Predict

Proportion
Discover
of
Effort

Describe

Collect

Stages of Maturity

Source: Booz Allen Hamilton

The Data Science Maturity Model

Start Here for the Basics 35

 When organizations start out, they have Data Silos. At this stage,
they have not carried out any broad Aggregate activities. They may
not have a sense of all the data they have or the data they need. The
decision to create a Data Science capability signals the transition into
the Collect stage.

All of your initial effort will be focused on identifying and aggregating

data. Over time, you will have the data you need and a smaller
proportion of your effort can focus on Collect. You can now begin to
Describe your data. Note, however, that while the proportion of time
spent on Collect goes down dramatically, it never goes away entirely.
This is indicative of the four activities outlined earlier – you will
continue to Aggregate and Prepare data as new analytic questions
arise, additional data is needed and new data sources become available.

Organizations continue to advance in maturity as they move through

the stages from Describe to Advise. At each stage they can tackle
increasingly complex analytic goals with a wider breadth of analytic
capabilities. As described for Collect, each stage never goes away
entirely. Instead, the proportion of time spent focused on it goes
down and new, more mature activities begin. A brief description
of each stage of maturity is shown in the table The Stages of Data
Science Maturity.

The Stages of Data Science Maturity

Stage Description Example

Focuses on collecting internal Gathering sales records and

Collect or external datasets. corresponding weather data.

Seeks to enhance or How are my customers

refine raw data as well distributed with respect to
Describe as leverage basic analytic location, namely zip code?
functions such as counts.

Are there groups within

Identifies hidden relationships
Discover or patterns.
my regular customers that
purchase similarly?

Utilizes past observations to Can we predict which products

Predict predict future observations. that certain customer groups
are more likely to purchase?

Defines your possible decisions,

Your advice is to target advertise
optimizes over those decisions,
Advise and advises to use the decision
to specific groups for certain
products to maximize revenue.
that gives the best outcome.

Source: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
The maturity model provides a powerful tool for understanding
and appreciating the maturity of a Data Science capability.
Organizations need not reach maximum maturity to achieve
success. Significant gains can be found in every stage. We believe
strongly that one does not engage in a Data Science effort, however, »»W here does your organization
unless it is intended to produce an output – that is, you have the fall in analytic maturity?
intent to Advise. This means simply that each step forward in
maturity drives you to the right in the model diagram. Moving Take the quiz!
to the right requires the correct processes, people, culture and 1. H
 ow many data sources do
operating model – a robust Data Science capability. What Does it you collect?
Take to Create a Data Science Capability? addresses this topic. a. Why do we need a bunch of data?
– 0 points, end here.
We have observed very few organizations actually operating at b. I don’t know the exact number.
the highest levels of maturity, the Predict and Advise stages. The – 5 points
tradecraft of Discover is only now maturing to the point that c. We identified the required data and
organizations can focus on advanced Predict and Advise activities. collect it. – 10 points
This is the new frontier of Data Science. This is the space in which
we will begin to understand how to close the cognitive gap between 2. Do you know what questions
your Data Science team is trying
humans and computers. Organizations that reach Advise will be to answer?
met with true insights and real competitive advantage.
a. Why do we need questions?
- 0 points
b. No, they figure it out for themselves.
– 5 points
c. Yes, we evaluated the questions that
will have the largest impact to the
business. – 10 points

3. Do you know the important factors

driving your business?
a. I have no idea. – 0 points
b. Our quants help me figure it out.
– 5 points
c. We have a data product for that.
– 10 points

4. Do you have an understanding of

future conditions?
a.I look at the current conditions and
read the tea leaves. – 0 points
b. We have a data product for that.
– 5 points

5. Do you know the best course

of action to take for your key
decisions?
a. I look at the projections and plan a
course. – 0 points
b. We have a data product for that.
– 5 points

Check your score:

0 – Data Silos, 5-10 – Collect,
10-20 – Describe, 20-30 – Discover,
30-35 – Predict, 35-40 - Advise

Source: Booz Allen Hamilton

Start Here for the Basics 37

 What Does it Take to Create
a Data Science Capability?

Data Science is all about building teams and culture.

Many organizations (both commercial and government) see the

potential in capitalizing on data to unlock operational efficiencies,
to create new services and experiences, and to propel innovation.
Unfortunately, too many business leaders invest in one-off technical
solutions— with a big price tag and mixed results— instead of
investing in building a strategic Data Science capability. A Data
Science capability embeds and operationalizes Data Science across
an enterprise such that it can deliver the next level of organizational
performance and return on investment. A Data Science capability
moves an organization beyond performing pockets of analytics to an
enterprise approach that uses analytical insights as part of the normal
course of business. When building a capability, it is important for an
organization to first identify its analytic goals (i.e., what it is trying
to achieve through analytics) and then assess its readiness to achieve
those goals – examining both technical readiness and organizational
readiness. An organization can then make strategic choices on how to
address gaps and begin to build their capability.

T H E F I ELD G U I D E to D A T A S C I E N C E
Building Your Data Science Team
A critical component to any Data Science capability is having the
right team. Data Science depends on a diverse set of skills as shown
in The Data Science Venn Diagram. Computers provide the
environment in which data-driven hypotheses are tested, and as such,
computer science is necessary for data manipulation and processing.
Mathematics provides the theoretical structure in which Data Science
problems are examined. A rich background in statistics, geometry,
linear algebra, and calculus are all important to understand the basis
for many algorithms and tools. Finally, domain expertise contributes
to an understanding of what problems actually need to be solved,
what kind of data exists in the domain, and how the problem space
may be instrumented and measured.

DOMAIN EXPERTISE
Provides understanding
of the reality in which a
problem space exists.

COMPUTER SCIENCE MATHEMATICS

Provides the environment Provides the theoretical
in which data products structure in which Data
are created. Science problems
are examined.

Source: Booz Allen Hamilton

The Data Science Venn Diagram (inspired by [12]

)

Remember that Data Science is a team sport. Most of the time, you
will not be able to find the rare “unicorns” - people with expertise
across all three of the skill areas. Therefore, it is important to build a
blended team that covers all three elements of the Data Science
Venn Diagram.

DOMAIN EXPERTISE
Provides understanding
of the reality in which a
problem space exists.
Start Here for the Basics 39
 BALANCING THE DATA
SCIENCE TEAM EQUATION

Balancing the composition of a Data Science team

is much like balancing the reactants and products in
a chemical reaction. Each side of the equation must
represent the same quantity of any particular element.
In the case of Data Science, these elements are the
foundational technical skills Computer Science (CS),
Mathematics (M) and Domain Expertise (DE). The
reactants, your Data Scientists, each have their own
unique skills compositions. You must balance the staff
mix to meet the skill requirements of the Data Science
team, the product in the reaction. If you don’t correctly
balance the equation, your Data Science team will not
have the desired impact on the organization.

2 CS M2 + 2 CS + M DE → CS4 M5 DE

In the example above, your project requires four parts

computer science, five parts mathematics and one part
domain expertise. Given the skills mix of the staff, five
people are needed to balance the equation. Throughout
your Data Science project, the skills requirements of
the team will change. You will need to re-balance the
equation to ensure the reactants balance with
the products.

Source: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
Understanding What Makes
a Data Scientist
Data Science often requires a significant investment of time across
a variety of tasks. Hypotheses must be generated and data must be
acquired, prepared, analyzed, and acted upon. Multiple techniques
are often applied before one yields interesting results. If that seems
daunting, it is because it is. Data Science is difficult, intellectually
taxing work, which requires lots of talent: both tangible technical
»»The Triple Threat Unicorn
skills as well as the intangible “x-factors.”

There are four independent yet comprehensive foundational Data Individuals who are great at
all three of the Data Science
Science competency clusters that, when considered together, convey
foundational technical skills are
the essence of what it means to be a successful Data Scientist. There
like unicorns – very rare and if
are also reach back competencies that complement the foundational you’re ever lucky enough to find
clusters but do not define the core tradecraft or attributes of the Data one they should be treated carefully.
Science team. When you manage these people:

›› Encourage them to lead your

Data Science Competency Framework team, but not manage it.
(see [13] for complete framework) Don’t bog them down with
responsibilities of management
Clusters Competencies Description that could be done by other staff.

Technical: Advanced Mathematics; The technical competency ›› Put extra effort into managing
“Knows How and Computer Science; Data cluster depicts the
What to do” Mining and Integration; foundational technical their careers and interests
Database Science; Research and specialty knowledge within your organization. Build
Design; Statistical Modeling; and skills needed for
Machine Learning; successful performance opportunities for promotion into
Operations Research; in each job or role. your organization that allow
Programming and Scripting
them to focus on mentoring other
Data Scientists and progressing
Data Science Collaboration and Teamwork; The characteristics in the
Communications; Data consulting competency the state of the art while also
Consulting:
“Can Do in Science Consulting; Ethics cluster can help Data advancing their careers.
and Integrity Scientists easily integrate
a Client and
into various market or
Customer domain contexts and partner ›› Make sure that they have the
Environment” with business units to opportunity to present and
understand the environment
and solve complex problems. spread their ideas in many
different forums, but also be
sensitive to their time.
Cognitive: Critical Thinking; Inductive The cognitive competency
“Able to Do or and Deductive Reasoning; cluster represents the type
Learn to Do” Problem Solving of critical thinking and
reasoning abilities (both
inductive and deductive) a
Data Scientist should have to
perform their job.

Personality: Adaptability/Flexibility; The personality competency

“Willing or Ambiguity Tolerance; Detail cluster describes the
Motivated to Do” Orientation; Innovation and personality traits that drive
Creativity; Inquisitiveness; behaviors that are beneficial
Perseverance; Resilience to Data Scientists, such as
and Hardiness; Self- inquisitiveness, creativity,
Confidence; Work Ethic and perseverance.

Reach Back Competencies for Data Science Teams

Business Acumen; Data Visualization; Domain Expertise; Program Management

Source: Booz Allen Hamilton

Start Here for the Basics 41

 The most important qualities of Data Scientists tend to be the
intangible aspects of their personalities. Data Scientists are by nature
curious, creative, focused, and detail-oriented.

›› Curiosity is necessary to peel apart a problem and examine the

interrelationships between data that may appear superficially
unrelated.
›› Creativity is required to invent and try new approaches to
solving a problem, which often times have never been applied in
»»Don’t judge a book by its such a context before.
cover, or a Data Scientist by ›› Focus is required to design and test a technique over days and
his or her degree in this case. weeks, find it doesn’t work, learn from the failure, and try again.
Amazing Data Scientists can
›› Attention to Detail is needed to maintain rigor, and to detect and
be found anywhere. Just look
avoid over-reliance on intuition when examining data.
at the diverse and surprising
sampling of degrees held by We have found the single most important attribute is flexibility in
Our Experts: overcoming setbacks - the willingness to abandon one idea and try
a new approach. Often, Data Science is a series of dead ends before,
›› Bioinformatics at last, the way forward is identified. It requires a unique set of
›› Biomedical Engineering personality attributes to succeed in such an environment. Technical
›› Biophysics skills can be developed over time: the ability to be flexible – and
patient, and persistent – cannot.
›› Business
›› Computer Graphics
›› Computer Science
›› English Finding the Athletes for Your Team
›› Forest Management
›› History Building a Data Science team is complex. Organizations must
simultaneously engage existing internal staff to create an “anchor” who
›› Industrial Engineering
can be used to recruit and grow the team, while at the same time
›› Information Technology undergo organizational change and transformation to meaningfully
›› Mathematics incorporate this new class of employee. Building a team starts with
›› National Security Studies identifying existing staff within an organization who have a high
aptitude for Data Science. Good candidates will have a formal
›› Operations Research
background in any of the three foundational technical skills we
›› Physics mentioned, and will most importantly have the personality traits
›› Wildlife & Fisheries necessary for Data Science. They may often have advanced (masters or
Management higher) degrees, but not always. The very first staff you identify should
also have good leadership traits and a sense of purpose for the
organization, as they will lead subsequent staffing and recruiting
efforts. Don’t discount anyone – you will find Data Scientists in the
strangest places with the oddest combinations of backgrounds.

T H E F I ELD G U I D E to D A T A S C I E N C E
Shaping the Culture
It is no surprise—building a culture is hard and there is just as
much art to it as there is science. It is about deliberately creating the
conditions for Data Science to flourish (for both Data Scientists and
the average employee). You can then step back to empower collective
ownership of an organic transformation.

Data Scientists are fundamentally curious and imaginative. We have

a saying on our team, “We’re not nosy, we’re Data Scientists.” These »»“I'm not nosey, I'm a Data
qualities are fundamental to the success of the project and to gaining Scientist”
new dimensions on challenges and questions. Often Data Science
projects are hampered by the lack of the ability to imagine something ›› Always remember that
new and different. Fundamentally, organizations must foster trust and unrelenting curiosity and
transparent communication across all levels, instead of deference to imagination should be the
authority, in order to establish a strong Data Science team. Managers hallmarks of Data Science. They
should be prepared to invite participation more frequently, and offer are fundamental to the success
of every Data Science project.
explanation or apology less frequently.

It is important to provide a path into the Data Science “club” and

to empower the average employee to feel comfortable and conversant
with Data Science. For something to be part of organizational
culture, it must be part of the fabric of the employee behavior.
That means employees must interact with and use data products
in their daily routines. Another key ingredient to shaping the
right culture is that all employees need a baseline of Data Science
knowledge, starting with a common lexicon, to facilitate productive
collaboration and instill confidence. While not everyone will be
Data Scientists, employees need to identify with Data Science and
be equipped with the knowledge, skills, and abilities to work with
Data Scientists to drive smarter decisions and deliver exponential
organizational performance.

Start Here for the Basics 43

 Selecting Your Operating Model
Depending on the size, complexity, and the business drivers,
organizations should consider one of three Data Science operating
models: Centralized, Deployed, or Diffused. These three models are
shown in the figure, Data Science Operating Models.

CENTRALIZED DEPLOYED DIFFUSED

Business units bring their Small Data Science teams Data Scientists are fully
problems to a centralized are forward deployed to embedded within the
Data Science team. business units. business units.

Source: Booz Allen Hamilton

Data Science Operating Models
(see [13] for complete descriptions)

Centralized Data Science teams serve the organization across all business
units. The team is centralized under a Chief Data Scientist and they all
co-locate together. The domain experts come to this organization for
brief rotational stints to solve challenges around the business. This model
provides greater efficiency with limited Data Science resources but can also
create the perceived need to compete with other business units for Data
Science talent. To address this challenge, it is important to place emphasis
on portfolio management and creating transparency on how organizations
will identify and select Data Science projects.

Deployed Data Science teams go to the business unit and reside there for
short- or long-term assignments. They are their own entity and they work
with the domain experts within the group to solve hard problems. In
the deployed model, Data Science teams collectively develop knowledge
across business units, with central leadership as a bridging mechanism for
addressing organization-wide issues. However, Data Science teams are
accountable to business unit leadership and their centralized leadership,
which could cause confusion and conflict. In this model, it is important
to emphasize conflict management to avoid competing priorities.

The Diffused Data Science team is one that is fully embedded with each
group and becomes part of the long-term organization. These teams work
best when the nature of the domain or business unit is already one focused
on analytics. In the Diffused Model, teams can quickly react to high-
priority business unit needs. However, the lack of central management can
result in duplicate software and tools. Additionally, business units with the
most money will often have full access to analytics while other units have
none—this may not translate to the greatest organizational impact. In this
model, it is important to establish cross-functional groups that promote
organization-wide governance and peer collaboration.

Full descriptions of each operating model can be found in Booz Allen’s Tips for
Building a Data Science Capability [13].

T H E F I ELD G U I D E to D A T A S C I E N C E
How to Generate Momentum
A Data Science effort can start at the grass roots level by a few folks
tackling hard problems, or as directed by the Chief Executive Officer,
Chief Data Officer, or Chief Analytics Officer. Regardless of how an
effort starts, political headwinds often present more of a challenge
than solving any technical hurdles. To help battle the headwinds, it is
important to generate momentum and prove the value a Data Science
team can provide. The best way to achieve this is usually through
a Data Science prototype or proof of concept. Proofs of concepts
can generate the critical momentum needed to jump start any Data
Science Capability Four qualities, in particular, are essential for every
Data Science prototype:

1. Organizational Buy-in: A prototype will only succeed if the

individuals involved believe in it and are willing to do what
they can to make it successful. A good way to gauge interest
is to meet with the middle managers; their views are usually
indicative of the larger group.

2. Clear ROI: Before choosing a prototype problem, ensure that

the ROI of the analytic output can be clearly and convincingly
demonstrated for both the project and the organization as a
whole. This outcome typically requires first reaching consensus
on how the ROI will be determined and measured, so that the
benefit can be quantified.

3. Necessary Data: Before selecting a prototype, you must first

determine exactly what data is needed, whether it will actually
be available, and what it will cost in terms of time and expense.
It is important to note that organizations do not need all the
possible data – they can still create successful analytics even
with some gaps.

4. Limited Complexity and Duration: The problem addressed

by the prototype should achieve a balance between being too
complex and too easy. Organizations new to Data Science often
try to show its value with highly complex projects. However,
the greater the complexity, the greater the risk of failure. At the
same time, if the problem is too easy to solve, senior leaders
and others in the organization may not see the need for Data
Science. Look for efforts that could benefit from large datasets,
or bringing together disparate datasets that have never been
combined before, as opposed to those that require complex
analytic approaches. In these cases, there is often low-hanging
fruit that can lead to significant value for the organization.

Start Here for the Basics 45

TA K E O F F the T R A I N I N G W H E E L S
THE PRACTITIONER’S GUIDE
TO DATA SCIENCE
Read this section to get beyond the hype and
learn the secrets of being a Data Scientist.
 Guiding Principles

Failing is good; failing quickly is even better.

The set of guiding principles that govern how we conduct the

tradecraft of Data Science are based loosely on the central tenets
of innovation, as the two areas are highly connected. These principles
are not hard and fast rules to strictly follow, but rather key tenets
that have emerged in our collective consciousness. You should use
these to guide your decisions, from problem decomposition
through implementation.

›› Be willing to fail. At the core of Data Science is the idea of

experimentation. Truly innovative solutions only emerge when
»»Tips From the Pros you experiment with new ideas and applications. Failure is an
acceptable byproduct of experimentation. Failures locate regions
It can be easier to rule out a solution that no longer need to be considered as you search for a solution.
than confirm its correctness. As a ›› Fail often and learn quickly. In addition to a willingness to fail, be
result, focus on exploring obvious ready to fail repeatedly. There are times when a dozen approaches
shortcomings that can quickly
must be explored in order to find the one that works. While you
disqualify an approach. This will allow
shouldn’t be concerned with failing, you should strive to learn from
you to focus your time on exploring
the attempt quickly. The only way you can explore a large number
truly viable approaches as opposed to
dead ends.
of solutions is to do so quickly.
›› Keep the goal in mind. You can often get lost in the details and
challenges of an implementation. When this happens, you lose
sight of your goal and begin to drift off the path from data to
analytic action. Periodically step back, contemplate your goal, and
evaluate whether your current approach can really lead you where
you want to go.

›› Dedication and focus lead to success. You must often explore

many approaches before finding the one that works. It’s easy to
become discouraged. You must remain dedicated to your analytic
goal. Focus on the details and the insights revealed by the data.
Sometimes seemingly small observations lead to big successes.
»»Tips From the Pros

If the first thing you try to do is to ›› Complicated does not equal better. As technical practitioners, we
create the ultimate solution, you will have a tendency to explore highly complex, advanced approaches.
fail, but only after banging your head While there are times where this is necessary, a simpler approach
against a wall for several weeks. can often provide the same insight. Simpler means easier and
faster to prototype, implement and verify.

T H E F I ELD G U I D E to D A T A S C I E N C E
The Importance of Reason

Beware: in the world of Data Science, if it walks like a duck

and quacks like a duck, it might just be a moose.

Data Science supports and encourages shifting between deductive

(hypothesis-based) and inductive (pattern-based) reasoning.
Inductive reasoning and exploratory data analysis provide a means
to form or refine hypotheses and discover new analytic paths.
Models of reality no longer need to be static. They are constantly
tested, updated and improved until better models are found.

The analysis of big data has brought inductive reasoning to the

forefront. Massive amounts of data are analyzed to identify
correlations. However, a common pitfall to this approach is confusing »»Correlation without
correlation with causation. Correlation implies but does not prove Causation
causation. Conclusions cannot be drawn from correlations until the
underlying mechanisms that relate the data elements are understood. A common example of this
Without a suitable model relating the data, a correlation may simply phenomenon is the high correlation
be a coincidence. between ice cream consumption and
the murder rate during the summer
months. Does this mean ice cream
consumption causes murder or,
conversely, murder causes ice cream
consumption? Most likely not, but
you can see the danger in mistaking
correlation for causation. Our job as
Data Scientists is making sure we
understand the difference.

Take off the Training Wheels 49

 ›› The Dangers of Rejection In the era of big gene and formulate a hypothesis
data, one piece to test whether there is a
of analysis that significant difference between
is frequently the means. Given that we were
overlooked is running thousands of these tests
the problem of at α = 0.05, we found several
finding patterns differences that were significant.
Paul Yacci when there The problem was that some
are actually no of these could be caused by
apparent patterns. In statistics random chance.
this is referred to as Type I error.
As scientists, we are always Many corrections exist to
on the lookout for a new or control for false indications of
interesting breakthrough that significance. The Bonferroni
could explain a phenomenon. correction is one of the most
We hope to see a pattern in our conservative. This calculation
data that explains something lowers the level below which you
or that can give us an answer. will reject the null hypothesis
The primary goal of hypothesis (your p value). The formula is
testing is to limit Type I error. alpha/n, where n equals the
This is accomplished by using number of hypothesis tests
small α values. For example, that you are running. Thus, if
a α value of 0.05 states that you were to run 1,000 tests of
there is a 1 in 20 chance that significance at α = 0.05, your
the test will show that there p value should be less than
is something significant when 0.00005 (0.05/1,000) to reject the
in actuality there isn’t. This null hypothesis. This is obviously
problem compounds when a much more stringent value.
testing multiple hypotheses. A large number of the previously
When running multiple significant values were no longer
hypothesis tests, we are likely significant, revealing the true
to encounter Type I error. As relationships within the data.
more data becomes available
The corrected significance gave
for analysis, Type I error
us confidence that the observed
needs to be controlled.
expression levels were due to
One of my projects required differences in the cellular gene
testing the difference between expression rather than noise. We
the means of two microarray were able to use this information
data samples. Microarray to begin investigating what
data contains thousands of proteins and pathways were
measurements but is limited active in the genes expressing
in the number of observations. the phenotype of interest. By
A common analysis approach solidifying our understanding
is to measure the same genes of the causal relationships, we
under different conditions. If focused our research on the
there is a significant enough areas that could lead to new
difference in the amount of discoveries about gene function
gene expression between the and, ultimately to improved
two samples, we can say that medical treatments.
the gene is correlated with a
particular phenotype. One way
to do this is to take the mean of
each phenotype for a particular

T H E F I ELD G U I D E to D A T A S C I E N C E
Reason and common sense are foundational to Data Science. Without these, data is
simply a collection of bits. Context, inferences and models are created by humans and
carry with them biases and assumptions. Blindly trusting your analyses is a dangerous
thing that can lead to erroneous conclusions. When you approach an analytic
challenge, you should always pause to ask yourself the following questions:

›› What problem are we trying ›› Does the analysis address the

to solve? Articulate the answer original intent? Make sure
as a sentence, especially when that you are not aligning the
communicating with the end- answer with the expectations
user. Make sure that it sounds of the client. Always speak
»»Tips From the Pros
like an answer. For example, the truth, but remember that
“Given a fixed amount of answers of “your baby is ugly”
Better a short pencil than a
human capital, deploying require more, not less, analysis.
long memory. End every day by
people with these priorities documenting where you are; you
will generate the best return ›› Is the story complete? The goal
may learn something along the way.
on their time.” of your analysis is to tell an
Document what you learned and why
actionable story. You cannot
you changed your plan.
›› Does the approach make sense? rely on the audience to stitch
Write out your analytic plan. the pieces together. Identify
Embrace the discipline of potential holes in your
writing, as it brings structure story and fill them to avoid
to your thinking. Back of surprises. Grammar, spelling
the envelope calculations are and graphics matter; your
an existence proof of your audience will lose confidence
approach. Without this kind in your analysis if your results
of preparation, computers are look sloppy.
power tools that can produce
lots of bad answers really fast. ›› Where would we head next?
No analysis is ever finished, »»Tips From the Pros
›› Does the answer make sense? you just run out of resources.
Can you explain the answer? Understand and explain what Test your answers with a friendly
Computers, unlike children, additional measures could audience to make sure your findings
do what they are told. Make be taken if more resources hold water.
sure you spoke to it clearly by are found.
validating that the instructions
you provided are the ones you
intended. Document your
assumptions and make sure
they have not introduced bias
in your work.
›› Is it a finding or a mistake?
Be skeptical of surprise
findings. Experience says that
it if seems wrong, it probably
is wrong. Before you accept
that conclusion, however,
make sure you understand
and can clearly explain why
it is wrong.

Take off the Training Wheels 51

 Component Parts of
Data Science

There is a web of components that interact to create your

solution space. Understanding how they are connected
is critical to your ability to engineer solutions to Data
Science problems.

The components involved in any Data Science project fall into a

number of different categories including the data types analyzed, the
analytic classes used, the learning models employed and the execution
models used to run the analytics. The interconnection across these
components, shown in the figure, Interconnection Among the Component
Parts of Data Science, speaks to the complexity of engineering Data
Science solutions. A choice made for one component exerts influence
over choices made for others categories. For example, data types
lead the choices in analytic class and learning models, while latency,
timeliness and algorithmic parallelization strategy inform the
execution model. As we dive deeper into the technical aspects of
Data Science, we will begin with an exploration of these components
and touch on examples of each.

Read this to get the quick and dirty:

When engineering a Data how the learning models embodied

Science solution, work from an will operate and evolve, and the
understanding of the components execution models that will govern
that define the solution space. how the analytic will be run.
Regardless of your analytic goal, You will be able to articulate a
you must consider the data types complete Data Science solution
with which you will be working, only after considering each of
the classes of analytics you will use these aspects.
to generate your data product,

T H E F I ELD G U I D E to D A T A S C I E N C E
 execution data types

data ing
models

execu

stream
serial
tion
ex
pa utio
ec

da tch
ra n
lle

ba
ta
l
str
e
exe ami tured
cut ng uc
ion str ta
da

batch d
execution unstructure
data

e tran
offlin g ana sformin
n i n lytic g
lear s

le
in
e an arn
nl al ing
g

o yt
in

ic
rn

s
a
le

pre lytics
lear ised

ana
ning

dic
supervised
learning
rv

tive
upe
uns

learning analytic
models classes

Source: Booz Allen Hamilton

Interconnection Among the Component Parts of Data Science

Take off the Training Wheels 53

Data Types
Data types and analytic goals go hand-in-hand much like the chicken
and the egg; it is not always clear which comes first. Analytic goals are
derived from business objectives, but the data type also influences the
goals. For example, the business objective of understanding consumer
product perception drives the analytic goal of sentiment analysis.
Similarly, the goal of sentiment analysis drives the selection of a
text-like data type such as social media content. Data type also
drives many other choices when engineering your solutions.

There are a number of ways to classify data. It is common to

characterize data as structured or unstructured. Structured data exists
when information is clearly broken out into fields that have an
explicit meaning and are highly categorical, ordinal or numeric.
A related category, semi-structured, is sometimes used to describe
structured data that does not conform to the formal structure of
data models associated with relational databases or other forms
of data tables, but nonetheless contains tags or other markers.
Unstructured data, such as natural language text, has less clearly
delineated meaning. Still images, video and audio often fall under
the category of unstructured data. Data in this form requires
preprocessing to identify and extract relevant ‘features.’ The features
are structured information that are used for indexing and retrieval,
or training classification, or clustering models.

Data may also be classified by the rate at which it is generated,

collected or processed. The distinction is drawn between streaming
data that arrives constantly like a torrent of water from a fire
hose, and batch data, which arrives in buckets. While there is
rarely a connection between data type and data rate, data rate has
significant influence over the execution model chosen for analytic
implementation and may also inform a decision of analytic class or
learning model.

Take off the Training Wheels 55

 Classes of Analytic Techniques
As a means for helping conceptualize the universe of possible analytic
techniques, we grouped them into nine basic classes. Note that
techniques from a given class may be applied in multiple ways to
achieve various analytic goals. Membership in a class simply indicates
a similar analytic function. The nine analytic classes are shown in the
figure, Classes of Analytic Techniques.
ANALYTIC CLASSES

TRANSFORMING LEARNING PREDICTIVE

Aggregation Enrichment Processing Regression Clustering Classification Recommend Simulation Optimization

Source: Booz Allen Hamilton

Classes of Analytic Techniques

»» Transforming Analytics
›› Aggregation: Techniques to summarize the data. These
include basic statistics (e.g., mean, standard deviation),
distribution fitting, and graphical plotting.
LEARNING STYLE TRAINING STYLE
›› Enrichment: Techniques for adding additional information
to the data, such as source information or other labels.
Unsupervised ››Supervised
Processing:Offline
Techniques that address data cleaning,
Online
preparation, and separation. This group also includes
common algorithm pre-processing activities such as
transformations and feature extraction.

»» Learning Analytics
›› Regression: Techniques for estimating relationships among
variables, including understanding which variables are
important in predicting future values.
SCHEDULING
›› Clustering:SEQUENCING
Techniques to segment the data into naturally
similar groups.

Batch
››Streaming
Classification:
Serial
Techniques
Parallel
to identify data element
group membership.
›› Recommendation: Techniques to predict the rating or
preference for a new entity, based on historic preference
or behavior.

»» Predictive Analytics
›› Simulation: Techniques to imitate the operation of a real-
world process or system. These are useful for predicting
behavior under new conditions.
›› Optimization: Operations Research techniques focused on
selecting the best element from a set of available alternatives
to maximize a utility function.

T H E F I ELD G U I D E to D A T A S C I E N C E
Learning Models
Analytic classes that perform predictions, such as regression,
clustering, classification and recommendation employ learning
models. These models characterize how the analytic is trained to
perform judgments on new data based on historic observation.
Aspects of learning models describe both the types of judgments
performed and how the models evolve over time, as shown in the
figure, Analytic Learning Models.

LEARNING STYLE TRAINING STYLE

Semi-
Unsupervised Supervised Offline Reinforcement Online
Supervised

Source: Booz Allen Hamilton

Analytic Learning Models

Learning models are typically described as belonging to the categories

of unsupervised or supervised learning. Supervised learning takes
place when a model is trained using a labeled dataset that has a known
class or category associated with each data element. The model relates
features found in training instances with labels so that predictions
can be made for unlabeled instances. Unsupervised learning involves »Reinforcement
» Learning
no a-priori knowledge about the classes into which data can be in Action
placed. Unsupervised learning uses the features in the dataset to
form groupings based on feature similarity. Semi-supervised learning The possibilities of Reinforcement
is a hybrid between these two approaches, using a small amount of Learning captured significant
labeled data in conjunction with a large amount of unlabeled data. attention with the publication of a
This is done to improve learning accuracy in cases where only a study in the journal Nature in which
small number of labeled observations are available for learning. a computer agent learned to play 49
different video games with accuracy
There are a variety of ways to train learning models. A useful rivaling a professional game tester [14].
distinction is between those that are trained in a single pass, which are The agent was able to achieve these
known as offline models, and those that are trained incrementally over results using only the raw screen
time, known as online models. Many learning approaches have online pixels and game score as input. This
or offline variants. The decision to use one or another is based on the approach represents the first artificial
intelligence agent that is capable
analytic goals and execution models chosen.
of learning complex tasks while
Generating an offline model requires taking a pass over the entire bridging between high-dimensional
training dataset. Improving the model requires making separate sensory inputs and actions.
passes over the data. These models are static in that once trained, their
predictions will not change until a new model is created through a
subsequent training stage. Offline model performance is easier to
evaluate due to this deterministic behavior. Deployment of the model
into a production environment involves swapping out the old model
for the new.

Online models dynamically evolve over time, meaning they only

require a single deployment into a production setting. The fact that

Take off the Training Wheels 57

 these models do not have the entire dataset available when being
trained is a challenge. They must make assumptions about the data
based on the examples observed; these assumptions may be sub-
optimal. The impact of sub-optimal predictions can be mitigated in
cases where feedback on ANALYTIC CLASSES
the model’s predictions is available. Online
models can rapidly incorporate feedback to improve performance.
TRANSFORMING LEARNING PREDICTIV
One such training style is known as Reinforcement Learning. Under
this approach, an algorithm takes action in an environment and
incrementally learns how to achieve goals based on the response to a
Aggregation Enrichment Processing Regression
function used to determine theClustering
quality ofClassification Recommend
its results. Reinforcement Simulation Optim

learning is generally applicable to complex, real-world tasks that

involve optimization, such as navigation or trading. Due to the
publication of many promising results from Reinforcement Learning
algorithms, the popularity of this technique has risen dramatically in
recent years along with Deep Learning.

Execution Models
Execution models describe how data is manipulated to perform
LEARNING
an analytic function. TheySTYLE TRAINING
may be categorized STYLE
across a number
of dimensions. Execution Models are embodied by an execution
framework, which orchestrates the sequencing of analytic
computation. In this sense, a framework might be as simple as a
Unsupervised Supervised Offline Online
programming language runtime, such as the Python interpreter, or
a distributed computing framework that provides a specific API for
one or more programming languages such as Hadoop, MapReduce
or Spark. Grouping execution models based on how they handle data
is common, classifying them as either batch or streaming execution
models. The categories of execution model are shown in the figure,
Analytic Execution Models.

SCHEDULING SEQUENCING

Batch Streaming Serial Parallel

Source: Booz Allen Hamilton

Analytic Execution Models

A batch execution model implies that data is analyzed in large

segments, that the analytic has a state where it is running and a state
where it is not running and that little state is maintained in memory
between executions. Batch execution may also imply that the analytic
produces results with a frequency on the order of several minutes or
more. Batch workloads tend to be fairly easy to conceptualize because

T H E F I ELD G U I D E to D A T A S C I E N C E
they represent discrete units of work. As such, it is easy to identify
a specific series of execution steps as well as the proper execution
frequency and time bounds based on the rate at which data arrives.
Depending on the algorithm choice, batch execution models are
easily scalable through parallelism. There are a number of frameworks
that support parallel batch analytic execution. Most famously,
Hadoop provides a distributed batch execution model in its
MapReduce framework.

Conversely, a streaming model analyzes data as it arrives. Streaming

execution models imply that under normal operation, the analytic
is always executing. The analytic can hold state in memory and
constantly deliver results as new data arrives, on the order of seconds
or less. Many of the concepts in streaming are inherent in the Unix- »»Tips From the Pros
pipeline design philosophy; processes are chained together by linking
the output of one process to the input of the next. As a result, many In order to understand system capacity
developers are already familiar with the basic concepts of streaming. in the context of streaming analytic
A number of frameworks are available that support the parallel execution, collect metrics including:
execution of streaming analytics such as Storm, S4 and Samza. the amount of data consumed, data
emitted, and latency. This will help
The choice between batch and streaming execution models often you understand when scale limits
hinges on analytic latency and timeliness requirements. Latency refers are reached.
to the amount of time required to analyze a piece of data once it
arrives at the system, while timeliness refers to the average age of an
answer or result generated by the analytic system. For many analytic
goals, a latency of hours and timeliness of days is acceptable and
thus lend themselves to the implementation enabled by the batch
approach. Some analytic goals have up-to-the-second requirements
where a result that is minutes old has little worth. The streaming
execution model better supports such goals.

Batch and streaming execution models are not the only dimensions
within which to categorize analytic execution methods. Another
distinction is drawn when thinking about scalability. In many cases,
scale can be achieved by spreading computation over a number of
computers. In this context, certain algorithms require a large shared
memory state, while others are easily parallelizable in a context
where no shared state exists between machines. This distinction has
significant impacts on both software and hardware selection when
building out a parallel analytic execution environment.

Take off the Training Wheels 59

 Fractal Analytic Model

Data Science analytics are a lot like broccoli.

Fractals are mathematical sets that display self-similar patterns. As

you zoom in on a fractal, the same patterns reappear. Imagine a stalk
of broccoli. Rip off a piece of broccoli and the piece looks much like
the original stalk. Progressively smaller pieces of broccoli still look like
the original stalk.

Data Science analytics are a lot like broccoli – fractal in nature in

both time and construction. Early versions of an analytic follow the
same development process as later versions. At any given iteration, the
analytic itself is a collection of smaller analytics that often decompose
into yet smaller analytics.

T H E F I ELD G U I D E to D A T A S C I E N C E
Iterative by Nature
Good Data Science is fractal in time — ­ an iterative process. Getting
an imperfect solution out the door quickly will gain more interest
from stakeholders than a perfect solution that is never completed. The
figure, The Data Science Product Lifecycle, summarizes the lifecycle of
the Data Science product.

Setup Try Do

Evaluate

Evaluate

Source: Booz Allen Hamilton

The Data Science Product Lifecycle

Set up the infrastructure, aggregate and prepare the data, and

incorporate domain expert knowledge. Try different analytic
techniques and models on subsets of the data. Evaluate the models,
refine, evaluate again, and select a model. Do something with your
models and results – deploy the models to inform, inspire action, and
act. Evaluate the business results to improve the overall product.

Take off the Training Wheels 61

 Smaller Pieces of Broccoli: A Data
Science Product
Components inside and outside of the Data Science product will
change with each iteration. Let’s take a look under the hood of a
Data Science product and examine the components during one
such iteration.

In order to achieve a greater analytic goal, you need to first decompose

the problem into sub-components to divide and conquer. The figure,
The Fractal Analytic Model, shows a decomposition of the Data Science
product into four component pieces.

GOAL CLASSES OF ANALYTICS

› Describe
Aggregation
› Discover
› Predict
› Advise Enrichment

Clustering

Classification
DATA ACTION
› Text › Productization
› Imagery › Data Monetization
› Waveform › Insights & Relationships
› Geo
› Time Series

COMPUTATION

Source: Booz Allen Hamilton

The Fractal Analytic Model

T H E F I ELD G U I D E to D A T A S C I E N C E
GOAL

You must first have some idea of your analytic goal and the end state
of the analysis. Is it to Discover, Describe, Predict, or Advise? It is
probably a combination of several of those. Be sure that before you
start, you define the business value of the data and how you plan to
use the insights to drive decisions, or risk ending up with interesting
but non-actionable trivia.

DATA

Data dictates the potential insights that analytics can provide. Data
Science is about finding patterns in variable data and comparing those
patterns. If the data is not representative of the universe of events you
wish to analyze, you will want to collect that data through carefully
planned variations in events or processes through A/B testing or
design of experiments. Datasets are never perfect so don’t wait for
perfect data to get started. A good Data Scientist is adept at handling
messy data with missing or erroneous values. Just make sure to spend
the time upfront to clean the data or risk generating garbage results.

COMPUTATION

Computation aligns the data to goals through the process of creating

insights. Through divide and conquer, computation decomposes
into several smaller analytic capabilities with their own goals, data,
computation and resulting actions, just like a smaller piece of broccoli
maintains the structure of the original stalk. In this way, computation
itself is fractal. Capability building blocks may utilize different
types of execution models such as batch computation or streaming,
that individually accomplish small tasks. When properly combined
together, the small tasks produce complex, actionable results.

ACTION

How should engineers change the manufacturing process to generate

higher product yield? How should an insurance company choose
which policies to offer to whom and at what price? The output of
computation should enable actions that align to the goals of the data
product. Results that do not support or inspire action are nothing but
interesting trivia.

Given the fractal nature of Data Science analytics in time and

construction, there are many opportunities to choose fantastic or
shoddy analytic building blocks. The Analytic Selection Process offers
some guidance.

Take off the Training Wheels 63

 The Analytic
Selection Process

If you focus only on the science aspect of Data Science you will
never become a data artist.

A critical step in Data Science is to identify an analytic technique that

will produce the desired action. Sometimes it is clear; a characteristic
of the problem (e.g., data type) points to the technique you should
implement. Other times, however, it can be difficult to know where
to begin. The universe of possible analytic techniques is large. Finding
your way through this universe is an art that must be practiced. We
are going to guide you on the next portion of your journey - becoming
a data artist.

T H E F I ELD G U I D E to D A T A S C I E N C E
Decomposing the Problem
Decomposing the problem into manageable pieces is the first step
in the analytic selection process. Achieving a desired analytic action
often requires combining multiple analytic techniques into a holistic,
end-to-end solution. Engineering the complete solution requires that
the problem be decomposed into progressively smaller sub-problems.

The Fractal Analytic Model embodies this approach. At any given

stage, the analytic itself is a collection of smaller computations that
decompose into yet smaller computations. When the problem is
decomposed far enough, only a single analytic technique is needed
to achieve the analytic goal. Problem decomposition creates multiple
sub-problems, each with their own goals, data, computations, and
actions. The concept behind problem decomposition is shown in the
figure, Problem Decomposition Using the Fractal Analytic Model.

GOAL CLASSES OF ANALYTICS

› Describe
Aggregation
› Discover
› Predict
› Advise Enrichment

Clustering

Classification
DATA ACTION
› Text › Productization
› Imagery › Data Monetization
› Waveform › Insights & Relationships
› Geo
› Time Series

GOAL

DATA ACTION

GOAL

DATA ACTION

Source: Booz Allen Hamilton

Problem Decomposition Using the Fractal Analytic Model

Take off the Training Wheels 65

 On the surface, problem decomposition appears to be a mechanical,
repeatable process. While this may be true conceptually, it is
really the performance of an art as opposed to the solving of an
engineering problem. There may be many valid ways to decompose
the problem, each leading to a different solution. There may be
hidden dependencies or constraints that only emerge after you begin
developing a solution. This is where art meets science. Although the
art behind problem decomposition cannot be taught, we have distilled
some helpful hints to help guide you. When you begin to think about
decomposing your problem, look for:

›› Compound analytic goals that create natural segmentation.

For example, many problems focused on predicting future
conditions include both Discover and Predict goals.
»»Tips From the Pros ›› Natural orderings of analytic goals. For example, when extracting
features you must first identify candidate features and then
One of your first steps should be to select the features set with the highest information value. These
explore available data sources that two activities form distinct analytic goals.
have not been previously combined.
Emerging relationships between data
sources often allow you to pick low ›› Data types that dictate processing activities. For example, text or
hanging fruit. imagery both require feature extraction.
›› Requirements for human-in-the-loop feedback. For example,
when developing alerting thresholds, you might need to solicit
analyst feedback and update the threshold based on their
assessment.
›› The need to combine multiple data sources. For example, you may
need to correlate two datasets to achieve your broader goal.
Often this indicates the presence of a Discover goal.

In addition to problem decomposition providing a tractable approach

to analytic selection, it has the added benefit of simplifying a highly
complex problem. Rather than being faced with understanding the
entire end-to-end solution, the computations are discrete segments
that can be explored. Note, however, that while this technique helps
the Data Scientist approach the problem, it is the complete end-to-
end solution that must be evaluated.

T H E F I ELD G U I D E to D A T A S C I E N C E
 ›› Identifying Spoofed Domains Identifying spoofed domains is important for an organization
to preserve their brand image and to avoid eroded customer
confidence. Spoofed domains occur when a malicious actor
creates a website, URL or email address that users believe is
associated with a valid organization. When users click the link,
visit the website or receive emails, they are subjected to some
type of nefarious activity.
Stephanie
Rivera

Discover spoofed
domains

Compare Test &

List of
recently Datasets Evaluation
Alert on spoofed domains to Discover likely
registered provide opportuinity to candidates for
company minimize brand image and spoofed domains
domains Simulate
Data consumer confidence damage
Spoofed
Collection
Data

Store
Generated
List of recently Domains
List of candidate
Describe closeness of registered spoofed domains
spoof to valid domains company domains Generate
Test &
Candidate
Evaluation
› List of Domains
candidate
spoofed
domains Quantitative
Calculate Set Threshold that balances
› List of measure of feature
Metric Threshold false positive and false
recently information value
negative rate
registered
company
domains
› List of candidate spoofed
domains
› List of recently registered Test & Quantitative threshold for
company domains Evaluation automated result ranking
› Quantitative measure of
feature information value

Source: Booz Allen Hamilton

Spoofed Domain Problem Decomposition

Our team was faced with the problem of alert speeds are vital. Result speed created
identifying spoofed domains for a commercial an implementation constraint that forced us to
company. On the surface, the problem sounded re-evaluate how we decomposed the problem.
easy; take a recently registered domain, check
to see if it is similar to the company’s domain Revisiting the decomposition process led us
and alert when the similarity is sufficiently high. to a completely new approach. In the end,
Upon decomposing the problem, however, the we derived a list of domains similar to those
main computation quickly became complicated. registered by the company. We then compared
that list against a list of recently registered
We needed a computation that determined domains. The figure, Spoofed Domain Problem
similarity between two domains. As we Decomposition, illustrates our approach. Upon
decomposed the similarity computation, testing and initial deployment, our analytic
complexity and speed became a concern. discovered a spoofed domain within 48 hours.
As with many security-related problems, fast

Take off the Training Wheels 67

Implementation Constraints
In the spoofed domains case study, the emergence of an
implementation constraint caused the team to revisit its approach.
This demonstrates that analytic selection does not simply mean
choosing an analytic technique to achieve a desired outcome. It also
means ensuring that the solution is feasible to implement.

The Data Scientist may encounter a wide variety of implementation

constraints. They can be conceptualized, however, in the context of
five dimensions that compete for your attention: analytic complexity,
speed, accuracy & precision, data size, and data complexity. Balancing
these dimensions is a zero sum game - an analytic solution cannot
simultaneously exhibit all five dimensions, but instead must make
trades between them. The figure, Balancing the Five Analytic
Dimensions, illustrates this relationship.

SPEED: The speed at which an

analytic outcome must be produced
(e.g., near real-time, hourly, daily) or the time

SPEED it takes to develop and implement the

analytic solution

ANALYTIC COMPLEXITY:
Algorithmic complexity (e.g., complexity class
and execution resources)
DATA ANALYTIC
COMPLEXITY COMPLEXITY ACCURACY & PRECISION: The ability
to produce exact versus approximate
solutions as well as the ability to provide a
measure of confidence

DATA SIZE: The size of the dataset

(e.g., number of rows)

ACCURACY
DATA COMPLEXITY: The data
type, formal complexity measures
DATA SIZE &
including measures of overlap and
PRECISION
linear separability, number of
dimensions /columns, and linkages
between datasets

Source: Booz Allen Hamilton

Balancing the Five Analytic Dimensions

Implementation constraints occur when an aspect of the problem

As we compiled this section, we
dictates the value for one or more of these dimensions. As soon as
talked extensively about ways to
one dimension is fixed, the Data Scientist is forced to make trades
group and classify implementation
among the others. For example, if the analytic problem requires constraints. After much discussion
actions to be produced in near real-time, the speed dimension is we settled on these five dimensions.
fixed and trades must be made among the other four dimensions. We present this model in hopes
Understanding which trades will achieve the right balance among that others weigh in and offer
the five dimensions is an art that must be learned over time. their own perspectives.

Take off the Training Wheels 69

 Some common examples of implementation
constraints include:
• Computation frequency. The solution may need to run on a regular
basis (e.g., hourly), requiring that computations be completed
within a specified window of time. The best analytic is useless if it
»»Tips From the Pros
cannot solve the problem within the required time.
When possible, consider approaches • Solution timeliness. Some applications require near real-time
that make use of previously computed results, pointing toward streaming approaches. While some
results. Your algorithm will run algorithms can be implemented within streaming frameworks,
much faster if you can avoid many others cannot.
re-computing values across
the full time horizon of data.
• Implementation speed. A project may require that you rapidly
develop and implement a solution to quickly produce analytic
insights. In these cases, you may need to focus on less complex
techniques that can be quickly implemented and verified.
• Computational resource limitations. Although you may be able
to store and analyze your data, data size may be sufficiently large
that algorithms requiring multiple computations across the full
»»Tips From the Pros dataset are too resource intensive. This may point toward needing
approaches that only require a single pass on the data (e.g., canopy
Our Data Science Product Lifecycle cluster as opposed to k-means clustering).
has evolved to produce results quickly
and then incrementally improve
the solution. • Data storage limitations. There are times when big data becomes
so big it cannot be stored or only a short time horizon can be
stored. Analytic approaches that require long time horizons may
not be possible.

Organizational policies and regulatory requirements are a major

source of implicit constraints that merit a brief discussion. Policies
are often established around specific classes of data such as Personally
Identifiable Information (PII) or Personal Health Information (PHI).
»»Tips From the Pros
While the technologies available today can safely house information
with a variety of security controls in a single system, these policies
Streaming approaches may be useful
force special data handling considerations including limited retention
for overcoming storage limitations.
periods and data access. Data restrictions impact the data size and
complexity dimensions outlined earlier, creating yet another layer of
constraints that must be considered.

T H E F I ELD G U I D E to D A T A S C I E N C E
 FILTERING
How do I identify

Guide to Analytic Selection data based on

its absolute or
relative values?

IMPUTATION
Your senses are incapable of perceiving the entire universe, so How do I fill in
we drew you a map. missing values
in my data?

The universe of analytic techniques is vast and hard to comprehend. DIMENSIONALITY

We created this diagram to aid you in finding your way from data REDUCTION 
and goal to analytic action. Begin at the center of the universe How do I reduce
the number of
(Data Science) and answer questions about your analytic goals and
dimensions
problem characteristics. The answers to your questions will guide you in my data?
through the diagram to the appropriate class of analytic techniques
and provide recommendations for a few techniques to consider.
NORMALIZATION &
TRANSFORMATION
How do I reconcile
duplication
representations
in the data?

1 DESCRIBE FEATURE
EXTRACTION 

Data
Science 2 DISCOVER PROCESSING 
How do I clean
and separate
my data?

3 PREDICT

4 ADVISE
1
DESCRIBE
How do I develop
an understanding
of the content of
my data?

TIP: There are several situations where dimensionality reduction may be needed:
 
›› Models fail to converge feature space
›› Models produce results equivalent to
random chance
›› You do not know which aspects of the
data are
2
›› You lack the computational power to ›› the most important DISCOVER
perform operations across the
Feature Extraction is a broad topic and is highly dependent upon the domain area.
  3
This topic could be the subject of an entire book. As a result, a detailed exploration
has been omitted from this diagram. See the Featuring Engineering and Feature PREDICT
Selection sections in the Life in the Trenches chapter for additional information.
TIP: Always check data labels for correctness. This is particularly true for time stamps,
  4 ENRICH
which may have reverted to system default values. How do
ADVISE
TIP: Smart enrichment can greatly speed-up computational time. It can also be a huge
  new info
differentiator between the accuracy of different end-to-end analytic solutions. to my d
Data
Source: Booz Allen Hamilton
Science
Take off the Training Wheels 71
 If you want to add or remove data based on its value, start with:
 › Relational algebra projection and selection
FILTERING If early results are uninformative and duplicative, start with:
How do I identify › Outlier removal › Gaussian filter
data based on › Exponential smoothing › Median filter
its absolute or
relative values? If you want to generate values from other observations in your dataset, start with:
› Random sampling
› Markov Chain Monte Carlo (MC)
IMPUTATION
How do I fill in If you want to generate values without using other observations in your dataset,
missing values start with:
› Mean › Regression models
in my data? › Statistical distributions

DIMENSIONALITY If you need to determine whether there is multi-dimensional correlation,

REDUCTION  start with:
› PCA and other factor analysis
How do I reduce
the number of If you can represent individual observations by membership in a group, start with:
dimensions › K-means clustering
in my data? › Canopy clustering

If you have unstructured text data, start with:

NORMALIZATION & › Term Frequency/Inverse Document Frequency (TF IDF)
TRANSFORMATION
If you have a variable number of features but your
How do I reconcile algorithm requires a fixed number, start with:
duplication › Feature hashing
representations
in the data? If you are not sure which features are the most important, start with:
› Wrapper methods
› Sensitivity analysis
FEATURE
If you need to facilitate understanding of the
EXTRACTION  probability distribution of the space, start with:
› Self organizing maps

If you suspect duplicate data elements, start with:

PROCESSING  › Deduplication
DES
How do I clean
If you want your data to fall within a specified range, start with:
and separate › Normalization
my data?
If your data is stored in a binary format, start with:
› Format Conversion
DISC
If you are operating in frequency space, start with: What
› Fast Fourier Transform (FFT), key r
› Discrete wavelet transform in the
If you are operating in Euclidian space, start with:
1 › Coordinate transform

DESCRIBE
How do I develop If you are unfamiliar with the dataset, start with
an understanding basic statistics: PR
of the content of › Count › Standard deviation › Box plots
my data? AGGREGATION › Mean › Range › Scatter plots
How do I collect
and summarize If your approach assumes the data follows a AD
my data? distribution, start with:
› Distribution fitting
2 If you want to understand all the information Sc
DISCOVER available on an entity, start with:
› “Baseball card” aggregation
3
PREDICT

4 ENRICHMENT  If you need to keep track of source information or other

How do I add user-defined parameters, start with:
ADVISE
new information › Annotation
to my data?
If you often process certain data fields together or use
Data
Science one field to compute the value of another, start with:
› Relational algebra rename,
› Feature addition (e.g., Geography, Technology, Weather)
Source: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
tart with:

ataset, CLASSIFICAT
If you want an ordered set of clusters with variable precision, start with:
How do I
› Hierarchical
predict group
 If you have a known number of clusters, start with: membership?
› X-means
› Canopy
› Apriori
tart with: If you have text data, start with:
› Topic modeling

If you have non-elliptical clusters, start with:

› Fractal
› DB Scan

CLUSTERING If you want soft membership in the clusters, start with:

How do I segment › Gaussian mixture models
the data to
 If you have an known number of clusters, start with:
find natural
› K-means
groupings?

1 If your data has unknown structure, start with:

› Tree-based methods
DESCRIBE
If statistical measures of importance are
needed, start with:
2 › Generalized linear models

DISCOVER REGRESSION If statistical measures of importance are not

What are the How do I needed, start with:
key relationships determine which › Regression with shrinkage (e.g., LASSO, Elastic net)
in the data? variables may › Stepwise regression
be important?

3 If you want to compare two groups

PREDICT › T-test

If you want to compare multiple groups

4 › ANOVA
ADVISE
HYPOTHESIS
TESTING
Data
Science How do I
test ideas?

TIP: Canopy clustering is good when you only want to make a single pass over the data.
 
TIP: Use canopy or hierarchical clustering to estimate the number of clusters you
 
should generate.
Source: Booz Allen Hamilton
 If you have known dependent relationships between variables
› Bayesian network

If you are unsure of feature importance, start with:

› Neural nets,
› Random forests Da
Scie
CLASSIFICATION  › Deep learning
with:
How do I
predict group If you require a highly transparent model, start with:
membership? › Decision trees 1
If you have <20 data dimensions, start with: DESC
› K-nearest neighbors
2
If you have a large dataset with an unknown classification DISCO
signal, start with:
› Naive bayes 3
Data PRED
Science If you want to estimate an unobservable state based on
observable variables, start with:
› Hidden markov model

1 4
 If you don't know where else to begin, start with:
DESCRIBE › Support vector machines (SVM) ADVISE
› Random forests What co
2 of action
should I
DISCOVER

3 If the data structure is unknown,

start with:
e, start with: › Tree-based methods
PREDICT
What are the
REGRESSION If you require a highly
ce are likely future
How do I predict transparent model, start with:
outcomes?
a future value? › Generalized linear models

If you have <20 data dimensions,

ce are not start with:
› K-nearest neighbors
LASSO, Elastic net)
4
ADVISE

If you only have knowledge of how people interact with items,

start with:
› Collaborative filtering
RECOMMENDATION 
If you have a feature vector of item characteristics, start with:
How do I predict relevant
› Content-based methods
conditions?
If you only have knowledge of how items are connected to one
another, start with:
› Graph-based methods

 TIP: It can be difficult to predict which classifier will work best on your dataset.
Always try multiple classifiers. Pick the one or two that work the best to refine and
explore further.
 TIP: These are our favorite, go-to classification algorithms.
 TIP: Be careful of the “recommendation bubble”, the tendency of recommenders to
recommend only what has been seen in the past.
You must ensure you add diversity to avoid this phenomenon.
 TIP: SVD and PCA are good tools for creating better features for recommenders.
Source: Booz Allen Hamilton
 If you have expert knowledge to capture
› Expert systems

s If you're looking for basic facts

› Logical reasoning
LOGICAL
REASONING
How do I sort
Data through different If your problem is represented by a
Science evidence? non-deterministic utility function, start with:
› Stochastic search

1 If approximate solutions are acceptable,

DESCRIBE start with:
› Genetic algorithms
2 › Simulated annealing
› Gradient search
DISCOVER
If your problem is represented by a
3 OPTIMIZATION deterministic utility function, start with:
PREDICT How do I identify the › Linear programming
best course of action › Integer programming
when my objective › Non-linear programming
4 can be expressed as
If you have limited resources to search with
a utility function?
› Active learning
ADVISE
What course If you want to try multiple models
of action › Ensemble learning
should I take?

unknown, If you must model discrete entities, start with:

› Discrete event simulation (DES)

If there are a discrete set of possible states, start with:

› Markov models
rt with:
odels If there are actions and interactions among autonomous
entities, start with:
mensions, › Agent-based simulation
SIMULATION
How do I characterize If you do not need to model discrete entities, start with:
a system that does not › Monte Carlo simulation
have a closed-form
representation? If you are modeling a complex system with feedback
mechanisms between actions, start with:
› Systems dynamics
ct with items,
If you require continuous tracking of system behavior,
start with:
› Activity-based simulation
cs, start with:
If you already have an understanding of what factors
govern the system, start with:
nnected to one › ODES
› PDES

If you have imprecise categories

› Fuzzy logic

Source: Booz Allen Hamilton

 Detailed Table of Analytics

Getting you to the right starting point isn’t enough. We also

provide a translator so you understand what you’ve been told.

Identifying several analytic techniques that can be applied to your

problem is useful, but their name alone will not be much help. The
Detailed Table of Analytics translates the names into something more
meaningful. Once you’ve identified a technique in the Guide to
Analytic Selection, find the corresponding row in the table. There you
will find a brief description of the techniques, tips we’ve learned and a
few references we’ve found helpful.

References we
Technique Description Tips From the Pros
love to read

Intelligent sample selection

Burr, Settles B. “Active
to improve performance
Learning: Synthesis Lectures
Active of model. Samples are Can be paired with a human in-the-loop
on Artificial Intelligence and
Learning selected to provide the to help capture domain knowledge.
Machine Learning.” Morgan
greatest information
& Claypool, 2012. Print.
to a learning model.

Macal, Charles, and Michael

In many systems, complex behavior
Simulates the actions North. “Agent-based Modeling and
Agent Based results from surprisingly simple rules.
and interactions of Simulation.” Winter Simulation
Simulation autonomous agents.
Keep the logic of your agents simple
Conference. Austin, TX. 2009.
and gradually build in sophistication.
Conference Presentation.

Bhattacharyya, Gouri K.,

Hypothesis testing for Check model assumptions before
and Richard A. Johnson.
ANOVA differences between utilizing, and watch out for Family Wise
Statistical Concepts and
more than two groups. error when running multiple tests.
Models. Wiley, 1977. Print.

Agrawal, Rakesh, and

Association Ramakrishnan Srikant.
Data mining technique
Utilize when you have a need to understand “Fast Algorithms for Mining
Rule Mining to identify the common
potential relationships between elements. Association Rules.” Proc. Of
(Apriori) co-occurances of items.
20th Intl. Conf. on VLDB. 1994.
Conference Presentation.

Models conditional Russel, Stuart, and Peter

Bayesian probabilities amongst Calculate by hand before using larger Norvig. “Artificial Intelligence:
Network elements, visualized as a models to ensure understanding. A Modern Approach.” Prentice
Directed Acyclic Graph. Hall, 2009 Print.
Compiled by: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
 References we
Technique Description Tips From the Pros
love to read

Also known as
'Recommendation,' suggest
or eliminate items from a
Use Singular Value Decomposition Owen, Sean, Robin Anil, Ted
set by comparing a history
Collaborative based Recommendation for cases Dunning, and Ellen Friedman.
of actions against items
Filtering performed by users. Finds
where there are latent factors in your Mahout in Action. New Jersey:
domain, e.g., genres in movies. Manning, 2012. Print.
similar items based on who
used them or similar users
based on the items they use.

Changing the coordinate system for data, for

Coordinate example, using polar or cylindrical coordinates, Abbott, Edwin A., Flatland: A
Provides a different may more readily highlight key structure in the Romance of Many Dimensions.
Transforma- perspective on data. data. A key step in coordinate transformations United Kingdom: Seely & Co.,
tion is to appreciate multidimensionality and to 1884. Print.
systematically analyze subspaces of the data.

Method that learns Bengio, Yoshua, and Yann

features that leads to LeCun. “Scaling Learning
Utilize a GPU to efficiently train complex
Deep Learning higher concept learning.
models.
Algorithms towards AI.” Large-
Usually very deep neural Scale Kernel Machines. New
network architectures. York: MIT Press, 2007. Print.

Applies controlled Montgomery, Douglas. Design

Fractional factorial designs can significantly
Design of experiments to quantify and Analysis of Experiments.
reduce the number of different types of
Experiments effects on system output
experiments you must conduct.
New Jersey: John Wiley
caused by changes to inputs. & Sons, 2012. Print.

Used to express Differential equations can be used to formalize Zill, Dennis, Warren Wright,
relationships between models and make predictions. The equations and Michael Cullen. Differential
Differential
functions and their themselves can be solved numerically and Equations with Boundary-Value
Equations derivatives, for example, tested with different initial conditions to study Problems. Connecticut: Cengage
change over time. system trajectories. Learning, 2012. Print.

Simulates a discrete Discrete event simulation is useful when Burrus, C. Sidney, Ramesh A.
sequence of events where analyzing event based processes such as Gopinath, Haitao Guo, Jan E.
each event occurs at a production lines and service centers to Odegard and Ivan W. Selesnick.
Discrete Event
particular instant in time. determine how system level behavior changes Introduction to Wavelets and
Simulation The model updates its state as different process parameters change. Wavelet Transforms: A Primer.
only at points in time when Optimization can integrate with simulation to New Jersey: Prentice Hall, 1998.
events occur. gain efficiencies in a process. Print.

Burrus, C.Sidney, Ramesh

A. Gopinath, Haitao Guo,
Discrete Transforms time series Offers very good time and frequency
Jan E. Odegard, and Ivan W.
data into frequency localization. The advantage over Fourier
Wavelet domain preserving transforms is that it preserves both frequency
Selesnick. Introduction to
Transform Wavelets and Wavelet Transforms:
locality information. and locality.
A Primer. New Jersey:
Prentice Hall, 1998. Print.

Dietterich, Thomas G. “Ensemble

Learning multiple models
Ensemble Be careful not to overfit data by having too many Methods in Machine Learning.”
and combining output to
Learning achieve better performance.
model parameters and overtraining. Lecture Notes in Computer
Science. Springer, 2000. Print.

Shortliffe, Edward H., and

Systems that use symbolic Bruce G. Buchanan. “A Model of
Expert Useful to have a human readable explanation of
logic to reason about facts. Inexact Reasoning in Medicine.”
Systems Emulates human reasoning.
why a system came to a conclusion.
Mathematical Biosciences.
Elsevier B.V., 1975. Print.

Chatfield, Chris, Anne B.

Koehler, J. Keith Ord, and
In comparison to a using moving average Ralph D. Snyder. “A New Look
Used to remove artifacts
Exponential where past observations are weighted equally, at Models for Exponential
expected from collection
Smoothing error or outliers.
exponential smoothing assigns exponentially Smoothing.” Journal of the Royal
decreasing weights over time. Statistical Society: Series D (The
Statistician). Royal Statistical
Society, 2001. Print.

Compiled by: Booz Allen Hamilton

Take off the Training Wheels 79

 References we
Technique Description Tips From the Pros
love to read

Describes variability among

correlated variables with the If you suspect there are inmeasurable Child, Dennis. The Essentials of
Factor
goal of lowering the number influences on your data, then you may want to Factor Analysis. United Kingdom:
Analysis of unobserved variables, try factor analysis. Cassell Educational, 1990. Print.
namely, the factors.

Transforms time series from

Filtering a time varying signal can be done Mitra, Partha P., and Hemant
time to frequency domain
Fast Fourier more effectively in the frequency domain. Also, Bokil. Observed Brain Dynamics.
efficiently. Can also be used
Transform for image improvement by
noise can often be identified in such signals by United Kingdom: Oxford
observing power at aberrant frequencies. University Press, 2008. Print.
spatial transforms.

Creates a standard
representation of data Ingersoll, Grant S., Thomas S.
There are a number of open source software
regardless of source format. Morton, and Andrew L. Farris.
Format packages that support format conversion and
For example, extracting raw Taming Text: How to Find,
Conversion UTF-8 encoded text from
can interpret a wide variety of formats. One
Organize, and Manipulate It. New
notable package is Apache Tikia.
binary file formats such as Jersey: Manning, 2013. Print.
Microsoft Word or PDFs.

Utilize when categories are not clearly defined. Zadeh L.A., "Fuzzy Sets.”
Logical reasoning that
Concepts such as "warm", "cold", and "hot" can Information and Control.
Fuzzy Logic allows for degrees of truth
mean different things at different temperatures California: University of
for a statement.
and domains. California, Berkeley, 1965. Print.

Parker, James R. Algorithms for

Gaussian Acts to remove noise Can be used to remove speckle Image Processing and Computer
Filtering or blur data. noise from images. Vision. New Jersey: John Wiley &
Sons, 2010. Print.

Expands ordinary linear MacCullagh, P., and John A.

Generalized regression to allow Use if the observed error in your system does Nelder. Generalized Linear
Linear Models for error distribution not follow the normal distribution. Models. Florida: CRC Press,
that is not normal. 1989. Print.

Evolves candidate Increasing the generation size adds diversity

models over generations in considering parameter combinations, but De Jong, Kenneth A. Evolutionary
Genetic by evolutionary requires more objective function evaluation. Computation - A Unified
Algorithms inspired operators of Calculating individuals within a generation Approach. Massachusetts:
mutation and crossover is strongly parallelizable. Representation of MIT Press, 2002. Print.
of parameters. candidate solutions can impact performance.

Kolda, Tamara G., Robert M.

Lewis, and Virginia Torczon.
Systematic search across “Optimization by Direct
A grid across the parameters is used to
discrete parameter Search: New Perspectives on
Grid Search values for parameter
visualize the parameter landscape and assess
Some Classical and Modern
whether multiple minima are present.
exploration problems. Methods.” SIAM Review. Society
for Industrial and Applied
Mathematics, 2003. Print.

One of the most powerful properties of Hidden

Markov Models is their ability to exhibit
Models sequential data by
some degree of invariance to local warping Bishop, Christopher M. Pattern
determining the discrete
Hidden Markov (compression and stretching) of the time axis. Recognition and Machine
latent variables, but
Models the observables may be
However, a significant weakness of the Hidden Learning. New York: Springer,
Markov Model is the way in which it represents 2006. Print.
continuous or discrete.
the distribution of times for which the system
remains in a given state.

Connectivity based
Provides views of clusters at multiple
clustering approach
resolutions of closeness. Algorithms Rui Xu, and Don Wunsch.
Hierarchical that sequentially builds
begin to slow for larger datasets due Clustering. New Jersey: Wiley-
Clustering bigger (agglomerative)
to most implementations exhibiting IEEE Press, 2008. Print.
or smaller (divisive)
O(N3) or O(N2) complexity.
clusters in the data.

Centroid based clustering

K-means algorithms, where with When applying clustering techniques, make
Rui Xu, and Don Wunsch.
K means the number sure to understand the shape of your data.
and X-means of clusters is set and X Clustering techniques will return poor results if
Clustering. New Jersey: Wiley-
Clustering IEEE Press, 2008. Print.
means the number of your data is not circular or ellipsoidal shaped.
clusters is unknown.

Compiled by: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
 References we
Technique Description Tips From the Pros
love to read

Linear, Set of techniques for Winston, Wayne L. Operations

Start with linear programs because algorithms
Non-linear, minimizing or maximizing a Research: Applications and
for integer and non-linear variables can take
and Integer function over a constrained
much longer to run.
Algorithms. Connecticut:
Programming set of input parameters. Cengage Learning, 2003. Print.

Problems that are intractable using analytic

Andrieu, Christophe, Nando
approaches can become tractable using MCMC,
A method of sampling de Freitas, Amaud Doucet,
Markov Chain when even considering high-dimensional
typically used in Bayesian and Michael I. Jordan. “An
problems. The tractability is a result of using
Monte Carlo models to estimate the joint
statistics on the underlying distributions of
Introduction to MCMC for
(MCMC) distribution of parameters Machine Learning.” Machine
interest, namely, sampling with Monte Carlo and
given the data. Learning. Kluwer Academic
considering the stochastic sequential process of
Publishers, 2003. Print.
Markov Chains.

Particularly useful for numerical integration, Fishman, George S. Monte

Set of computational
Monte Carlo solutions of differential equations, computing Carlo: Concepts, Algorithms, and
techniques to generate
Methods random numbers.
Bayesian posteriors, and high dimensional Applications. New York: Springer,
multivariate sampling. 2003. Print.

Predicts classes following Assumes that all variables are independent,

Bayes Theorem that so it can have issues learning in the context of Ingersoll, Grant S., Thomas S.
states the probability of highly interdependent variables. The model can Morton, and Andrew L. Farris.
Naïve Bayes an outcome given a set of be learned on a single pass of data using simple Taming Text: How to Find,
features is based on the counts and therefore is useful in determining Organize, and Manipulate It. New
probability of features given whether exploitable patterns exist in large Jersey: Manning, 2013. Print.
an outcome. datasets with minimal development time.

Training a neural network takes substantially

Learns salient features in Haykin, Simon O. Neural
longer than evaluating new data with an already
Neural data by adjusting weights Networks and Learning
trained network. Sparser network connectivity
Networks between nodes through a
can help to segment the input space and
Machines. New Jersey:
learning rule. Prentice Hall, 2008. Print.
improve performance on classification tasks.

Maimon, Oded, and Lior

Rockach. Data Mining and
Method for identifying and Be cautious when removing outliers. Sometimes Knowledge Discovery Handbook:
Outlier
removing noise or artifacts the most interesting behavior of a system is at A Complete Guide for Practitioners
Removal from data. times when there are aberrant data points. and Researchers. The
Netherlands: Kluwer Academic
Publishers, 2005. Print.

Many large datasets contain correlations Wallisch, Pascal, Michael E.

between dimensions; therefore part of the Lusignan, Marc D. Benayoun,
Principal Enables dimensionality
dataset is redundant. When analyzing the Tanya I. Baker, Adam Seth
reduction by identifying
Components highly correlated
resulting principal components, rank order Dickey, and Nicholas G.
Analysis them by variance as this is the highest Hatsopoulos. Matlab for
dimensions.
information view of your data. Use skree plots to Neuroscientists. New Jersey:
infer the optimal number of components. Prentice Hall, 2009. Print.

Bergstra J. and Bengio Y.

Randomly adjust
Use as a benchmark for how well a search Random Search for Hyper-
Random parameters to find a
algorithm is performing. Be careful to use a Parameter Optimization,
Search better solution than
good random number generator and new seed. Journal of Machine Learning
currently found.
Research 13, 2012.

Tibshirani, Robert. “Regression

Regression A method of variable Shrinkage and Selection via the
There are different methods to select the
selection and prediction Lasso.” Journal of the Royal
with Shrinkage combined into a possibly
lambda parameter. A typical choice is cross
Statistical Society. Series B
(Lasso) validation with MSE as the metric.
biased linear model. (Methodological). Toronto: Royal
Statistical Society, 1996. Print.

Saltelli, A., Marco Ratto, Terry

Insensitive model parameters during an Andres, Francesca Campolongo,
Involves testing individual
optimization are candidates for being set to Jessica Cariboni, Debora Gatelli,
Sensitivity parameters in an analytic
constants. This reduces the dimensionality Michaela Saisana, and Stefano
Analysis or model and observing the
of optimization problems and provides an Tarantola. Global Sensitivity
magnitude of the effect.
opportunity for speed up. Analysis: the Primer. New Jersey:
John Wiley & Sons, 2008. Print.

Compiled by: Booz Allen Hamilton

Take off the Training Wheels 81

 References we
Technique Description Tips From the Pros
love to read

Named after a controlled

The standard annealing function allows for
cooling process in
initial wide exploration of the parameter space
metallurgy, and by Bertsimas, Dimitris, and John
Simulated followed by a narrower search. Depending on
analogy using a changing Tsitsiklis. “Simulated Annealing.”
Annealing temperature or annealing
the search priority the annealing function can
Statistical Science. 1993. Print.
be modified to allow for longer explorative
schedule to vary
search at a high temperature.
algorithmic convergence.

A method of variable
selection and prediction.
Akaike's information
criterion AIC is used as
Caution must be used when considering Hocking, R.R. “The Analysis
the metric for selection.
Stepwise Stepwise Regression, as over fitting often and Selection of Variables in
The resulting predictive
Regression model is based upon
occurs. To mitigate over fitting try to limit the Linear Regression.” Biometrics.
number of free variables used. 1976. Print.
ordinary least squares,
or a general linear model
with parameter estimation
via maximum likelihood.

General-purpose Witten, Ian H., Eibe Frank,

Stochastic optimization for learning of Applied in cases where the objective and Mark A. Hall. Data Mining:
Gradient neural networks, support function is not completely differentiable Practical Machine Learning Tools
Descent vector machines, and when using sub-gradients. and Techniques. Massachusetts:
logistic regression models. Morgan Kaufmann, 2011. Print.

Hsu, Chih-Wei, Chih-Chung

Projection of feature vectors
Chang, and Chih-Jen Lin. “A
Support Vector using a kernel function into Try multiple kernels and use k-fold cross
Practical Guide to Support Vector
Machines a space where classes are validation to validate the choice of the best one.
Classification.” National Taiwan
more separable.
University Press, 2003. Print.

Typically used in text mining. Assuming a

Term corpus of news articles, a term that is very Ingersoll, Grant S., Thomas S.
Frequency A statistic that measures frequent such as “the” will likely appear many Morton, and Andrew L. Farris.
Inverse the relative importance of a times in many documents, having a low value. A Taming Text: How to Find,
Document term from a corpus. term that is infrequent such as a person’s last Organize, and Manipulate It. New
Frequency name that appears in a single article will have a Jersey: Manning, 2013. Print.
higher TD IDF score.

Topic Modeling Blei, David M., Andrew Y. Ng,

Identifies latent topics Employ part-of-speech tagging to eliminate and Michael I. Jordan. “Latent
(Latent
in text by examining words other than nouns and verbs. Use raw Dirichlet Allocation.” Journal
Dirichlet word co-occurrence. term counts instead of TF/IDF weighted terms. of Machine Learning Research.
Allocation) 2003. Print.

James, G., D. Witten, T. Hastie,

Models structured as graph and R. Tibshirani. “Tree Based
Tree Based Can be used to systematize a process or act as
trees where branches Methods.” An Introduction to
Methods indicate decisions.
a classifier.
Statistical Learning. New York:
Springer, 2013. Print.

Bhattacharyya, Gouri K., and

Hypothesis test used to Make sure you meet the tests assumptions and
Richard A. Johnson. Statistical
T-Test test for differences between watch out for Family Wise error when running
Concepts and Models. Wiley,
two groups. multiple tests.
1977. Print.

Feature set reduction John, George H., Ron

method that utilizes Kohavi, and Karl Pfleger.
performance of a set of “Irrelevant Features and the
features on a model, as Subset Selection Problem.”
Wrapper Utilize k-fold cross validation
a measure of the feature Proceedings of ICML-94, 11th
Methods set’s performance. Can
to control over fitting.
International Converence
help identify combinations on Machine Learning. New
of features in models that Brunswick, New Jersey. 1994.
achieve high performance. Conference Presentation.

Compiled by: Booz Allen Hamilton

T H E F I ELD G U I D E to D A T A S C I E N C E
LI F E in T H E T R E N C H E S
NAVIGATING NECK DEEP IN DATA
Our Data Science experts have learned
and developed new solutions over the years
from properly framing or reframing analytic
questions. In this section, we list a few
important topics to Data Science coupled
with firsthand experience from our experts.
 Going Deep into
Machine Learning

Machines are getting better at learning by mimicking the

human brain.

Think about where you were 10 years ago. Could computers understand
and take action based upon your spoken word? Recently, speech-to-text
quality has improved dramatically to nearly perfect accuracy, much to
the delight of many mobile phone users. In other complex tasks, similar
magic capabilities have emerged. The world-record high scores in 29
video games are now held by a machine learning algorithm with no
specific knowledge of Atari or computer games in general.

These impressive feats were made possible by deep learning, a different

way of approaching machine learning problems. Most approaches to
machine learning require humans to encode logic and rules to create
features, which are then fed into machine learning models. In some
domains, such as audio, text, image, and signal processing, effective
feature engineering requires considerable human expertise to achieve
decent model performance. Deep learning avoids the necessity
of human-encoded features and instead incorporates the feature
engineering, feature selection, and model fitting into one step.

Deep learning is an extension of an idea originating in the

1950s called neural networks, which are loosely inspired by our
understanding of how neurons in our brains operate. Recent
hardware developments, originally designed for faster renderings
of graphics, birthed a renaissance in neural networks. The latest
graphical processing units, or GPUs, have more than 3,000 processing
cores that are well suited for parallel processing of complex matrix
manipulations required for rendering graphics – and for executing
computations on neural networks.

In the late 2000s, the combination of GPUs, advances in algorithms,

and collections of big data reignited interest in neural networks. GPUs
enabled computers to process much larger networks in much less time,
and clever advances in algorithms made the model fitting process more
efficient. Large collections of image, video, and text data provided
content for the deep learning algorithms to learn. The ability to train
larger networks with more data drove the exploration of new neural
network architectures featuring additional hidden layers and widening
the breadth of the networks.

Presently, deep learning has moved beyond academic applications and is

finding its way into our daily lives. Deep learning powers speech-to-text
on our mobile phones and smart devices, image search provided by major
tech companies, language translation services for text and spoken word,
and even drug discovery within advanced pharmaceutical companies.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› National Data Science Bowl
The first-ever National Data Science Bowl offered Data Scientists
a platform through which individuals could harness their passion,
unleash their curiosity and amplify their impact to affect change
on a global scale. The competition presented participants with
more than 100,000 underwater images provided by the Hatfield
Marine Science Center. Participants were challenged to develop a
classification algorithm that would enable researchers to monitor
Aaron Sander ocean health at a speed and scale never before possible.

More than 1,000 teams submitted a Deep Sea, Happy Lantern Festival,
total of approximately 15,000 solutions and Poisson Process, all used CNNs
over the 90 days of the competition. in their solutions. Their results
A large proportion of the participants’ increased algorithm accuracy by 10%
implemented solutions used deep over the state of the art. Without their
learning-based approaches, specifically algorithms, it would have taken marine
Convolutional Neural Nets (CNNs). researchers more than two lifetimes to
The competition forum exploded manually complete the classification
with competitors collectively sharing process. The work submitted by all
knowledge and collaborating to advance the participants represents major
the state-of-the-art in computer vision. advances for both the marine research
Participants tested new techniques and Data Science communities.
for developing CNNs and contributed
»Visit
» www.DataScienceBowl.com
to the development of open source to learn more about the first-ever
software for creating CNN models. National Data Science Bowl
The top three competitors, Team

OUTPUT
INPUT LAYER HIDDEN LAYER
LAYER

Source: Booz Allen Hamilton

A Representation of Deep Learning

Life in the Trenches 87

 Feature Engineering

Feature engineering is a lot like oxygen. You can’t do without

it, but you rarely give it much thought.

Feature engineering is the process by which one establishes the

representation of data in the context of an analytic approach. It is the
foundational skill in Data Science. Without feature engineering, it
would not be possible to understand and represent the world through
»What
» differentiates a great a mathematical model. Feature engineering is a challenging art. Like
model from a bad one? other arts, it is a creative process that manifests uniquely in each
Data Scientist. It will be influenced substantially by the scientist’s
In most cases, the inputs to the model experiences, tastes and understanding of the field.
matter even more than the choice
of the algorithm. The traditional When faced with the problem of feature engineering, there are
approach has a two-step process several paths that one may initially take. Generally speaking, better
where heuristics and subject-matter features can be developed with more knowledge of the domain. One
expertise are used to find a good set of approach to feature engineering is to begin by describing smaller
features, and then algorithms optimize
elements of the domain and continuously constructing more intricate
model parameters. Deep learning
features as the model becomes more complex. These more complicated
combines these steps into one.
features can be defined by considering other attributes of the domain,
Feature engineering, feature selection,
and model parameter estimation
aggregating features together into different groups or using advanced
are accomplished simultaneously. statistical and mathematical techniques to devise new features. The
This reduces the need for highly- ultimate judge in this process is the performance of the machine
specialized domain knowledge and learning algorithm which makes decisions based on this feature vector.
often results in better models.
Consider the example of email spam classification. Because the
For example, in the context of images domain is a set of emails, one possible choice of an initial feature
from the natural world, deep networks
vector is the integer array that counts the number of times a
may learn low-level features such as
given word appears in the email. This is called the "bag of words"
lines at various angles and curved
assumption, where the order that words appear in a text is ignored.
lines. Middle layers may combine the
lower-level features into more complex
If an algorithm with this feature vector does not adequately
geometric shapes and patterns. distinguish between spam and non-spam emails, a feature could
Higher-level layers combine the mid- be added that counts the number of misspelled words in the text.
level features into more complicated This new feature uses the spam recognition domain knowledge that
features that begin to resemble faces asserts many spam emails misspell words. These misspelled words
and shapes of animals. Applied to alert filters saying that the existence of certain words automatically
other types of data, such as audio label an email as spam. If this new feature is not enough, there are
and text, deep neural networks learn still additional features to be considered such as whether or not a
increasingly sophisticated features first or last name is used in the email.
at each layer in a similar manner.

Feature engineering represents a complex, but crucial aspect of Data

Science. The Learning Optimal Features sidebar goes into detail about
feature learning – an automated approach to feature engineering that
applies machine learning techniques.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› Chemoinformatic Search
On one assignment, my of the molecules were available. We
team was confronted created a model of how molecular
with the challenge of structural similarities might affect
developing a search their properties. We began by
engine over chemical
finding all the sub-graphs of each
compounds. The goal
molecule with length n, resulting
of chemoinformatic
search is to predict
in a representation similar to
Ed Kohlwey
the properties that the bag-of-words approach from
a molecule will exhibit as well as to natural language processing.
provide indices over those predicted We summarized each molecule
properties to facilitate data discovery fragment in a type of fingerprint
in chemistry-based research. These called a “Counting Bloom Filter.”
properties may either be discreet (e.g.,
“a molecule treats disease x well”) Next, we used several exemplars from
or continuous (e.g., “a molecule may the set to create new features. We
be dissolved up to 100.21 g/ml”). found the distance from each member
of the full training set to each of the
Molecules are complex 3D structures, exemplars. We fed these features into
which are typically represented as a non-linear regression algorithm
a list of atoms joined by chemical to yield a model that could be used
bonds of differing lengths with varying on data that was not in the original
electron domain and molecular training set. This approach can be
geometries. The structures are conceptualized as a “hidden manifold,”
specified by the 3-space coordinates whereby a hidden surface or shape
and the electrostatic potential defines how a molecule will exhibit a
surface of the atoms in the molecule. property. We approximate this shape
Searching this data is a daunting using a non-linear regression and a
task when one considers that set of data with known properties.
naïve approaches to the problem Once we have the approximate
bear significant semblance to the shape, we can use it to predict the
Graph Isomorphism Problem.[15] properties of new molecules.
The solution we developed was Our approach was multi-staged and
based on previous work in molecular complex – we generated sub-graphs,
fingerprinting (sometimes also called created bloom filters, calculated
hashing or locality sensitive hashing). distance metrics and fit a linear-
Fingerprinting is a dimensionality regression model. This example
reduction technique that dramatically provides an illustration of how many
reduces the problem space by stages may be involved in producing a
summarizing many features, often sophisticated feature representation.
with relatively little regard to the By creatively combining and building
importance of the feature. When “features on features,” we were able
an exact solution is likely to be to create new representations of data
infeasible, we often turn to heuristic that were richer and more descriptive,
approaches such as fingerprinting. yet were able to execute faster and
produce better results.
Our approach used a training set
where all the measured properties

Life in the Trenches 89

 Feature Selection

Models are like honored guests; you should only feed them the
good parts.

Feature selection is the process of determining the set of features with

the highest information value to the model. Two main approaches are
filtering and wrapper methods. Filtering methods analyze features
using a test statistic and eliminate redundant or non-informative
features. As an example, a filtering method could eliminate features
that have little correlation to the class labels. Wrapper methods
utilize a classification model as part of feature selection. A model is
trained on a set of features and the classification accuracy is used to
measure the information value of the feature set. One example is that
of training a neural network with a set of features and evaluating the
accuracy of the model. If the model scores highly on the test set, then
the features have high information value. All possible combinations of
features are tested to find the best feature set.

There are tradeoffs between these techniques. Filtering methods

are faster to compute since each feature only needs to be compared
against its class label. Wrapper methods, on the other hand, evaluate
feature sets by constructing models and measuring performance.
This requires a large number of models to be trained and evaluated
(a quantity that grows exponentially in the number of features).
Why would anyone use a wrapper method? Feature sets may perform
better than individual features.[16] With filter methods, a feature
with weak correlation to its class labels is eliminated. Some of
these eliminated features, however, may have performed well when
combined with other features.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› Cancer Cell Classification
On one project, was used as feedback into
our team was the Genetic Algorithm. When
challenged a set of features contained
to classify no useful information, the
cancer cell model performed poorly and
profiles. The a different feature set would
overarching be explored. Over time, this
Paul Yacci
goal was method selected a set of
to classify different types features that performed with
of Leukemia, based on high accuracy. The down-
Microarray profiles from 72 selected feature set increased
samples[17] using a small speed and performance as
set of features. We utilized well as allowed for better
a hybrid Artificial Neural insight into the factors that
Network (ANN)[18] and Genetic may govern the system. This
Algorithm[19] to identify subsets allowed our team to design
of 10 features selected from a diagnostic test for only a
thousands.[20] We trained the few genetic markers instead
ANN and tested performance of thousands, substantially
using cross-fold validation. reducing diagnostic test
The performance measure complexity and cost.

Life in the Trenches 91

 Ensemble Models

None of us is as smart as all of us, but some are smarter

than others.

In 1906, Sir Francis Galton attended a fair at which there was a

contest to guess the weight of an ox. Galton had the idea to collect the
guesses of the 787 entrants and compute the mean. To his surprise,
the mean was only one pound off from the ox’s real weight of 1,198
pounds. Together, the estimates made by many amateurs formed a
prediction that was more accurate than that of individual experts.

Galton’s “wisdom of crowds” extends to Data Science in the form of

ensemble learning, which is colloquially and somewhat equivalently
called ensembling, blending, or stacking. An ensemble takes the
predictions of many individual models and combines them to
make a single prediction. Like the people guessing the ox’s weight,
Data Science models have unique strengths and weaknesses (i.e.,
determined by their design), and are influenced by varied perspectives
based on past experience (i.e., the data they have observed).

Ensembles overcome individual weaknesses to make predictions with

more accuracy than their constituent models. These models need not
stem from different methodologies; an ensemble might employ the
same method with different parameters or weights (e.g., boosting),
feature subsets (e.g., random forests), or sub-samples of data (e.g.,
bagging). The ensembling methodology may be as simple as averaging
two outputs, or as complex as using a “meta model” to learn an
optimal combination.

An ensemble’s ability to reduce individual errors arises from the

diversity of its members. If one model over fits the data, it is balanced
by a different model that under fits the data. If one subset is skewed
by outlier values, another subset is included without them. If one
method is unstable to noisy inputs, it is bolstered by another method
that is more robust.

In practice, ensembling typically improves a model by a few percent.

The price of this accuracy is paid in complexity. The accuracy vs.
complexity tradeoff can make it difficult to know when ensembling
is justified. On one hand, ensembles appear to be a fit for high-stakes
problems—think detecting cancer in MRI images vs. detecting
unripe blueberries on a conveyer belt. On the other hand, high-stakes
problems mandate higher standards for auditing model functionality.

The Data Scientist must manage a balance between ensemble

interpretability and black-box complexity. If this seems easy, it isn’t!
Put yourself in the driver’s seat of the machine learning code for a
self-driving car. If a well-behaved regression model makes a right
decision 99.5% of the time, but a complex, less-explainable ensemble
is right 99.8% of the time, which would you pick?

T H E F I ELD G U I D E to D A T A S C I E N C E
 ›› Ensemble
The Value of
Models
Several years ago, the Kaggle Photo Quality Prediction
competition posed the question “Given anonymized information
on thousands of photo albums, predict whether a human
evaluator would mark them as 'good'.” Participants were
supplied a large collection of user-generated photos. The goal
was to create an algorithm that could automatically pick out
particularly enjoyable or impressive photos from the collection.
Will Cukierski

Over the course of the competition, 207 people submitted entries. The log likelihood metric was
used to evaluate the accuracy of the entries. Scores for the top 50 teams ranged from 0.18434 to
0.19884, where lower is better. Kaggle data scientist Ben Hamner used the results to illustrate the
value of ensembling by means of averaging the top 50 scores. The figure below shows the results.

Ensemble
Private Leaderboard
Log likelihood (lower is better)

Rank=1
Final Team Rank

Value of Ensembling for the Kaggle Photo Quality Prediction Competition

(results courtesy of Ben Hamner, Kaggle CTO)

The-blue line shows the individual scores all points across the top 50 teams. However,
for each of the top 50 teams. The orange after we increase the number of models in
line shows the ensembled score for the top the ensemble beyond 15, we begin to see
n teams, where n ranges from 1 to the value the ensembled score increase. This occurs
on the axis. For example, the ensemble point because we are introducing less accurate (i.e.,
for Final Team Rank 5 is an ensemble of the potentially overfit) models into the ensemble.
entries for teams 1 through 5. As shown in The results of this simple experiment quantify
the graph, the ensembled score is lower than the value of creating an ensemble model, while
any single individual score. The diversity of reinforcing the idea that we must be thoughtful
models included within the ensemble causes when selecting the individual models contained
the respective errors to cancel out, resulting within the ensemble.
in an overall lower score. This holds true for

Life in the Trenches 93

 Data Veracity

We’re Data Scientists, not data alchemists. We can’t make

analytic gold from the lead of data.

While most people associate data volume, velocity, and variety with
big data, there is an equally important yet often overlooked dimension
– data veracity. Data veracity refers to the overall quality and
correctness of the data. You must assess the truthfulness and accuracy
of the data as well as identify missing or incomplete information. As
the saying goes, “Garbage in, garbage out.” If your data is inaccurate or
missing information, you can’t hope to make analytic gold.

Assessing data truthfulness is often subjective. You must rely on your

experience and an understanding of the data origins and context.
Domain expertise is particularly critical for the latter. Although data
accuracy assessment may also be subjective, there are times when
quantitative methods may be used. You may be able to re-sample from
the population and conduct a statistical comparison against the stored
values, thereby providing measures of accuracy.

The most common issues you will encounter are missing or

incomplete information. There are two basic strategies for dealing
with missing values – deletion and imputation. In the former, entire
observations are excluded from analysis, reducing sample size and
potentially introducing bias. Imputation, or replacement of missing
or erroneous values, uses a variety of techniques such as random
»
»Tips From the Pros sampling (hot deck imputation) or replacement using the mean,
statistical distributions or models.
Find an approach that works,
implement it, and move on. You
can worry about optimization and
tuning your approaches later during
incremental improvement.

T H E F I ELD G U I D E to D A T A S C I E N C E
 ›› Time Series Modeling On one of our projects, the team was faced with correlating the time
series for various parameters. Our initial analysis revealed that the
correlations were almost non-existent. We examined the data and
quickly discovered data veracity issues. There were missing and
null values, as well as negative-value observations, an impossibility
given the context of the measurements (see the figure, Time Series
Data Prior to Cleansing). Garbage data meant garbage results.
Brian Keller

Source: Booz Allen Hamilton

Time Series Data Prior to Cleansing

Because sample size was already small, deleting distortion in the underlying signal, and we quickly
observations was undesirable. The volatile nature abandoned the approach.
of the time series meant that imputation through
One of our team members who had experience in
sampling could not be trusted to produce values
signal processing suggested a median filter. The
in which the team would be confident. As a result,
median filter is a windowing technique that moves
we quickly realized that the best strategy was an
through the data point-by-point, and replaces it
approach that could filter and correct the noise in
with the median value calculated for the current
the data.
window. We experimented with various window
We initially tried a simplistic approach in which we sizes to achieve an acceptable tradeoff between
replaced each observation with a moving average. smoothing noise and smoothing away signal. The
While this corrected some noise, including the figure, Time Series Data After Cleansing, shows the
outlier values in our moving-average computation same two time series after median filter imputation.
shifted the time series. This caused undesirable

Source: Booz Allen Hamilton

Time Series Data After Cleansing

The application of the median filter approach was By addressing our data veracity issues, we were
hugely successful. Visual inspection of the time able to create analytic gold. While other approaches
series plots reveals smoothing of the outliers may also have been effective, implementation speed
without dampening the naturally occurring peaks constraints prevented us from doing any further
and troughs (no signal loss). Prior to smoothing, analysis. We achieved the success we were after and
we saw no correlation in our data, but afterwards, moved on to address other aspects of the problem.
Spearman’s Rho was ~0.5 for almost all parameters.

Life in the Trenches 95

 Application of
Domain Knowledge

We are all special in our own way. Don’t discount what

you know.

Knowledge of the domain in which a problem lies is immensely

valuable and irreplaceable. It provides an in-depth understanding
of your data and the factors influencing your analytic goal. Many
times domain knowledge is a key differentiator to a Data Science
team’s success. Domain knowledge influences how we engineer and
select features, impute data, choose an algorithm, and determine
success. One person cannot possibly be a domain expert in every
field, however. We rely on our team, other analysts and domain
experts as well as consult research papers and publications to build
an understanding of the domain.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› Motor Vehicle TheftOn one project, Our team began by parsing and
our team explored verifying San Francisco crime data.
how Data Science We enriched stolen car reporting with
could be applied to general city data. After conducting
improve public safety. several data experiments across both
According to the space and time, three geospatial and
FBI, approximately one temporal hotspot emerged (see
Armen $8 Billion is lost figure, Geospatial and Temporal Car
Kherlopian annually due to Theft Hotspots). The domain expert
automobile theft. Recovery of the on the team was able to discern
one million vehicles stolen every that the primary geospatial hotspot
year in the U.S. is less than 60%. corresponded to an area surrounded
Dealing with these crimes represents by parks. The parks created an urban
a significant investment of law mountain with a number of over-foot
enforcement resources. We wanted access points that were conducive to
to see if we could identify how to car theft.
reduce auto theft while efficiently
using law enforcement resources.

GEOSPATIAL HOTSPOTS

TEMPORAL HOTSPOT
3am 6am 9am Noon 3pm 6pm 9pm
Sun

Sat

Fri

Thu

Wed

Tue

Mon
3am 6am 9am Noon 3pm 6pm 9pm
Source: Booz Allen Hamilton

Geospatial and Temporal Car Theft Hotspots

Our team used the temporal hotspot information in tandem with the insights
from the domain expert to develop a Monte Carlo model to predict the likelihood
of a motor vehicle theft at particular city intersections. By prioritizing the
intersections identified by the model, local governments would have the
information necessary to efficiently deploy their patrols. Motor vehicle thefts
could be reduced and law enforcement resources could be more efficiently
deployed. The analysis, enabled by domain expertise, yielded actionable insights
that could make the streets safer.

Life in the Trenches 97

 The Curse of
Dimensionality

There is no magical potion to cure the curse, but there is PCA.

The “curse of dimensionality” is one of the most important results

in machine learning. Most texts on machine learning mention this
phenomenon in the first chapter or two, but it often takes many years
of practice to understand its true implications.

Classification methods, like most machine learning methods, are

subject to the implications of the curse of dimensionality. The basic
intuition in this case is that as the number of data dimensions
increases, it becomes more difficult to create generalizable
classification models (models that apply well over phenomena not
observed in the training set). This difficulty is usually impossible to
overcome in real world settings. There are some exceptions in domains
where things happen to work out, but usually you must work to
minimize the number of dimensions. This requires a combination
of clever feature engineering and use of dimensionality reduction
techniques (see Feature Engineering and Feature Selection Life in the
Trenches). In our practical experience, the maximum
number of dimensions seems to be ~10 for linear model-based
approaches. The limit seems to be in the tens of thousands for more
sophisticated methods such as support vector machines, but the limit
still exists nonetheless.

A counterintuitive consequence of the curse of dimensionality is

that it limits the amount of data needed to train a classification
model. There are roughly two reasons for this phenomenon. In one
case, the dimensionality is small enough that the model can be
trained on a single machine. In the other case, the exponentially
expanding complexity of a high-dimensionality problem makes it
(practically) computationally impossible to train a model. In our
experience, it is quite rare for a problem to fall in a “sweet spot”
between these two extremes.

Rather than trying to create super-scalable algorithm

implementations, focus your attention on solving your immediate
problems with basic methods. Wait until you encounter a problem
where an algorithm fails to converge or provides poor cross-validated
results, and then seek new approaches. Only when you find that
alternate approaches don’t already exist, should you begin building
new implementations. The expected cost of this work pattern is lower
than over-engineering right out of the gate.

Put otherwise, “Keep it simple, stupid”.

T H E F I ELD G U I D E to D A T A S C I E N C E
›› Baking the Cake I was once given a time series set of roughly
1,600 predictor variables and 16 target variables
and asked to implement a number of modeling
techniques to predict the target variable
values. The client was challenged to handle the
complexity associated with the large number of
variables and needed help. Not only did I have
Stephanie a case of the curse, but the predictor variables
Rivera
were also quite diverse. At first glance, it looked
like trying to bake a cake with everything in the cupboard.
That is not a good way to bake or to make predictions!

The data diversity could be Using this approach, I was

partially explained by the fact able to condition upon week,
that the time series predictors without greatly increasing
did not all have the same the dimensionality. For the
periodicity. The target time other predictors, I used
series were all daily values a variety of techniques,
whereas the predictors were including projection and
daily, weekly, quarterly, and correlation, to make heads
monthly. This was tricky to or tails of the predictors. My
sort out, given that imputing approach successfully reduced
zeros isn’t likely to produce the number of variables,
good results. For this specific accomplishing the client’s goal
reason, I chose to use neural of making the problem space
networks for evaluating the tractable. As a result, the cake
weekly variable contributions. turned out just fine.

Life in the Trenches 99

 Model Validation

Repeating what you just heard does not mean that you
learned anything.

Model validation is central to construction of any model. This answers

the question “How well did my hypothesis fit the observed data?”
If we do not have enough data, our models cannot connect the dots.
On the other hand, given too much data the model cannot think
outside of the box. The model learns specific details about the training
data that do not generalize to the population. This is the problem of
model over fitting.

Many techniques exist to combat model over fitting. The simplest

method is to split your dataset into training, testing and validation
sets. The training data is used to construct the model. The model
constructed with the training data is then evaluated with the testing
data. The performance of the model against the testing set is used to
further reduce model error. This indirectly includes the testing data
within model construction, helping to reduce model over fit. Finally,
the model is evaluated on the validation data to assess how well the
model generalizes.

A few methods where the data is split into training and testing sets
include: k-fold cross-validation, Leave-One-Out cross-validation,
bootstrap methods, and resampling methods. Leave-One-Out cross-
validation can be used to get a sense of ideal model performance
over the training set. A sample is selected from the data to act as the
testing sample and the model is trained on the rest of the data. The
error on the test sample is calculated and saved, and the sample is
returned to the dataset. A different sample is then selected and the
process is repeated. This continues until all samples in the testing set
have been used. The average error over the testing examples gives a
measure of the model’s error.

There are other approaches for testing how well your hypothesis
reflects the data. Statistical methods such as calculating the coefficient
of determination, commonly called the R-squared value are used to
identify how much variation in the data your model explains. Note
that as the dimensionality of your feature space grows, the R-squared
»Do
» we really need a case study value also grows. An adjusted R-squared value compensates for this
to know that you should
phenomenon by including a penalty for model complexity. When
check your work?
testing the significance of the regression as a whole, the F-test
compares the explained variance to unexplained variance. A regression
result with a high F-statistic and an adjusted R-squared over 0.7 is
almost surely significant.

T H E F I ELD G U I D E to D A T A S C I E N C E
P U T T I N G it A L L T O G E T H E R
›› Streamlining Medication Review
Analytic Challenge

The U.S. Food and Drug Administration (FDA) is responsible for advancing public
health by supporting the delivery of new treatments to patients; assessing the safety,
efficacy and quality of regulated products; and conducting
research to drive medical innovation. Although the FDA
houses one of the world’s largest repositories of regulatory
and scientific data, reviewers are not able to easily leverage
data-driven approaches and analytics methods to extract »»Our Case Studies
information, detect signals and uncover trends to enhance
regulatory decision-making and protect public health. In Hey, we have given you a lot of
addition, a rapid increase in the volume, velocity and variety really good technical content. We
of data that must be analyzed to address and respond to know that this section has the look
and feel of marketing material,
regulatory challenges, combined with variances in data
but there is still a really good story
standards, formats, and quality, severely limit the ability
here. Remember, storytelling comes
of FDA Center for Drug Evaluation and Research (CDER) in many forms and styles, one of
regulatory scientists to conduct cross-study, cross-product, which is the marketing version. You
retrospective, and meta-analysis during product reviews. should read this chapter for what it
is – great information told with a
Booz Allen Hamilton was engaged to research, develop, marketing voice.
and evaluate emerging informatics tools, methods, and
techniques to determine their ability to address regulatory
challenges faced by the FDA Center for Drug Evaluation
and Research (CDER). The main goal was to enable the CDER community to fully
utilize the agency’s expansive data resources for efficient and effective drug review
through the design and development of informatics capabilities based on Natural
Language Processing (NLP), data integration, and data visualization methodologies.

Our Approach

To support transformational change datasets, and geographic visualization

at CDER, we designed and developed capabilities to support the inspection of
a set of informatics prototypes for pharmaceutical manufacturing facilities.
the analysis and modeling of complex
Product Safety Analytics. The review,
structured, unstructured, and
surveillance, and analysis of adverse
fragmented datasets. We developed
events throughout the product lifecycle
multiple prototypes to enable the
require significant resources. The ability
evaluation of emerging informatics
to identify actionable insights that lead
tools, methods, and techniques and
to informed decision-making requires
their ability to enable a critical-value
significant investment of effort, including
driver – e.g., co-locate complex,
the active and passive surveillance
heterogeneous data to identify patterns
of adverse events. To address these
and foster development of strategies
challenges, we developed a Product
to protect public health. For example,
Safety Dashboard that compares adverse
we implemented NLP algorithms
events listed in the product label (i.e.,
to compare adverse events across

T H E F I ELD G U I D E to D A T A S C I E N C E
package inserts) with data from the of the product quality-review process.
FDA Adverse Event Reporting System Integration of disparate data sources is
(FAERS). Using NLP, we extracted the first step in building a comprehensive
adverse events from the product label profile of manufacturers, facilities, and
to create a structured table of label the products associated with individual
data out of unstructured text. This facilities. To address these challenges,
dashboard allows safety evaluators to we developed a Facility Inventory Report
view whether or not a reported adverse to show the geographic location of
event is already known, without having facilities and their associated metadata.
to access an external data source This geovisualization tool processes and
and read through product labels. transforms raw data into a user-friendly
visual interface with mapping features
Product Quality Analytics.
to enhance the surveillance capabilities
To support CDER’s mission of reviewing
of CDER and provide reviewers with the
and managing product quality, novel
ability to establish connections between
methodologies and tools are needed
facility data and product quality.
to improve the efficiency and efficacy

Our Impact

Since the FDA is responsible for regulating 25 cents of every dollar that Americans
spend, the agency’s ability to fully use regulatory datasets and meaningfully integrate
previously incompatible data to rapidly detect product quality and safety issues is
critical for safeguarding public health. NLP approaches provide CDER with the ability
to more efficiently search a broader range of textual data and enhances the ability to
gain insight from additional data forms that may seem unrelated. Data integration
and visualization directly increase the efficiency of researchers by reducing their time
spent on searching for frequently-performed aggregate or granular calculations,
and by proactively presenting the most frequently desired data to the reviewer
through thoughtful and contextual dashboards designed to reveal patterns and
trends in disparate data sources. These new capabilities position the FDA to enhance
regulatory decision-making, drive advances in personalized medicine, and enable
earlier detection of safety signals in the general population.

Putting it all Together 105

›› Reducing Flight Delays
Analytic Challenge

Domestic airline departure delays are estimated to cost the U.S. economy $32.9 billion
annually. The Federal Aviation Administration’s (FAA’s) Traffic Flow Management
System (TFMS) is used to strategically manage flights and includes a flight departure-
delay prediction engine which applies simple heuristics to predict flight delays.
However, the limited predictive power of these heuristics constrains the FAA’s ability
to act in accordance with its existing departure-delay management plan. In response,
the FAA’s NextGen Advanced Concepts and Technology Development Group wanted to
create a predictive probabilistic model to improve aircraft departure time predictions.
This new model would help the FAA understand the causes of departure delays and
develop policies and actions to improve the reliability of departure time predictions for
real-time air traffic flow management.

Our Approach

The commercial aviation industry is rich departure deviation. The most critical
in flight operations data, much of which steps in model development were the
is publicly available through government selection of optimal algorithms to
websites and a few subscription discretize model variables, and the
vendors. Booz Allen Hamilton leveraged selection of appropriate machine learning
these sources to gather over 4 TB of techniques to learn the model from the
data detailing tarmac and airspace data. The team followed information
congestion, weather conditions, network theory principles to discretize model
effects, Traffic Management Initiatives, variables to maximize the model’s
and airline and aircraft-specific predictive power, and to represent the
attributes for every commercial flight data as closely as possible with the
departing from U.S. airports between least amount of network complexity.
2008 and 2012. This data included over 50 Booz Allen segmented the model
million flights and around 100 variables variables into three different categories
for each flight. The data included based on the time to flight departure:
composite variables (e.g. incoming flight 24 hours, 11 hours, and one hour.
delay) that were constructed from the Certain flight variables could only
raw data to capture relevant dynamics be known for specific pre-departure
of flight operations. Data acquisition, times. For example, the tarmac and
processing, quality control, and accuracy airspace congestion variables for
between disparate datasets were a flight are only known just before
important steps during this process. the flight, and hence those variables
The team applied supervised learning feature only in the one hour category.
algorithms to develop Bayesian Belief Departure delays were predicted for
Network (BBN) models to predict flight each of the three time horizons.

T H E F I ELD G U I D E to D A T A S C I E N C E
Our Impact

For a typical airport, the model delivers a delay prediction improvement of between
100,000 and 500,000 minutes annually over previous FAA predictions. The model can
be used by a range of aviation stakeholders, such as airlines, to better understand
and predict network flight delays. This can improve the airlines’ operational decisions
to include more proactive schedule adjustments during times of disruption (e.g.
weather or sector load). The improved reliability of departure prediction will improve
FAA's predictions for airports, sectors, and other resources, and has the potential to
enable improved real-time traffic flow management, which can significantly reduce
airline departure, delays, and the associated economic costs. This means a more
efficient and effective air transportation network.

Putting it all Together 107

›› Making Vaccines Safer
Analytic Challenge

The U.S. Food and Drug Administration (FDA) Center for Biologics Evaluation and
Research (CBER) is responsible for protecting public health by assuring the safety
and efficacy of biologics, including vaccines, blood and blood products. CBER’s
current surveillance process, which requires resource-intensive manual review by
expert Medical Officers, does not scale well to short-term workload variation and
limits long-term improvements in review cycle-time. In addition, the large volume of
Adverse Event (AE) reports received by the Agency makes it difficult for reviewers to
compare safety issues across products and patient populations.

CBER engaged Booz Allen Hamilton to develop advanced analytics approaches for
the triage and analysis of AE reports. The main goal was to leverage (1) Natural
Language Processing (NLP) to alleviate resource pressures by semi-automating
some of the manual review steps through techniques, such as text classification
and entity extraction, and (2) network visualizations to offer alternative interactions
with datasets and support AE pattern recognition. By integrating NLP and network
analysis capabilities into the Medical Officer’s review process, Booz Allen successfully
provided decision-makers with important information concerning product risks and
possible mitigations that can reduce risk.

Our Approach

We designed and developed a set extracting diagnosis, cause of death, and

of prototypes for the analysis and time to onset, and presenting relevant
visualization of complex structured information for review, the text-mining
and unstructured AE datasets. We tool streamlines and enhances the post
developed tools that leverage NLP market surveillance process.
and network analysis to extend and
Adverse Event Network Analytics.
enhance CBER’s efforts to monitor the
Visualizing relationships between
safety of biologics and manage safety
vaccines and AEs can reveal new
throughout the product lifecycle.
patterns and trends in the data, leading
Adverse Event Text Mining Analytics. reviewers to uncover safety issues.
Reviewing AE reports involves sorting To assist CBER Medical Officers and
based on the likelihood of a relationship researchers in identifying instances
between a product and reported adverse where certain vaccines or combinations
events. Since much of an AE report is of vaccines might have harmful
unstructured text, and most reports effects on patients, we developed an
are not directly related to the use of AE network analysis tool using open
the implicated biologic, manual review source tools. This network analyzer
is time consuming and inefficient. To allows users to select partitions of the
address these challenges, we enhanced FDA Vaccine Adverse Event Reporting
and integrated tools for text mining System (VAERS) database, generate a
and NLP of AE reports using open co-occurrence matrix, view networks
source tools, including Python and R. By and network metrics (e.g., betweenness,

T H E F I ELD G U I D E to D A T A S C I E N C E
closeness, degree, strength), and handle large amounts of data generated
interact with network nodes to gain by a Monte Carlo Markov Chain analysis
insights into product safety issues. of the spread of influenza, developed
a large database analysis strategy
Other Analytics Solutions. In addition,
involving the application of classification
Booz Allen refactored, modularized,
algorithms to simulated genomic data,
and expended the capabilities of CBER’s
and implemented a Statistical Analysis
computer simulation model of the Bovine
Software (SAS) macro that automatically
Spongiform Encephalopathy (BSE)
compares the relative potency of a given
agent to improve estimates of variant
lot of vaccine using a matched set of dose
Creutzfeldt-Jakob disease (vCJD) risk
response curves.
for blood products, developed code to

Our Impact

New methods for post market surveillance of biologics are critical for FDA reviewers
who must determine whether reported adverse events are actually a result of a
biologic product. With more than 10 million vaccines administered each year to
children less than one-year old, CBER reviewers are under pressure to quickly
evaluate potential safety signals through manual evaluation of AE reports, review
of scientific literature, and analysis of cumulative data using frequency calculations
or statistical algorithms. Booz Allen’s support resulted in the development
of innovative and data-driven approaches for the analysis of structured and
unstructured AE reports. We increased the speed of existing text mining tools by
two thousand times, allowing CBER reviewers to run a text mining algorithm to
extract information contained in VAERS reports in seconds, instead of hours. We
also increased the productivity of Medical Officers through the implementation
of text mining and network analysis tools. These new capabilities allow CBER to
streamline the post market review process, extract knowledge from scientific data,
and address public concerns regarding vaccine safety more quickly and efficiently.

Putting it all Together 109

›› Forecasting the Relative Risk
for the Onset of Mass Killings to
Help Prevent Future Atrocities
Analytic Challenge

Mass atrocities are rare yet devastating crimes. They are also preventable. Studies
of past atrocities show that we can detect early warning signs of atrocities and that
if policy makers act on those warnings and develop preventive strategies, we can
save lives. Yet despite this awareness, all too often we see warning signs missed and
action taken too late, if at all, in response to threats of mass atrocities.

The Early Warning Project, an initiative of the United States Holocaust Memorial
Museum (Holocaust Museum), aims to assess a country’s level of risk for the onset
of future mass killings. Over time, the hope is to learn which models and which
indicators are the best at helping anticipate future atrocities to aid in the design and
implementation of more targeted and effective preventive strategies. By seeking to
understand why and how each countries’ relative level of risk rises and falls over
time, the system will deepen understanding of where new policies and resources can
help make a difference in averting atrocities and what strategies are most effective.
This will arm governments, advocacy groups, and at-risks societies with earlier and
more reliable warning, and thus more opportunity to take action, well before mass
killings occur.

The project’s statistical risk assessment seeks to build statistical and machine
learning algorithms to predict the onset of a mass killing in the succeeding 12 months
for each country with a population larger than 500,000. The publically available system
aggregates and provides access to open source datasets as well as democratizes the
source code for analytic approaches developed by the Holocaust Museum staff and
consultants, the research community, and the general public. The Holocaust Museum
engaged Booz Allen to validate existing approaches as well as explore new and
innovative approaches for the statistical risk assessment.

Our Approach

Taking into account the power of identifying new datasets, building new
crowdsourcing, Booz Allen put out a call machine learning models, and creating
to employees to participate in a hack-a- frameworks for ensemble modeling
thon—just the start of the team’s support and interactive results visualization.
as the Museum refined and implemented Following the hack-a-thon, Booz
the recommendations. More than 80 Allen Data Scientists worked with
Booz Allen Hamilton software engineers, Holocaust Museum staff to create
data analysts, and social scientists a data management framework to
devoted a Saturday to participate. automate the download, aggregation,
Interdisciplinary teams spent 12 hours and transformation of the open

T H E F I ELD G U I D E to D A T A S C I E N C E
source datasets used by the statistical supporting greater engagement by
assessment. This extensible the Data Science community.
framework allows integration of new
datasets with minimal effort, thereby

Our Impact

Publically launched in the fall of 2015, the Early Warning Project can now leverage
advanced quantitative and qualitative analyses to provide governments, advocacy
groups and at-risk societies with assessments regarding the potential for mass
atrocities around the world. Integration of the project’s statistical risk assessment
models and expert opinion pool created a publicly available source of invaluable
information and positioned Data Science at the center of global diplomacy.

The machine learning models developed during the hack-a-thon achieved

performance on par with state of the art approaches as well as demonstrated the
efficacy of predictions 2-5 years into the future. Teams also identified approaches for
constructing test/validation sets that support more robust model evaluation. These
risk assessments are an important technological achievement in and of themselves,
but what this initiative means for the Data Science community’s position in global
diplomatic dialogue marks an entirely new era for those on the frontiers of Big Data.

The data management framework developed from the lessons learned of the hack-
a-thon represents a great leap forward for the Holocaust Museum. The periodicity
of aggregating and transforming data was reduced from twice per year to once
per week. In addition to providing the community with more up-to-date data, the
reduced burden on researchers enables them to spend more time analyzing data and
identifying new and emergent trends. The extensible framework will also allow the
Holocaust Museum to seamlessly integrate new datasets as they become available or
are identified by the community as holding analytic value for the problem at hand.

Through this project, the Holocaust Museum was able to shift the dynamic from
monitoring ongoing violence to determining where it is likely to occur 12 to 24 months
into the future by integrating advanced quantitative and qualitative analyses to assess
the potential for mass atrocities around the world. The Early Warning Project is an
invaluable predictive resource supporting the global diplomatic dialogue. While the
focus of this effort was on the machine learning and data management technologies
behind the initiative, it demonstrates the growing role the Data Science community is
playing at the center of global diplomatic discussions.

Putting it all Together 111

›› Predicting Customer Response
Analytic Challenge

It is very challenging to know how a customer will respond to a given promotional

campaign. Together with the InterContinental Hotels Group (IHG), Booz Allen
Hamilton explored approaches to predict customer-by-customer response to a state-
of-the-art promotional campaign in order to better understand and increase return on
investment (ROI).

In the past, conventional statistics have worked well for analyzing the impact of
direct marketing promotions on purchase behavior. Today, modern multi-channel
promotions often result in datasets that are highly dimensional and sometimes
sparse, which strains the power of conventional statistical methods to accurately
estimate the effect of a promotion on individual purchase decisions. Because of
the growing frequency of multi-channel promotions, IHG was driven to investigate
new approaches. In particular, IHG and Booz Allen studied one recent promotional
campaign using hotel, stay, and guest data for a group of loyalty program customers.

Our Approach

Working closely with IHG experts, Booz Allen investigated three key elements related
to different stages of analytic maturity:

Describe: Using initial data mining, what a probabilistic Bayesian Belief Network
insights or tendencies in guest behaviors (BBN) can learn the pairwise relationships
can be identified after joining multiple, between all individual customer attributes
disparate datasets? and their impact on promotional
return, Booz Allen investigated how
Discover: Can we determine which
this technique could be used to model
control group members would be likely to
each treated customer without an exact
register for a promotion if offered? If so,
controlled look-alike.
can we also quantify their registration?

Predict: How would a hotel guest that Specifically, Booz Allen developed a BBN
received the promotion have responded to predict customer-by-customer impacts
if they were not offered the promotion? driven by promotional campaign offers,
How would a hotel guest that did not subsequently estimating the aggregated
receive the promotion have responded ROI of individual campaigns. We used six
if they were offered the promotion? machine learning techniques (support
vector machine, random forest, decision
For the promotion that was the focus tree, neural network, linear model, and
of this case study, not everything about AdaBoost) in unison with the BBN to
customers could be controlled as required predict how each customer would be
by traditional statistics. However, because influenced by a promotional offer.

T H E F I ELD G U I D E to D A T A S C I E N C E
Our Impact

The probabilistic model was capable of predicting customer response to the

promotion, without relying on a costly control group. This finding represents millions
of dollars of savings per promotional campaign. This analysis is an industry-
first for the travel and hospitality sector, where demand for more data-driven
approaches to optimize marketing ROI at an individual level is rapidly growing.

Because Booz Allen and IHG’s approach enabled estimation of ROI for each
hypothetical customer, even when no exact look-alikes exist, there are a number
of valuable future applications. One such application is optimal campaign design—
the ability to estimate the promotional attributes for an individual customer
that are likely to drive the greatest incremental spend. Another application is
efficient audience selection - which would reduce the risk of marketing “spam”
that prompts costly unsubscriptions and can negatively impact a hotel’s brand.

Putting it all Together 113

CLOSING T I M E
 ›› THE F U TURE of
DATA S C I E N C E
Data Science is rapidly evolving as it touches every aspect of our
lives on a daily basis. As Data Science changes the way we interact
with, and explore our world, the algorithms and applications of
Data Science continue to advance. We expect this trend to continue
as Data Science has an increasingly profound effect on humanity.
We describe here some of the trends and developments we
anticipate emerging in the field of Data Science over the coming years.
Kirk Borne
The advancements in some Data Science algorithms will deliberately
track the evolution of data structures and data models that Data
Scientists are using to represent their domains of study. One of the
clearest examples of this linkage is in the development of massive-
scale graph analytics algorithms, which are deployed on graph
databases (including network data and semantically linked databases).
It is sometimes said “all the world is a graph,” and consequently the
most natural data structure is not a table with rows and columns, but
a network graph with nodes and edges. Graph analytics encompasses
traditional methods of machine learning, but with a graph-data twist.

Another growth area in Data Science algorithms is in the domain

of geospatial temporal predictive analytics, which can be applied
to any dataset that involves geospatial location and time – that
describes just about everything in our lives! We expect increasingly
sophisticated deployments of this methodology in the areas of law
enforcement, climate change, disaster management, population
health, sociopolitical change, and more.

It is obvious that bigger, faster, and more complex datasets will

require faster (hyperfast!) analytics. We anticipate advanced
Data Science algorithms that take advantage of technological
advancements in quantum machine-learning, in-memory data
operations, and machine learning on specialized devices (e.g.,
the GPU, Raspberry Pi, or the next-generation mobile handheld
“supercomputer”). In such commodity devices, we expect to see
development of more embedded machine learning (specifically, deep
learning) algorithms that perform time-critical data-to-insights
transformations at the point of data collection. Such use cases will
be in great abundance within the emerging Internet of Things (IoT),
including the industrial IoT and the internet of everything.

Advances in cognitive machine learning are on the horizon,

including open source and configurable algorithms that exploit
streaming real-time data’s full content, context, and semantic
meaning. The ability to use the 360-degree view of a situation will
enable the delivery of the right action, at the right time, at the right
place, in the right context – this is the essence of cognitive analytics.
Another way to view cognitive analytics is that, given all of the
data and the context for a given object or population, the algorithm
identifies the right question that you should be asking of your data
(which might not be the question that you traditionally asked).
T H E F I ELD G U I D E to D A T A S C I E N C E
 Another area of Data Science evolution that tracks with the growth
in a particular data type is that of unstructured data, specifically text.
The growth of such unstructured data is phenomenal, and demands
richer algorithms than those used on numerical data, since there are
many more shades of meaning in natural language than in tables of
numbers. The new Data Science algorithms for unstructured data
will be applied in multiple directions. Natural Language Generation
will be used to convert data points into text, which can be used
to generate the data’s story automatically. Structured Database
Generation will transform text documents or other unstructured
data into data points (i.e., converting qualitative data into machine-
computable quantitative data).

All of these technical advancements, plus others that we cannot

yet imagine, will be brought to bear on new domains. Some of the
hottest, most critical domains in which Data Science will be applied
in the coming years include:
›› Cybersecurity including advanced detection, modeling, prediction, and
prescriptive analytics

›› Healthcare including genomics, precision medicine, population health, healthcare

delivery, health data sharing and integration, health record mining, and wearable
device analytics

›› IoT including sensor analytics, smart data, and emergent discovery alerting and response
›› Customer Engagement and Experience including 360-degree view, gamification,
and just-in-time personalization

›› Smart X, where X = cities, highways, cars, delivery systems, supply chain, and more
›› Precision Y, where Y = medicine, farming, harvesting, manufacturing, pricing, and more
›› Personalized Z, where Z = marketing, advertising, healthcare, learning, and more
›› Human capital (talent) and organizational analytics
›› Societal good

Algorithms
THE FUTURE OF Applications
DATA SCIENCE

› Massive-scale Graph › Cybersecurity

› Geospatial Temporal Predictive Analytics › Healthcare
› Hyperfast Analytics › Internet of Things
› Embedded Deep Learning › Customer Engagement & Experience
› Cognitive Machine Learning › Smart Everything
› Natural Language Generation › Human Capital
› Structured Database Generation › Data for Societal Good

Closing Time 117

›› P A R T I N G
THOUGHTS
Data Science capabilities are creating data analytics that are
touching every aspect of our lives on a daily basis. From visiting
the doctor, to driving our cars, to shopping for services Data
Science is quietly changing the way we interactive with and
explore our world. We hope we have helped you truly understand
the potential of your data and how to become extraordinary
thinkers by asking the right questions. We hope we have helped
continue to drive forward the science and art of Data Science.
Most importantly, we hope you are leaving with a newfound
passion and excitement for Data Science.

Thank you for taking this journey with us. Please join our
conversation and let your voice be heard. Email us your ideas
and perspectives at [email protected] or submit them
via a pull request on the Github repository.

Tell us and the world what you know. Join us. Become an
author of this story.

Closing Time 119

 ›› R E F E R E N C E S
1. Commonly attributed to: Nye, Bill. Reddit Ask Me Anything (AMA).
July 2012. Web. Accessed 15 October 2013. SSRN: <http://www.reddit.
com/r/IAmA/comments/x9pq0/iam_bill_nye_the_science_guy_ama>

2. Fayyad, Usama, Gregory Piatetsky-Shapiro, and Padhraic Smyth. “From

Data Mining to Knowledge Discovery in Databases.” AI Magazine 17.3
(1996): 37-54. Print.

3. “Mining Data for Nuggets of Knowledge.” Knowledge@Wharton,1999.

Web. Accessed 16 October 2013. SSRN: <http://knowledge.wharton.
upenn.edu/article/mining-data-for-nuggets-of-knowledge>

4. Cleveland, William S. “Data Science: An Action Plan for Expanding

the Technical Areas of the Field of Statistics.” International Statistical
Review 69.1 (2001): 21-26. Print.

5. Davenport, Thomas H., and D.J. Patil. “Data Scientist: The Sexiest Job
of the 21st Century.” Harvard Business Review 90.10 (October 2012):
70–76. Print.

6. Smith, David. “Statistics vs Data Science vs BI.” Revolutions, 15

May 2013. Web. Accessed 15 October 2013. SSRN:<http://blog.
revolutionanalytics.com/2013/05/statistics-vs-data-science-vs-bi.html>

7. Brynjolfsson, Erik, Lorin M. Hitt, and Heekyung H. Kim. “Strength

in Numbers: How Does Data-Driven Decision Making Affect Firm
Performance?” Social Science Electronic Publishing, 22 April 2011. Web.
Accessed 15 October 2013. SSRN: <http://ssrn.com/abstract=1819486
or http://dx.doi.org/10.2139/ssrn.1819486>

8. “The Stages of an Analytic Enterprise.” Nucleus Research. February 2012.

Whitepaper.

9. Barua, Anitesh, Deepa Mani, and Rajiv Mukherjee. “Measuring the

Business Impacts of Effective Data.” University of Texas. Web. Accessed
15 October 2013. SSRNL: <http://www.sybase.com/files/White_Papers/
EffectiveDataStudyPt1-MeasuringtheBusinessImpactsofEffectiveDa
ta-WP.pdf>

10. Zikopoulos, Paul, Dirk deRoos, Kirshnan Parasuraman, Thomas Deutsch,

David Corrigan and James Giles. Harness the Power of Big Data: The IBM
Big Data Platform. New York: McGraw Hill, 2013. Print. 281pp.

T H E F I ELD G U I D E to D A T A S C I E N C E
11. Booz Allen Hamilton. Cloud Analytics Playbook. 2013. Web. Accessed 15
October 2013. SSRN: <http://www.boozallen.com/media/file/Cloud-
playbook-digital.pdf>

12. Conway, Drew. “The Data Science Venn Diagram.” March 2013.
Web. Accessed 15 October 2013. SSRN: <http://drewconway.com/
zia/2013/3/26/the-data-science-venn-diagram>

13. Booz Allen Hamilton. Tips for Building a Data Science Capability. 2015.
Web Accessed 2 September 2015. SSRN: < https://www.boozallen.com/
content/dam/boozallen/documents/2015/07/DS-Capability-Handbook.pdf>

14. Mnih et al. 2015. Human-level control through deep reinforcement

learning. Nature. 518: 529-533.

15. Torán, Jacobo. “On the Hardness of Graph Isomorphism.” SIAM Journal
on Computing. 33.5 (2004): 1093-1108. Print.

16. Guyon, Isabelle and Andre Elisseeff. “An Introduction to Variable and
Feature Selection.” Journal of Machine Learning Research 3 (March
2003):1157-1182. Print.

17. Golub T., D. Slonim, P. Tamayo, C. Huard, M. Gaasenbeek, J. Mesirov,

H. Coller, M. Loh, J. Downing, M. Caligiuri, C. Bloomfield, and E.
Lander. “Molecular Classification of Cancer: Class Discovery and Class
Prediction by Gene Expression Monitoring.” Science. 286.5439 (1999):
531-537. Print.

18. Haykin, Simon O. Neural Networks and Learning Machines. New Jersey:
Prentice Hall, 2008. Print.

19. De Jong, Kenneth A. Evolutionary Computation - A Unified Approach.

Massachusetts: MIT Press, 2002. Print.

20. Yacci, Paul, Anne Haake, and Roger Gaborski. “Feature Selection of
Microarray Data Using Genetic Algorithms and Artificial Neural
Networks.” ANNIE 2009. St Louis, MO. 2-4 November 2009.
Conference Presentation.

Closing Time 121

›› About
B O O Z A LLEN
H A M I L T O N
Booz Allen Hamilton has been at the forefront of strategy and
technology for more than 100 years. Today, the firm provides
management and technology consulting and engineering services to
leading Fortune 500 corporations, governments, and not-for-profits
across the globe. Booz Allen partners with public and private sector
clients to solve their most difficult challenges through a combination
of consulting, analytics, mission operations, technology, systems
delivery, cybersecurity, engineering, and innovation expertise.

With international headquarters in McLean, Virginia, the firm

employs more than 22,500 people globally, and had revenue of $5.27
billion for the 12 months ended March 31, 2015. To learn more, visit
www.boozallen.com. (NYSE: BAH)

Closing Time 123