Analytics May/June 2016 by Lionheart Publishing Inc.

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

DRIVING BETTER BUSINESS DECISIONS

M AY / J UNE 2016 BROUGHT TO YOU BY:

MANKIND & MACHINES • It takes both to get most from big data & analytics • Will computers replace knowledge workers?

ALSO INSIDE: • Self-service analytics: requisite skills • Making wise decisions with the IoT • Cloud-supported machine learning • Corporate profile: Steelcase Inc.

Executive Edge Mu Sigma strategy guru Tom Pohlmann on analytics job one: closing the data-todecisions gap

INS IDE STO RY

Humans vs. machines Can data scientists (the darlings of the Analytics Era) and other knowledge workers who are on almost every company’s wish list in terms of employment really be an endangered species and in danger of going the way of the dodo bird and losing their jobs to robots, machine learning and cognitive computing? That’s the question “Competing on Analytics” author Tom Davenport and Harvard University Press Senior Editor Julia Kirby addressed in a panel discussion at the recent INFORMS Conference on Business Analytics & Operations Research in Orlando, Fla. Their answer, with a few caveats: not in my lifetime. Whew! (Those of a more tender age are on their own as far as what the more distant future may hold.) The packed Davenport/Kirby session included Vijay Mehrotra, who explores the provocative topic in more detail in his “Analyze This!” column. Humans vs. machines is an old tale, popularized by the legendary John Henry, whose prowess as a “steel-driving man” was measured in a race against a steam-powered hammer. According to the legend, John Henry won the race, but collapsed and died just after the finish.

A N A LY T I C S - M A G A Z I N E . O R G

Speaking of humans vs. machines, Chris Mazzei, global chief analytics officer at Ernst & Young, says that the key to fulfilling the promise of analytics and getting the most from big data and analytics is focusing on the human element. Writes Mazzei: “For many companies, the people- and processrelated change management issues have prevented analytics from fully delivering on its potential.” For the rest of the story, click here. Meanwhile, with apologies to Tina Turner of “Proud Mary” fame, the big data wheel keeps on turning, cranking out unfathomable amounts of data. Several authors address that sticky point in this issue, including Tom Pohlmann, head of Values & Strategy at Mu Sigma, who argues that the No. 1 job of analytics is to close the data-to-decisions gap. Along those same lines, Sean Martin, founder and chief technology officer of Cambridge Semanitics, outlines the requisite business skills needed to get the most out of data lakes. Smart machines will no doubt play an ever-increasing role going forward, but humans will have their finger on the button that means the most. ❙

W W W. I N F O R M S . O R G

C O N T E N T S

DRIVING BETTER BUSINESS DECISIONS

MAY/JUNE 2016 Brought to you by

FEATURES 36

FULFILLING THE PROMISE OF ANALYTICS By Chris Mazzei Strategy, leadership and consumption: The keys to getting the most from big data and analytics focus on the human element.

HOW TO GET THE MOST OUT OF DATA LAKES By Sean Martin A handful of requisite business skills that facilitate self-service analytics at unparalleled speed.

MAKING WISE DECISIONS WITH THE IOT By Tayfun Keskin and Haluk Demirkan Consumer fantasy comes true with the Internet of Things, but are organizations ready for even bigger, wider, deeper data?

CLOUD-SUPPORTED MACHINE LEARNING SERVICES By Lakshmi D. Baskar, Neil Lobo, Praveen Ananth and Palaniappa Krishnan Exploring and comparing the potential of big data analytics on selected cloud providersâ&#x20AC;&#x2122; platforms.

CORPORATE PROFILE: STEELCASE INC. By Tim Merkle Advanced analytics team helps global company unlock human promise by creating great work experiences, wherever work happens.

68 4

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

XLMINER®: Data Mining Everywhere Predictive Analytics in Excel, Your Browser, Your Own App

XLMiner® in Excel – part of Analytic Solver® Platform – is the most popular desktop tool for business analysts who want to apply data mining and predictive analytics. And now it’s available on the Web, and in SDK (Software Development Kit) form for your own applications.

Forecasting, Data Mining, Text Mining in Excel. XLMiner does it all: Text processing, latent semantic analysis, feature selection, principal components and clustering; exponential smoothing and ARIMA for forecasting; multiple regression, k-nearest neighbors, and ensembles of regression trees and neural networks for prediction; discriminant analysis, logistic regression, naïve Bayes, k-nearest neighbors, and ensembles of classification trees and neural nets for classification; and association rules for affinity analysis.

have in Excel, and generate the same reports, displayed in your browser or downloaded for local use.

XLMiner SDK: Predictive Analytics in Your App. Access all of XLMiner’s parallelized forecasting, data mining, and text mining power in your own application written in C++, C#, Java, R or Python. Use a powerful object API to create and manipulate DataFrames, and combine data wrangling, training a model, and scoring new data in a single operation “pipeline”.

Find Out More, Start Your Free Trial Now. Visit www.solver.com to learn more, register and download Analytic Solver Platform or XLMiner SDK. And visit www.xlminer.com to learn more and register for a free trial subscription – or email or call us today.

XLMiner.com: Data Mining in Your Web Browser. Use a PC, Mac, or tablet and a browser to access all the forecasting, data mining, and text mining power of XLMiner in the cloud. Upload files or access datasets already online. Use the same Ribbon and dialogs you

The Leader in Analytics for Spreadsheets and the Web Tel 775 831 0300 • info@solver.com • www.solver.com

DRIVING BETTER BUSINESS DECISIONS

DEPARTMENTS

2 8 16 20 24 28 74 76 80

Inside Story Executive Edge Analyze This! Healthcare Analytics INFORMS Initiatives Newsmakers Conference Preview Five-Minute Analyst Thinking Analytically

President Edward H. Kaplan, Yale University President-Elect Brian Denton, University of Michigan Past President L. Robin Keller, University of California, Irvine Secretary Pinar Keskinocak, Georgia Tech Treasurer Sheldon N. Jacobson, University of Illinois Vice President-Meetings Ronald G. Askin, Arizona State University Vice President-Publications Jonathan F. Bard, University of Texas at Austin Vice President Sections and Societies Esma Gel, Arizona State University Vice President Information Technology Marco Lübbecke, RWTH Aachen University Vice President-Practice Activities Jonathan Owen, CAP, General Motors Vice President-International Activities Grace Lin, Institute for Information Industry Vice President-Membership and Professional Recognition Susan E. Martonosi, Harvey Mudd College Vice President-Education Jill Hardin Wilson, Northwestern University Vice President-Marketing, Communications and Outreach Laura Albert McLay, University of Wisconsin-Madison Vice President-Chapters/Fora Michael Johnson, University of Massachusetts-Boston INFORMS OFFICES www.informs.org • Tel: 1-800-4INFORMS Executive Director Melissa Moore Meetings Director Laura Payne Director, Public Relations & Marketing Jeffery M. Cohen Headquarters INFORMS (Maryland) 5521 Research Park Drive, Suite 200 Catonsville, MD 21228 Tel.: 443.757.3500 E-mail: informs@informs.org ANALYTICS EDITORIAL AND ADVERTISING

Analytics (ISSN 1938-1697) is published six times a year by the Institute for Operations Research and the Management Sciences (INFORMS), the largest membership society in the world dedicated to the analytics profession. For a free subscription, register at http://analytics.informs.org. Address other correspondence to the editor, Peter Horner, peter.horner@mail.informs.org. The opinions expressed in Analytics are those of the authors, and do not necessarily reflect the opinions of INFORMS, its officers, Lionheart Publishing Inc. or the editorial staff of Analytics. Analytics copyright ©2016 by the Institute for Operations Research and the Management Sciences. All rights reserved.

A N A LY T I C S - M AGA Z I N E . O RG

Lionheart Publishing Inc., 506 Roswell Street, Suite 220, Marietta, GA 30060 USA Tel.: 770.431.0867 • Fax: 770.432.6969

President & Advertising Sales John Llewellyn john.llewellyn@mail.informs.org Tel.: 770.431.0867, ext. 209 Editor Peter R. Horner peter.horner@mail.informs.org Tel.: 770.587.3172 Assistant Editor Donna Brooks donna.brooks@mail.informs.org Art Director Alan Brubaker alan.brubaker@mail.informs.org Tel.: 770.431.0867, ext. 218 Advertising Sales Aileen Kronke aileen@lionhrtpub.com Tel.: 770.431.0867, ext. 212

Su pp ac Pow orts he e T Sp r B abl ar I a ea k n u ExcelBig Dd , at a

ANALYTIC SOLVER PLATFORM ®

From Solver to Full-Power Business Analytics in

Solve Models in Desktop Excel or Excel Online.

Plus Forecasting, Data Mining, Text Mining.

From the developers of the Excel Solver, Analytic Solver Platform makes the world’s best optimization software accessible in Excel. Solve your existing models faster, scale up to large size, and solve new kinds of problems. Easily publish models from Excel to share on the Web.

Analytic Solver Platform samples data from Excel, PowerPivot, and SQL databases for forecasting, data mining and text mining, from time series methods to classification and regression trees and neural networks. And you can use visual data exploration, cluster analysis and mining on your Monte Carlo simulation results.

Conventional and Stochastic Optimization. Fast linear, quadratic and mixed-integer programming is just the starting point in Analytic Solver Platform. Conic, nonlinear, non-smooth and global optimization are just the next step. Easily incorporate uncertainty and solve with simulation optimization, stochastic programming, and robust optimization – all at your fingertips.

Find Out More, Download Your Free Trial Now. Analytic Solver Platform comes with Wizards, Help, User Guides, 90 examples, and unique Active Support that brings live assistance to you right inside Microsoft Excel. Visit www.solver.com to learn more, register and download a free trial – or email or call us today.

Fast Monte Carlo Simulation and Decision Trees. Analytic Solver Platform is also a full-power tool for Monte Carlo simulation and decision analysis, with 50 distributions, 40 statistics, Six Sigma metrics and risk measures, and a wide array of charts and graphs.

The Leader in Analytics for Spreadsheets and the Web Tel 775 831 0300 • info@solver.com • www.solver.com

EXE CU TIVE E D G E

Analytics job one: Closing the data-todecisions gap Data can no longer trump decisions, especially if you coordinate your company’s data and analytics efforts.

BY TOM POHLMANN

Are you marketing to the right customer? Is your product priced correctly? Have you found a more preemptive way to detect fraud? Everywhere you turn, data and analytics sit at the heart of answering such common but still vexing problems. But it’s the word data that seems to garner most of the attention, in part because of anecdotes like these: Experts predict that there will be a 4,300 percent increase in annual data generation by 2020, representing close to 40 zettabytes of data, which IDC found to be the equivalent of 57 times greater than all the grains of sand on all the earth’s beaches [1, 2]. When my firm conducts decision science workshops, in the lead-up to each event we survey participants and ask a simple open-ended question: “What’s your top challenge with respect to data and analytics?” Two types of responses emerge: concerns over data and concerns over analysis and decision-making, but with the former outweighing the latter by more than 2 to 1.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Your Analytics App – Everywhere

Use Solver, Risk Solver, XLMiner in Excel Online, Google Sheets Or Turn YOUR Excel Model into a Web or Mobile App in Seconds

The easiest way to build an analytic model – in Excel – is now the easiest way to deploy your analytic application to Web browsers and mobile devices – thanks to the magic of Frontline Solvers® and our RASON® server.

Use our Analytics Tools in your Web Browser. Solve linear, integer and nonlinear optimization models with Frontline’s free Solver, and run Monte Carlo simulation models with our free Risk Solver® tool, in Excel Online and Google Sheets. Use our free XLMiner® Analysis ToolPak tool for statistical analysis, matching the familiar Analysis ToolPak in desktop Excel.

Build Your Own Apps with RASON Software. RASON – RESTful Analytic Solver® Object Notation – is a new modeling language for optimization and simulation that’s embedded in JSON (JavaScript Object Notation). With support for linear, nonlinear and stochastic optimization, array and vector-matrix operations, and dimensional tables linked to external databases, the RASON language gives you all the power you need.

Your Excel Model Can Be a Web/Mobile App. The magic begins in Excel with Frontline Solvers V2016: Our Create App button converts your Excel optimization or simulation model to a RASON model, embedded in a Web page, that accesses our cloud servers via a simple REST API. You’re ready to run analytics in a browser or mobile device! Or if you prefer, run your RASON model on your desktop or server, with our Solver SDK®. Either way, you’re light-years ahead of other software tools.

Find Out More, Sign Up for a Free Trial Now. Visit www.solver.com/apps to learn more, and visit rason.com to sign up for a free trial of RASON and our REST API. Or email or call us today.

The Leader in Analytics for Spreadsheets and the Web Tel 775 831 0300 • info@solver.com • www.solver.com

EXE CU TIVE E D G E

Your background, be it from IT, marketing or a new analytics team, will affect your priorities. But across these disciplines and across most companies where we’ve worked, I see so many who believe that data must be completely buttoned up before it can be used to make informed decisions. These days that mindset can lead to costly delays and misallocated investments. We get it. Data has exploded, and companies must still dedicate a lot of resources to managing it. But data can no longer trump decisions. You need to treat them as equals. Especially if you play a role in coordinating your company’s data and analytics efforts, consider doing the following: 1. Balance data creation with insights consumption. We now know that data is proliferating at a rate that is impossible to keep up with. Production of data is outpacing our ability to use it, although not for a lack of trying. The avalanche of data has

led many top-level execs and the people working under them to view analytics tools as a way to keep up. The thinking is that by creating more metrics, more dashboards, more visualizations and more reports, you can yield more insights. However, these “insights” may not all say the same thing or they can overload you with too much information, including much that is irrelevant. For example, you can analyze click rates in 15 different ways, but that does not necessarily help you understand why people click and why they don’t. This chaos makes it difficult for business leaders to make smart, data-driven decisions vs. ones that are based more on gut feel. You could look to America’s costly drug war over the past 40 years as another piece of evidence that suggests how vast resources aren’t always arrayed in a manner that addresses the heart of a problem. Every business is home to thousands of decision supply chains that include identifying a problem, gathering data,

Table 1: Top data and analytics challenges (N=150 Mu Sigma clients and prospects). Data Challenges (69% of responses)

Analytics Challenges (31%)

Disparate sources of data

Prioritizing where to apply analytics

Legacy vs. new data systems

Scaling our analytics tools and capabilities

Usable data in a consistent format

Best practice sharing across silos

Adding new data when still dealing with existing

Identifying the correct metrics

Master data management

Creating actionable insights

Information governance

Change management and talent

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

generating and evangelizing insights, and then taking decisions on those insights. As an analytics leader, itâ&#x20AC;&#x2122;s your job to formalize this supply chain and inject transparency with a focus on measurable business outcomes. So before creating the next flavor of dashboard, obsess over who and how the resulting insights will be used. This helps keep analytics overload at bay and prevents you from being at the mercy of hundreds of separate decision-makers.

is to slow down and before answering, make sure the questions themselves are right. This is particularly important when dealing with data because one can

2. Focus your team on asking great questions more than finding fast answers. Your team is expected to quickly provide accurate answers when presented with questions. How many leads resulted from this campaign? What specific actions did visitors take on the site before becoming customers? How will the introduction of a new product affect our service quality? When faced with a barrage of questions in a high-pressure situation, it is human nature to blurt out a response. No one likes to seem ill-informed or ignorant. Rather than focusing on answers (which may end up wrong anyway), the better approach A NA L Y T I C S

M A Y / J U N E 2 016

EXE CU TIVE E D G E

always manipulate and interpret data to yield an answer that supports a particular position. Mu Sigma’s founder calls this “choking the data.” The key to really unlocking the value in your data is to ask the right questions. This may involve stepping back and taking a more expansive view of the problem rather than keeping it narrowly bounded. For example, exploring if you might have a broader customer experience problem rather than a pricing problem with a particular SKU. You may have to challenge basic assumptions and return to a first principles point of view, so that you can be confident in your answers. 3. Teach the organization how to fish; don’t just give ’em the fish. You may be under constant pressure to respond to hordes of problem requests. As a result, your teams dedicate so much time to call-and-response that they do not find the time needed to develop their analytical skills and put their creativity to use in a more proactive manner. One remedy that I’ve witnessed is to begin by allocating two-thirds of a team’s efforts to reacting to inbound problems and the rest to more proactive capability building – but over time, flipping that ratio. Here are some core tenets of quality problem-solving, each of which could require a fisherman’s guide unto its own. For one, always begin with measurable 12

A N A LY T I C S - M A G A Z I N E . O R G

outcomes in mind and chart the behavioral changes required to drive those outcomes. Next, make sure you approach problems in a consistent, meta-data driven manner. Third, map and understand the interactions between business problems. Finally, integrate different forms of analytics – descriptive, inquisitive, predictive and prescriptive – into each business problem, to avoid blinkered views. 4. Empower the organization to analyze data and contribute to the greater knowledge pool. You can’t be everywhere at once. Similarly, there is no one governing model for analytics that helps one group streamline all the information necessary to make the right decisions. In fact, governance of analytics can be a confusing, intricate mess, and it is now tangled up with information governance and sometimes data governance as well. Across the board, most governance focuses on lowering risk through decision rights, policies and standards, and this makes sense given the multifold threats inherent in data. So what’s the solution? Look no further than the “Citizen Analyst,” a term coined by Gartner and IBM, which has been used to define someone who may not have a formal background in data science, but uses the wealth of available informational resources and tools to conduct analyses on their own. W W W. I N F O R M S . O R G

THE

NATION’S FIRST

Associate in Applied Science (A.A.S.) degree in Business Analytics on campus or online.

Credential options • Enroll in one or several: • AAS degree

Why Study Business Analytics?

The Business Analytics curriculum is designed to provide students with the knowledge and the skills necessary for employment and growth in analytical professions. Business Analysts process and analyze essential information about business operations and also assimilate data for forecasting purposes. Students will complete course work in business analytics, including general theory, best practices, data mining, data warehousing, predictive modeling, project operations management, statistical analysis, and software packages. Related skills include business communication, critical thinking and decision making.The curriculum is hands-on, with an emphasis on application of theoretical and practical concepts. Students will engage with the latest tools and technology utilized in today’s analytics fields.

Accelerated Executive Program

Our accelerated learning options allow students to complete certificate credentials in two semesters part time or one semester full time. Accelerated options are available for the Business Intelligence and the Business Analyst certificates.

Questions? Tanya Scott

Director, Business Analytics

919-866-7106 tescott1@waketech.edu

• Certificates: Business Intelligence, Business Analyst, Finance Analytics, Marketing Analytics, and Logistics Analytics

Flexibility • Open-door enrollment • Courses are offered in the fall and spring • Courses can be taken online or on campus • Competitively priced tuition

Gain skills in: • Data gathering • Collating • Cleaning • Statistical Modeling • Visualization • Analysis • Reporting • Decision making

Use data and analysis tools: • Advanced Excel • Tableau • Analytics Programming • SAS Enterprise Guide • SAS Enterprise Miner • SPSS Modeler • MicroStrategy

• Presentation

Funded in full by a $2.9 million Dept. of Labor Trade Adjustment Assistance Community College & Career Training (DOLTAACCCT) grant.

businessanalytics.waketech.edu

EXE CU TIVE E D G E Table 2: Five Parallel Work Streams. Data Orientation Create analytics and insights

Drive consumption of analytics

Find answers quickly

Ask the right questions

Solve specific analytical problems

Improve the approach to problem-solving

Govern and mitigate risks

Empower the frontlines

Scale technology effectively

Drive a culture of experimentation

You can use these citizens to your advantage by empowering them with resources they need to do their jobs better. They may be able to create new Tableau cockpits, hire their own data scientists or procure their own modeling tools. As an analytics leader, you can’t beat them, so you’d best join them by offering tools or services, including flexible capacity, and helping them secure funding or providing training and advice for their own problem-solving work. 5. Don’t allow your focus on technology to undermine the culture you’ve fought to create. Those who stay mired in the tangible worlds of data, software and infrastructure can easily lose touch with important intangibles like company culture. As an analytics leader you must balance your focus on technology with a focus on culture. The journey from data to decision science requires introspection. Your team will perform at its best and receive the same from others when it adopts a learning mindset that allows employees to capitalize on change rather than trying to manage or react to it.

Decision Orientation

A N A LY T I C S - M A G A Z I N E . O R G

It’s also important to be open to experimentation and encourage team members to fail fast and often (but cheaply). In addition, be sure to incorporate a tight feedback loop into your analytics work, and focus on the art of problem-solving rather than a never-ending list of projects. You’ll find this helps inspire a mix of creativity and efficiency. These are just a few ideas that can help you bridge the gap between efforts to seize control of the data deluge in your organization, and spending more time on better decisions and business outcomes. Table 2 provides a short summary of the parallel threads that you’ll need to sustain. Do you have other ideas or interesting tactics to share? I’d love to hear them. ❙ Tom Pohlmann (Tom.Pohlmann@mu-sigma.com) is head of Values & Strategy at Mu Sigma, where he is responsible for client experience and programs to ensure that Mu Sigma’s work reflects its unique values and belief system. REFERENCES 1. CSC: http://www.csc.com/insights/ flxwd/78931-big_data_universe_beginning_ to_explode 2. IDC: https://www.emc.com/collateral/analystreports/idc-digital-universe-united-states.pdf

W W W. I N F O R M S . O R G

Dynamic Ideas llc

The Analytics Edge provides a unified, insightful, modern and entertaining treatment of analytics. The book covers the science of using data to build models, improve decisions, and ultimately add value to institutions and individuals. Most of the chapters start with a real world problem and data set, then describe how analytics has provided an edge in addressing that particular problem. The topics covered in the book include:

• • • • • • •

IBM Deep Blue and IBM Watson Sports Analytics and the MIT Blackjack Team The Framingham Heart Study and Kidney Allocation Google’s Search Engine and Recommendation Systems Fraud Detection and Predictive Policing Revenue Management and Emergency Room Operations Asset Management and Options Pricing

In the last chapters of the book, we give a brief overview of the analytics techniques used throughout the chapters, and provide exercises to help the reader put analytics into action.

www.dynamic-ideas.com Phone: 603 882-399 Fax: 603 386-6197 Email: orders@altlog.net

DI_ANALYTICS_AD.indd 1

4/8/16 8:07 PM

ANALY ZE TH I S !

‘Steps’ to analytics success

The near-term impact of today’s increasingly smart machines is a serious concern.

BY VIJAY MEHROTRA

Machine learning was a hot topic at the recent INFORMS Business Analytics and Operations Research Conference in Orlando, Fla. Everyone seemed to be talking about sophisticated automated technologies for combing through mountains of data, making predictions, discovering insights and optimizing decisions. In his keynote presentation, Sam Eldersveld described robots that had been introduced by Amazon in its distribution centers, drawing parallels between its control systems and other machine-learning applications. Illustrating these points with video footage, Eldersveld wryly observed that “this is kinda creepy.” There is surely something creepy about the “Rise of the Machines” scenarios and dystopian futures that science fiction writers and filmmakers are fond of showing us. But for many, the near-term impact of today’s increasingly smart machines is a far more serious concern. A recent article in The Economist observed that “the worry that AI could do to white-collar

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

jobs what steam power did to blue-collar ones during the Industrial Revolution is worth taking seriously [1]” In their new book “Only Humans Need Apply” [2], Thomas Davenport and Julia Kirby openly acknowledge that “knowledge workers’ jobs are at risk” and observe that “experts engaging in the current debate … fall into two camps – those who say we are heading inexorably into permanent high levels of unemployment and those who are certain new job types will spring up to replace all the ones that go by the wayside.” However, rather than offering yet another opinion on the macroeconomic impact of these ever smarter machines, the authors instead endeavor to describe the relative strengths of smart machines (rigorously following rules, doing repetitive tasks quickly and consistently, rapidly reviewing large volumes of data in search of patterns) and humans (contextual understanding, integration of information, complex communications, empathy and creativity). The core of the book is then focused on particular “steps” that individuals might take to leverage these relative strengths in a world where the pace of technological change continues to increase. Their steps are based on two fundamental ideas. First, the authors observe that it is specific tasks, rather A NA L Y T I C S

than particular jobs, that are vulnerable to being automated. Secondly, as more and more routinized tasks are automated, technology itself will produce opportunities for higher-order activities that leverage humans’ unique strengths. “Augment, don’t automate!” is one of the books key mantras, reflecting the authors’ metaphor of smart machines as wheels for the human mind. Most of you reading this column are engaged in what the authors call “Stepping Forward”: developing the cognitive technologies of the future. The book’s chapter on Stepping Forward reads like an ethnography of today’s tech industry, highlighting not only technical jobs in software engineering, data science and research but also roles in product management, marketing, consulting and entrepreneurship that are essential for these technologies from R&D to the potential customers. This chapter also includes a nice feature on Zahar Balaporia, a longtime INFORMS member and leader within the Analytics Society of INFORMS, profiling his work in facilitating the deployment of smart systems at trucking giant Schneider National. Balaporia is cited as an archetypical internal automation leader, an increasingly important role that interfaces directly with those who are “Stepping In” and “Stepping Up.” M A Y / J U N E 2 016

ANALY ZE TH I S !

Those that Step In are the frontline employees and managers who develop a deep understanding of how new cognitive technologies work, along with an appreciation of their potential and an awareness of their limitations. These are the power users and change agents who have the skills, the motivation and the inclination to dig under the hood to understand the model logic, identify its weaknesses from a business perspective, and communicate potentially valuable improvements. These are people working hard in the trenches to help their organizations capture the promised ROI associated with new technologies. In contrast, Step Up people are business leaders who constantly scan the horizon to understand emerging technologies and potential applications, make high-level decisions about which smart solutions to invest in, and manage the myriad challenges of organizational change, all while looking for the next set of big technology-driven opportunities. These are big jobs, few in number but hugely important. As the authors succinctly state, “they decide what smart people do, what smart machines do, and how they work together.” The conference in Orlando featured a session with Craig Brabec, a prototypical Step Up person (he recently joined McKesson as senior vice president for 18

A N A LY T I C S - M A G A Z I N E . O R G

data and analytics after being chief of analytics at Caterpillar). In his session, he thoughtfully discussed emerging supply chain management technologies, the challenges of earning credibility for new technologies with other business executives, the value of looking to other industries for ideas, and the importance of creative partnerships with early-stage technology vendors. As I listened to Brabec speak, it became clear to me that we should expect more and more C-Level executives to emerge from the ranks of those who have Stepped Up in the way that Brabec has, and I suspect that Davenport and Kirby would agree. The other two steps described in the book are “Stepping Narrowly” and “Stepping Aside.” The authors describe Stepping Narrowly as finding and cultivating a deep niche for which the opportunity is too small to attract substantial investment for automation. My prototype here is my friend’s small Midwestern law firm, a worldwide leader in cases associated with garage door opener technologies. The book points out that today’s online search capabilities enable such experts to develop, maintain, promote and deliver services based on their specialized expertise. To Step Aside is to find and/or create a role comprised largely of tasks for which humans have a distinct long-term W W W. I N F O R M S . O R G

advantage over smart machines. The authors cite entertainers, artisans, therapists, writers and designers as examples. They also point out that smart machines can enable Step Aside people to leverage their human strengths by freeing them up from repetitive tasks and enabling them to do vastly more of what they do best. Despite all the disruption that increasingly smart machines are having on the world already, the authors are largely optimistic that organizations will adapt in ways that are positive for people: “As we move more fully into the age of machines … the key to firms’ competitiveness is not the efficiency that automation provides but the distinctiveness that augmentation [of people] allows … they [organizations] will have to attract highly capable people, engage them, and retain them.” I am more ambivalent. As our economy recovers from the Great Recession, job growth has continued to be sluggish, and wages for the middle class have been largely stagnant for quite a long time [3]. Late in the book, Davenport describes a recent conversation with an insurance company executive whose desire to deploy automated claims processing technology is driven by the opportunity to please Wall Street by decreasing labor A NA L Y T I C S

costs. This mindset seems typical of executives in today’s world, and as a result, many of yesterday’s entry-level positions have already been automated (or offshored). Not coincidentally, these issues have also been a big part of this year’s presidential campaign. As I came to the end of the book, I was struck by the amount of creativity, judgment and entrepreneurial skill that are going to be required of individuals going forward. My broader challenge is to figure out how to pass these lessons on to my students (as well as my 12-yearold daughter) to help them avoid getting Stepped On or Stepped Over. My sense is that they should surely read “Only Humans Need Apply” – and also “The Startup of You” [4]. It’s not just the machines that need to learn. ❙ Vijay Mehrotra (vmehrotra@usfca.edu) is a professor in the Department of Business Analytics and Information Systems at the University of San Francisco’s School of Management and a longtime member of INFORMS.

REFERENCES 1. http://www.economist.com/news/ briefing/21650526-artificial-intelligencescares-peopleexcessively-so-rise-machines 2. https://www.harpercollins.com/ 9780062438614/only-humans-need-apply 3. For more on this, see https://www.ted. com/talks/erik_brynjolfsson_the_key_ to_growth_race_em_with_em_the_ machines?language=en 4. http://www.thestartupofyou.com/

M A Y / J U N E 2 016

HEALT H CARE A N A LY T I C S

HIMSS 2016 report: A good year ahead for data aficionados Four key conference takeaways and trends to keep an eye on.

The Healthcare Information Management Systems Society (HIMSS) organizes the largest health information technology conference in North America every year. This yearâ&#x20AC;&#x2122;s event, held in Las Vegas, attracted about 42,000 attendees from around the world. About 1,300 health IT exhibitors showcased their products and services, and more than 300 education programs were presented. I attended this event to get a glimpse of the industry with a focus on healthcare analytics. It was encouraging to hear success stories in health analytics around the country. Following are four key conference takeaways and trends to keep an eye on: 1. Precision medicine is close to reality.

BY RAJIB GHOSH

Vanderbilt University presented a great talk on how analytics is playing a key role in developing precision medicine for cancer patients. Vanderbilt University has emerged as a national leader in precision medicine with 215,000 genetic samples linked to electronic health records without patient identifiers. Vanderbilt will also lead the direct volunteers pilot studies under the

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Secretary Burwell delivered her opening keynote speech in front of a large audience at HIMSS 2016. first grant in the federal Precision Medicine Initiative Cohort Program. Verily, formerly known as Google Life Sciences, will serve as an advisor for the project. While precision medicine is a relatively new medical model, some academic medical centers are already applying aspects of precision medicines powered by data and analytics in some clinical areas to gain more insights. Precision medicine, when matured, can help clinicians detect which early stage cancer patients are fully out of risk of recurrence and who will need follow-up radiotherapy or chemotherapy. The latter in particular is disruptive for patients and may not be used unless necessary. Precision medicine can also identify early A NA L Y T I C S

risk of cancer by detecting anomalies and gene mutations. 2. Population health management powered by analytics is everywhere. Population health management â&#x20AC;&#x201C; with data aggregation platforms playing a key role â&#x20AC;&#x201C; was a major theme this year at HIMSS as evidenced by the many presentations on successful implementations of such programs. The presentations highlighted health systems that brought data from electronic health record systems, practice management systems and in some cases financial systems to centralized data warehouses and then applied various analytics and visualization tools to M A Y / J U N E 2 016

HEALT H CARE A N A LY T I C S

garner knowledge. Visualized knowledge was presented to the physicians at the point of care or to the quality improvement teams via dashboards. One physician group, Crystal Run Healthcare, applied data analytics to identify variability in cost and outcome for its diabetes patient cohort. The group identified best practices across its organization from physicians who achieved better outcomes at a lower cost and then applied that for all. In another instance, North Shore University’s data analytics team leveraged its clinical data warehouse and predictive analytics to predict the risk of patients passing on a contagious and antibiotic-resistant Staphylococcus during an outbreak. They cut the number of patients who needed to be tested by half without any increase of bacterial MRSA (Methicillin-resistant Staphylococcus aureus) cases.

Healthcare data is more precious than credit card data because medical records can’t be reissued. Healthcare data fetches more money on the black market. Unlike the financial industry, healthcare organizations did not face much cybercrime in the past since not much data was digitized. Therefore, they generally do not have the level of maturity and sophistication required when it comes to prevention. Cybercrime against healthcare organizations is a relatively new phenomenon and such organizations have found themselves quite vulnerable. Cybersecurity agencies in United States and Canada recently issued a warning about ransomware attacks against hospitals following a series of such events. As the volume of data grows in healthcare, cybersecurity threats will continue to rise, which will certainly keep data security and privacy officers awake at night.

3. Cybercrime is on the rise as healthcare data becomes rich and precious.

4. Data interoperability is one of the top priorities for ONC.

Another growing but scary trend is the increase of cybercrime activities in the healthcare space. According to 10Fold Communication, in 2015 three of the top seven largest breaches of data occurred in the healthcare industry, including more than 80 million patient and non-patient data records breached from Anthem in February last year.

Without data interoperability, population health management’s promise can never materialize. Data needs to flow from one system to another to achieve a longitudinal view of patient health. The Office of the National Coordinator of Health IT (ONC), therefore, has made this one of its top priorities. Secretary of Health and Human Services Sylvia Burwell kicked

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

off HIMSS 2016 with her opening keynote addressing the topic of interoperability. That set the tone for the rest of the conference. Electronic health record vendors are now actively working to make data liquidity happen through various initiatives such as direct messaging, the Sequoia Project and the CommonWell Health Alliance. If the pressure from ONC Welcome to the data era in medicine. and buyers continues to rise, we should see major improvements in inup with the Institute for Healthcare Imteroperability during the next three years. provement’s (IHI) “Triple Aim” initiative. More data liquidity will also drive up deI am not saying that all the potential for mands for analytics. improving healthcare with analytics has Last year I heard a lot of buzz about been fully realized, but I do think the Afpopulation analytics, but there wasn’t a fordable Care Act and the value-based whole lot of substance barring a few trailhealthcare delivery paradigm is finally blazers. Most vendors offering population showing the results that policymakers inhealth management solutions seemed tended when the bill was signed into law quite immature. Healthcare organizations by President Obama six years ago. ❙ were trying to get out of the hangover of Rajib Ghosh (rghosh@hotmail.com) is an electronic health record implementation independent consultant and business advisor with 20 and adoption challenges. This year the years of technology experience in various industry verticals where he had senior-level management picture has changed substantially. Not roles in software engineering, program management, only did I find engaging conversations on product management and business and strategy development. Ghosh spent a decade in the U.S. “post-EHR” scenarios, but I also heard healthcare industry as part of a global ecosystem stories of successful enterprise data of medical device manufacturers, medical software companies and telehealth and telemedicine solution warehouse program implementation by providers. He’s held senior positions at Hill-Rom, many organizations. Solta Medical and Bosch Healthcare. His recent work interest includes public health and the field of The healthcare industry has shifted IT-enabled sustainable healthcare delivery in the its focus toward care coordination and United States as well as emerging nations. outcome management, which matches A NA L Y T I C S

M A Y / J U N E 2 016

INFO RM S IN I T I AT I VE S

CAP recertification, new society & IAAA Recertification entails continued activity in the field and continued knowledge of improvements, developments and innovation in the field.

TIME TO RENEW CAP CERTIFICATION For many individuals who have earned their CAPÂŽ certification, itâ&#x20AC;&#x2122;s time to renew in order to avoid possibly having to retake the exam. Renewal is less hassle, it costs less and it saves wear and tear on your nerves. Why does anyone have to renew? Passing the exam proved you have the knowledge, but this is a professional certification that requires you continue to maintain and even exceed the level of knowledge you displayed when first taking the exam. Recertification entails continued activity in the field and continued knowledge of improvements, developments and innovation in the field. How does one renew? It is relatively easy. Continued employment in the field can be counted toward recertification, as can educational courses, seminars or workshops, etc. If you are pursuing an additional degree, some courses may be applicable, and the hours you spend there may be used for recertification. If you present at a conference, write published articles or serve as an officer in a relevant association, the

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

hours you spend on those activities may be counted. You need to amass 30 hours of professional development units (1 hour of activity) to recertify. For specifics on what areas of learning and/or activity are acceptable, consult page 28 of the Candidate Handbook, the CAP website or contact certification@informs.org. In related news, Manuel Bolivar of the Univer- Manuel Bolivar (left) with Aaron Burciaga. sidad de los Andes recently became the Highlights at the conference includfirst person in Colombia to earn CAP ed a recognition breakfast for the finalcertification. ists and winners of the various awards sponsored by the Analytics Society. The Society supported the INFORMS ANALYTICS SECTION BECOMES THE ANALYTICS SOCIETY Practitioners Colloquium, which helps The Analytics Section has become practice-oriented students transition INFORMSâ&#x20AC;&#x2122; newest society, so its now into successful real-world careers, and called the Analytics Society of INFORMS. its volunteers offered much appreciated Achieving society status recognizes the mentorship for these students throughnew societyâ&#x20AC;&#x2122;s significantly increased out the conference. membership base (well over a thousand members) and its vastly expanded scope INNOVATIVE APPLICATIONS IN ANALYTICS AWARD of activities, all of which were on display at the 2016 INFORMS Conference The Innovative Applications in Anaon Business Analytics & Operations lytics Award (IAAA), which recognizes Research in Orlando, Fla., in April. The creative and unique developments, apAnalytics Society remains the fastest plications or combinations of analytical growing subdivision of INFORMS. techniques used in practice, has garnered A NA L Y T I C S

M A Y / J U N E 2 016

INFO RM S IN I T I AT I VE S

significant interest in its brief life. The IAAA for 2016 was awarded to an MIT-led team for its submission entitled, “An Analytics Approach to the Clock Drawing Test for Cognitive Impairment,” during the recent INFORMS Conference on Analytics & Operations Research in Orlando. Fla. The Innovative Applications in Analytics Award is the flagship competition of the Analytics Society of INFORMS. The purpose of the award is to recognize the creative and unique application of a

combination of analytical techniques in a new area. The prize promotes the awareness and value of the creative combination of analytics techniques in unusual applications to provide insights and business value. Other finalist submissions included runner-up “Analytics for the Engagement Life Cycle of IBM’s Highly Valued IT Service Contracts” and third-place finisher “Driving Organic Growth with Zilliant SalesMax.” ❙

INFORMS Education Resource Library EVERYTHING YOU NEED TO KNOW ABOUT UNIVERSITY ANALYTICS PROGRAMS IN ONE CONVENIENT SITE • Professors - Search case studies and data sets that you can use in your classroom • Program Directors - Compare notes and learn how to improve - or create! - your school’s analytics program • Students - Find the right school for you: search analytics programs across the U.S. • Industry - Partner with universities and identify top analytics grads

AVAILABLE NOW at http://education.informs.org

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Most enterprises rely on analytic experts to solve their toughest challenges. But to get ahead of the competition, you need to put powerful optimization capabilities in the hands of your business users â&#x20AC;&#x201D; so they can deliver value with no friction or delays.

Take FICO Optimization Modeler for a test drive now, and transform your most complex business challenges into smarter, faster, more profitable customer decisions.

DOWNLOAD YOUR FREE TRIAL

NE W S M AK E R S

Edelman gala and awards galore ORION project demonstrates how operations research can be used to realize hundreds of millions of dollars in cost savings.

UPS WINS 2016 EDELMAN AWARD FOR O.R. ACHIEVEMENT The Institute for Operations Research and the Management Sciences (INFORMS) named UPS as the winner of the 2016 Franz Edelman Award for Achievement in Operations Research and the Management Science at an Oscar-like gala held in Orlando, Fla., in conjunction with the INFORMS Conference on Analytics & Operations Research. The award, considered the “Super Bowl of Operations Research,” honored UPS’s On-Road Integrated Optimization and Navigation (ORION) project that demonstrated how operations research (O.R.) can be used to realize hundreds of millions of dollars in cost savings. “We are very proud to have won the Franz Edelman Award, which represents the highest levels of performance and sophistication in operations research,” says Mark Wallace, UPS senior vice president of global engineering and sustainability. “This honor recognizes many years, countless resources and teams of people dedicated to deployment of one of the largest

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

UPS receives accolades for earning the 2016 Edelman Award. operations research projects in the world. ORION will lead to hundreds of millions of dollars in cost savings, reduced fuel usage, and it supports our efforts to strive towards improved sustainability.â&#x20AC;? UPS, the leading logistics provider in the world, and long known for its focus on efficiency improvement, in 2003 first instituted an ambitious effort to modernize its pickup and delivery operations. This commitment evolved into a suite of systems that are collectively known as package flow technologies (PFT) and an advanced optimization system known as ORION. ORION uses the data foundation of PFT to provide an optimized manifest to its drivers to help them meet the complex demands of providing service with greater efficiency. This is accomplished A NA L Y T I C S

by building more efficient routes, reducing the miles driven and reducing vehicle fuel consumption. More specifically, when fully deployed at the end of 2016 to 55,000 drivers, UPS estimates a reduction of 100 million miles driven annually and a savings of 10 million gallons of fuel per year. That amounts to eliminating more than 4,000 trips around the world per year. ORION also contributes to UPSâ&#x20AC;&#x2122;s sustainability efforts by reducing the CO2 emissions by 100,000 metric tons each year. The company expects ORION to save it $300 million to $400 million annually when fully implemented. The key for ORION is its use of analytics and operations research to identify and capitalize on the small efficiencies on the front lines, and to M A Y / J U N E 2 016

NE W S M AK E R S

use the power of data and heuristics to constantly improve performance. In day-to-day implementation, ORION takes the guesswork out of finding the most efficient and effective routes and schedules for delivery. Drivers face less pressure to make routing decisions and can instead focus on safety and serving customers. Customers also have greater control, since ORION allows them to use their own computers or smart phones to postpone or redirect packages with UPS My Choice to designated UPS Access Point locations, offering them greater convenience and security options. The company considers ORION its foundation for a new generation of advanced planning systems. Because of its sheer size and scope, it has come to be regarded as one of the largest operations research projects in the world. Other finalists and projects included: • 360i for “360is Digital Nervous System” • BNY Mellon for “Transition State and End State Optimization Used in the BNY Mellon U.S. Tri-Party Repo Infrastructure Reform Program” • Chilean Professional Soccer Association (ANFP) for “Operations Research Transforms Scheduling of Chilean Soccer Leagues and South American World Cup Qualifiers” 30

A N A LY T I C S - M A G A Z I N E . O R G

• New York City Police Department (NYPD) for “Domain Awareness System (DAS)” • U.S. Army CommunicationsElectronics Command (CECOM) for “Bayesian Networks for U.S. Army Electronics Equipment Diagnostic Applications: CECOM Equipment Diagnostic Analysis Tool, Virtual Logistics Assistance Representative” GENERAL MOTORS AWARDED INFORMS PRIZE General Motors, which is using big data and advanced analytics to predict failure of certain automotive components and systems before customers are affected, was named the winner of the 2016 INFORMS Prize for operations research and the management sciences. This year’s INFORMS Prize was presented at the 2016 INFORMS Conference on Analytics & Operations Research in Orlando, Fla. Industry-first proactive alert messages sent to customers through GM’s OnStar system covering potential issues with a vehicle’s battery, fuel pump or starter can transform an emergency repair into planned maintenance. A recent example of applying operations research and management science to the most complex issues the company W W W. I N F O R M S . O R G

faces led to the INFORMS Prize. “Over the last seven decades, OR/MS techniques have been used to improve our understanding of everything from prognostics to traffic science and supply chain logistics to manufacturing productivity, product development and vehicles telematics and prognostics,” says Gary Smyth, executive director of GM Global R&D Laboratories. “These approaches to problem-solving permeate almost everything we do.” GM has hundreds of OR/MS practitioners worldwide who play a vital role in everything from designing, building, selling and servicing vehicles to purchasing, logistics and quality. The team is constantly developing new business models and vetting emerging opportunities. Another example of management science positively impacting the business is helping to understand what products and features customers most want to create, as well as price features and option packages that would sell best. That work extends to determine the ideal number of vehicles and what vehicle variants Chevrolet, Buick, GMC and Cadillac dealers in the United States should stock. In addition, advanced analytics help dealers achieve GM’s goal of creating customers for life. The company A NA L Y T I C S

recently received the 2015 overall manufacturer loyalty award from IHS. The impact OR/MS is now having to the business can be traced to 2007, when GM created a center of expertise for operations research to promote best practices and transfer new technologies. GM has since expanded to include partner teams in product development, supply chain, finance, information technology and other teams. The INFORMS Prize honors effective integration of operations research in organizational decision-making. The award is given to an organization such as GM that has repeatedly applied the principles of O.R. in pioneering, varied, novel and lasting ways. CARNEGIE MELLON SCHOOLS RECEIVE UPS GEORGE D. SMITH PRIZE The School of Information Systems & Management and the School of Public Policy and Management at the H. John Heinz III College, Carnegie Mellon University, received the 2016 UPS George D. Smith Prize from INFORMS. The award recognizes excellence in preparing students to become practitioners of analytics and operations research. The award was presented at the 2016 INFORMS Conference on Business Analytics & Operations Research held April 10-12 in Orlando, Fla. M A Y / J U N E 2 016

NE W S M AK E R S

The other finalists included the Institute for Advanced Analytics, North Carolina State University, and the Operations Research Program, United States Air Force Academy. The UPS George D. Smith Prize is awarded to an academic department or program for effective and innovative preparation of students to be good practitioners of operations research, management science or analytics. Established in 2011, the award is named in memory of the late UPS chief executive officer who was a patron of operations researchers at the Fortune 500 company. STANFORD TEAM WINS SYNGENTA CROP CHALLENGE

future planting plan,” which modeled a system for predicting soybean seed variety selection. “It has been a wonderful experience working with Syngenta on this project, and we are excited about the impact our work can have on improving crop yields and addressing food security challenges,” says Xiaocheng Li. The Challenge tasked participants to develop a model that predicts the seed varieties farmers should plant next season to maximize yield. The inaugural competition aimed to address the challenge of global food security by fueling innovation among experts applying advanced analytics in biochemistry and agriculture. “Global food security is one of the greatest challenges facing the next generation, and there is a significant need to engage a broader talent base into agriculture,” says Joseph Byrum,

Syngenta and the Analytics Society of INFORMS named Xiaocheng Li, Huaiyang Zhong and associate professors David Lobell and Stefano Ermon – a team from Stanford University – as the winners of the inaugural Syngenta Crop Challenge in Analytics. The team was awarded a $5,000 prize for their entry, “Hierarchy modeling of soybean variety yield and decision making for Finalists and officials of the inaugural Syngenta Crop Challenge. 32

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

S S E N I BUS TICS Y L A N A

Online

MBA

BECOME MORE AT THE

Beacom School of Business Best Value MBA Ranked Top

Top Rated College by Forbes & Princeton Review

AFFORDABILITY & ACCREDITATION by Best Master’s Degree

Online MBA Ranked Top

IN THE WORLD

by Princeton Review

MBA – General MBA – Business Analytics MBA – Health Services Administration

Get started at

www.usd.edu/onlinemba cde@usd.edu • 800-233-7937

NE W S M AK E R S

Syngenta head of soybean seeds product development and lead for the Syngenta Crop Challenge in Analytics committee. “This competition clearly demonstrated that people outside and adjacent to the industry can make noteworthy contributions.” The finalists made their presentations at the INFORMS Conference on Business Analytics & Operations Research in Orlando, Fla. Programs were evaluated based on the rigor and validity of the process used to determine seed varieties, the quality of the proposed solution and the finalists’ ability to clearly articulate the solution and its methodology. The runner up, “Decision assist tool for seed variety selection to provide best yield in known soil and uncertain future weather conditions” (authored by Nataraju Vusirikala, Mehul Bansal, Prathap Siva Kishore Kommi) received a $2,500 prize. The third place entry, “Balancing weather risk and crop yield for soybean variety selection” (authored by Bhupesh Shetty, Ling Tong and Samuel Burer), received a $1,000 prize.

“The submissions from the Syngenta Crop Challenge in Analytics represent best in class science,” Byrum adds. “What is striking is the overall professionalism, quality and effort that the finalists put into the presentations. The teams were clearly committed and had a deep connection to the challenge.” Syngenta, a global agribusiness headquartered in Switzerland, donated the prize money from its 2015 Franz Edelman Award win in support of a commitment to run the Syngenta Crop Challenge for the next four years. “Syngenta is a great example of a company using operations research to better both its own performance as well as to help better society,” says Melissa Moore, executive director of INFORMS. Next year’s Crop Challenge will be announced this May with submissions due in January 2017. For more details about the Syngenta Crop Challenge and to register for the upcoming Challenge, visit www.ideaconnection.com/syngentacrop-challenge. ❙

Request a no-obligation INFORMS Member Benefits Packet For more information, visit: http://www.informs.org/Membership

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Nashville 2016 You are invited to present to 5,000 plus attendees, join intriguing plenary presentations, panel discussions, and tutorials, or submit an abstract for one of our numerous oral and poster tracks.

2016 INFORMS ANNUAL MEETING November 13-16 | Nashville, Tennessee

Come to the “Music City” to learn, share your expertise and experiences, build your professional network, ﬁnd prospective employers or employees, and reconnect with colleagues and meet new people. INFORMS is looking forward to hosting at the Music City Center and Omni Nashville in the heart of the city. You’ll be steps away from eclectic live music, wonderful restaurants, and Nashville history. We look forward to seeing all of you once again in 2016!

May 15 - Abstract Submission Deadline August 1 - Poster Competition Submission Deadline September 1 - Poster Submission Deadline

http://meetings.informs.org/nashville2016

B IG DATA & A NA LY T I C S

Fulfilling the promise of analytics Get the most from big data and analytics: Second of two articles based on an EY-Forbes study emphasizing the human element.

BY CHRIS MAZZEI hile data and analytics have been part of business for a long time, itâ&#x20AC;&#x2122;s only in recent years that the value they provide has captured the attention of senior executives and managers. That is because there has been an explosion of systems and devices that generate data coupled with greatly reduced costs to collect, store and analyze that data. The result: big data. Today, big data and analytics (BD&A) are changing the way decisions are made

A N A LY T I C S - M A G A Z I N E . O R G

â&#x20AC;&#x201C; from everyday business challenges to differentiating a company in an effort to gain competitive advantage. Decisionmakersâ&#x20AC;&#x2122; eyes and imaginations are now open to new opportunities that might have slipped under the radar in the past. However, for all the benefits BD&A promises, it is disruptive. And like all disruptive concepts, it can turn any organization on its head. In fact, because of the disruptive nature of BD&A, many companies are struggling to derive full value from their initiatives and capabilities. A

W W W. I N F O R M S . O R G

Photo Courtesy of 123rf.com | rawpixel

Most organizations do not have an effective and aligned business strategy for competing in a digital, analyticsenabled world. recent EY and Forbes Insights survey of 564 executives in large global enterprises found that most organizations still do not have an effective and aligned business strategy for competing in a digital, analytics-enabled world. For many companies, the people- and process-related change management issues have prevented analytics from fully delivering on its potential. Illustrating this challenge, the survey found that while 78 percent of organizations said data and analytics are changing the nature of competitive advantage, and despite the 66 percent who said they are investing $5 million or more in

A NA L Y T I C S

analytics, only 12 percent of organizations described themselves as analytics leaders. The reason so few companies class themselves as leaders, or are able to drive competitive differentiation, is because of the 89 percent that admitted that change management is the biggest barrier to realizing analyticsâ&#x20AC;&#x2122; value. What separates these leaders in analytics excellence from those organizations still struggling with their programs? The EY-Forbes survey found that the most advanced companies â&#x20AC;&#x201C; those within the top 10 percent in the survey results â&#x20AC;&#x201C; use BD&A in their decision-making â&#x20AC;&#x153;all

M A Y / J U N E 2 016

B IG DATA & A NA LY T I C S of the time” or “most of the time” and said they considered their organization as “advanced” or “leading” in applying BD&A to business issues and opportunities. Further, these companies reported a “significant” shift in their ability to meet competitive challenges. In other words, the top 10 percent of companies are not just producing data and analytics, they use the analyticsdriven insights at the point at which decisions are made. The ability of an organization to utilize, or as we refer to it, “consume,” analytics like this is not easy. Most companies have rich BD&A production strategies and processes in place – a topic that warrants another article for another day, but driving BD&A consumption throughout the organization has proven to be challenging. For many companies, being able to consume analytics requires an entirely new mindset – moving from viewing BD&A as simply a technology issue to viewing it as a powerful business tool. To do this, companies must consider the human element of analytics. The human element focuses on why, despite all the data available, analytics insights are not used. Are employees not aware of the data? Do they not understand it? Are they not trained on how to access it? Are the right analytics insights reaching the right people at the right

A N A LY T I C S - M A G A Z I N E . O R G

time? Or, perhaps they are but the insights are dismissed because they tell an inconvenient or unbelievable truth. Many companies are hitting the reset button on their BD&A initiatives to take the human element into account. But this is often easier said than done. To ensure that organizations receive the highest return on investment from their programs, three key areas should be considered: • Strategy. What strategy should an organization adopt in the face of disappointing returns on analytics investments? How does a company shift from analytics being a technical issue to a strategic business imperative? • Leadership. What leadership does a company need to have in place? Both at the senior level and at the operational level throughout the organization? • Consumption. How are analytics insights consumed, both at the individual and the organizational level? CREATING STRATEGY THAT CAN WIN IN A BD&A-ENABLED WORLD Without a strategic approach providing high-level guidance, analytics efforts are rudderless. The first step in steering the ship is to articulate a vision for the role of analytics.

W W W. I N F O R M S . O R G

Perhaps unsurprisingly, when we asked “what best describes the state of your organization’s overall strategy toward data and analytics?” we found that the most successful companies – the top 10 percent – say that analytics is central to their overall strategy. These organizations may have data scientists and analysts on staff, have identified targets for analysis and even talk about being an “analytics-driven enterprise” in their mission statements and annual reports. But enabling the organization to leverage analytics requires more than introducing technology or launching new programs. It is about having the organizational alignment, the governance and the culture to harness the transformative potential of analytics to guide the organization to success. Organizations leading the way in this area are formulating and acting on their visions in a tangible way. The top 10 percent of companies, “the best” in the survey, scored an average of 22.6 out of a total score of 25, compared with 12.6 percent for “the rest.” Clearly, the ability of organizations to define themselves in terms of analytics capabilities ranges widely. The transformation to an analytics-driven enterprise requires the integration of many components and a common purpose and vision. This is where strong leadership can make a difference.

A NA L Y T I C S

ANALYTICS LEADERSHIP REQUIRES A RENAISSANCE PROFESSIONAL When analytics needs to be at the heart of the business strategy, it’s logical that the analytics leader needs to be proficient in business as well as the technical aspects of analytics. While it might be logical, many organizations have struggled to install the right person at the head of their analytics programs. That’s in no small part too because it’s tough to find an individual with the rare blend of talents required. An analytics leader must have sound knowledge of data management fundamentals such as data extraction, data quality and developing data architecture. They must have an advanced understanding of the mathematic disciplines that underpin analytics as well as the enabling technologies. They must be able to effectively translate, using visualizations, what the data are saying into a compelling story that will create action among other senior leaders. And they must have an intimate knowledge of the business and the industry landscape, coupled with a hunger to innovate new products, process, or, in order to achieve competitive differentiation, an entirely new business strategy. The survey results illustrate this need for a multi-talented individual, with the top 10 percent of firms listing new revenue streams, sector knowledge/experience, statistical proficiency, and data extraction

M A Y / J U N E 2 016

B IG DATA & A NA LY T I C S and transformation as some of the most important things they look for in their analytics leader. Finding an individual with all the necessary skills (watch this video for more) is not easy. That’s why those who do combine these skills are in high demand, with the corresponding high salaries. ENHANCING BD&A CONSUMPTION ACROSS THE ENTERPRISE Analytics consumption takes place at two levels in an organization: senior-level executives and other decision-makers gain insights to help them understand their markets, product or service positioning and operations. Individual employees at all levels and locations throughout an enterprise can

use analytics to help improve their own decision-making. The top 10 percent of companies in the EY-Forbes survey scored an average of 77 percent for organizational consumption, with the rest of the companies scored an average of only 51 percent. As with strategy and leadership, the top companies are “getting it right” when it comes to data consumption, and this is reflected in their bottom lines. Industry sectors leading the way in enabling analytics consumption organization-wide include technology and consumer products and retail. But even for leaders, BD&A is as much an art as it is a science. A successful analytics environment does not

Figure 1: Based on a recent EY-Forbes study, for a majority of organizations the tools and technologies they employ to leverage data and analytics are either immature or have yet to be standardized. Source: EY study: “2015 EY/Forbes Insights Data & Analytics Impact Index: Don’t Forget the Human Element”

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Figure 2: Data and analytics skills are becoming essential to a myriad of jobs and business roles. Source: EY study: “2015 EY/Forbes Insights Data & Analytics Impact Index: Don’t Forget the Human Element”

depend on technology alone; it requires marshaling human capital to deliver the right insights at the right time. As noted earlier, any transformative initiative requires support from the top, but employees at all levels must also buy into the effort. They must also be trained so they can effectively understand and use analytics (see Figures 1 and 2 for how leading companies recognize, monitor and support their staff). The value of analytics comes from the behavioral alignment required to “consume” analytics in order to move from insights to action to creating value. A BRIGHTER FUTURE Competitive advantage over the coming years will depend on how well companies in all industries embed analytics into their enterprise-wide business strategy and decision-making. When a company

A NA L Y T I C S

puts analytics at its heart, seismic changes can take place. The pace of that change in any given organization will depend on two factors: First, the sense of urgency with the organization, typically driven by how rapidly the competitive landscape of the organizations is changing. And second, having the right leader in place who can implement the right organizational structure and governance that will embed analytics at the point decisions are made. Only then can organizations finally realize the promise of BD&A and secure the competitive advantage they need to succeed. ❙ As EY’s (Ernst & Young) global chief analytics officer, Chris Mazzei leads the Global Analytics Center of Excellence that serves as a catalyst for transformation both internally, within EY, to embed analytics into its service offerings across all business lines, as well as externally for EY’s clients by delivering analytics offerings that help organizations grow, optimize and protect value.

M A Y / J U N E 2 016

EXPLOITIN G A NA LY T I C S

How to get the most out of

data lakes Three requisite business skills that facilitate self-service analytics at unparalleled speed.

BY SEAN MARTIN one of the premier indicators of the flourishing self-service movement within the datasphere, semantically enriched smart data lakes provide a means for business users to access and analyze big data on an unprecedented, enterprise-wide scale. The enduring relevance of these platforms is almost entirely based on the value that end-users derive from them. Maximizing that utility necessitates a set of skills both familiar and foreign to the business.

A N A LY T I C S - M A G A Z I N E . O R G

The full appreciation of the skills necessary for optimizing data lakes is best understood by elucidating previous competencies that are now obsolete and contemporary ones that have usurped them. Such a transformation heralds a long-awaited displacement of technological reliance by a renewed emphasis on domain knowledge and data savvy. The result is heightened productivity and increased performance thresholds fueled by self-service analytics proficiency.

W W W. I N F O R M S . O R G

A NA L Y T I C S

M A Y / J U N E 2 016

Photo Courtesy of ShutterStock.com | jörg röse-oberreich

ABANDONED TECHNOLOGICAL SKILLS Deployments of non-semantic data lakes and relational means of ingesting, transforming and analyzing data have always required an inordinate focus on technological competence. It’s a recurring paradox: Business users were unversed in integration requirements, ETL (extract, transform, load) processes and the intricacies of data modeling, yet their abilities to fulfill business objectives with data were inherently circumscribed by them. Utilizing these traditional means of accessing business intelligence and analytics required end users to be responsible – or depend on someone else – for facility in Deployments of non-semantic data require an inordinate focus on the finer points of SQL, writing technological competence. code, data modeling, Hadoop skills and processes every time business requireMapReduce jobs. ments or data sources changed, deployInvariably, the dearth of such intricaments of analytics became less frequent, cies on the part of the business resulted the esteem for data-driven processes dein an over-reliance on IT departments creased, and tension between the busifor analytics and business intelligence ness and IT fomented. (BI). More recently, data scientists have REFINED BUSINESS SKILLS been deployed to account for this skills shortage in source integration, loading Enriching data lakes with semanand analytics results. With these profestic technologies obsoletes the relisionals tasked with retooling models and ance on technological competencies

DATA L AK ES

and provides the business self-service data access and analytics. Consequentially, there is a greater emphasis on business skills for utilizing these options, which never expose nor require any end-user cognizance of underlying semantic components such as RDF, OWL, SPARQL, taxonomies or vocabularies. Initial configuration of smart data lakes necessitates IT or data scientist involvement as do continued efforts to load new data sources. However, most critical facets of linking, contextualizing and data preparation is done upfront and incorporated into an evolving semantic model that facilitates self-service analytics at unparalleled speeds, emphasizing the following businessoriented competencies: Relating domain expertise to data. Although it combines aspects of data-driven processes and IT involvement, this particular skill is unique to the business and typifies the way that smart data lakes enhance its ability to perform. Most business users are proficient in their domains. Maximizing smart data lake utility, however, requires pairing that expertise with an ability to use data to improve it and business objectives. Specifically, the business needs to understand what types of information relate to objectives, how it is linked to additional information, 44

A N A LY T I C S - M A G A Z I N E . O R G

and what effect synthesizing that data creates for achieving outcomes. These users need to identify which particular attributes and properties of data benefit their business processes most, and formulate questions around them resulting in decisive action. Tool manipulation: dashboards and BI. This competency represents the extent of the technological involvement on behalf of the business in optimizing smart data lakes. Its prominence is somewhat tempered by the self-service movement’s simplification of visualizations for end users prior to the popularity of these hubs. Nonetheless, business users need proficiency in manipulating the various forms of “publishing” and viewing the results of analytics endeavors facilitated by data lakes, since they no longer have to request them from IT. Competencies in creating and tailoring dashboards, visualization tools or even previously used BI and analytics platforms are needed to determine the impact of analytics on business processes. These skills can be as basic as looking at a web browser interface for results; any variety of platforms works with these repositories. Embracing exploratory analytics. This competency is partly based on analytics adroitness and partly predicated on the business user’s choice. Perhaps W W W. I N F O R M S . O R G

even more than skill in relating domain expertise to data, it represents the full extent to which end-user ability can influence the unparalleled analytics potential smart data lakes provide. Simply understanding that newfound scope of analytics â&#x20AC;&#x201C; and exploiting it â&#x20AC;&#x201C; requires business user skill, particularly for those that are accustomed to traditional limitations of non-semantic data lakes and relational methods. Smart data lakes enable the business to traverse all of its organizationâ&#x20AC;&#x2122;s data,

A NA L Y T I C S

not just those in their particular domain. The expanded scope of possibilities that such data yields for analytics is only restrained by the ingenuity and drive of the user. The encompassing nature of such analytics, and the expedience at which questions are answered only to beget more questions, entails a different conception of the possibilities and relevance of analytics itself. The business must adopt this exploratory analytics mindset to approach the true yield that the data lake concept offers.

M A Y / J U N E 2 016

DATA L AK ES

All of the skills business users need to optimize smart data lake deployments pertain to analytics. This fact is largely entrenched in the reality that most of the other facets of data lakes (integrating sources, contextualizing and linking data, adhering to governance practices) are automated via the incorporation of the semantic model at the heart of these platforms. Those that are not, such as loading new sources and types of data, are done by IT and are easily added to the semantic model without the typical delay associated with this process. Consequentially, business users have the luxury of concentrating on analytics to achieve departmental objectives. MAXIMIZING DATA LAKES: EXPLOITING ANALYTICS The majority of the skills requisite for business users of smart data lakes revolve around business itself. This notion especially applies to the integration of domain knowledge with data and the basic dexterity required for tool manipulation. The lack of technological skills needed on the part of the business, however, should not be mistaken for some sort of sleight of hand. The reason these users can now traverse their organization’s entire information assets with an exploratory analytics mindset that may prove revolutionary is that the underlying semantics technologies perceive 46

A N A LY T I C S - M A G A Z I N E . O R G

the context and relationships between data elements – the end user does not. Instead, the business user merely reaps the benefits. Similarly, it’s the technology – graph models based on the semantic technology standards, not the business user – that is responsible for linking data sources based on those relationships for integration efforts, thus allowing the business to reap the benefits again. The same concept applies to the ability to parse through those data elements in their native forms at speeds that empower those users to leverage more of their data quicker than previously possible. The business user is not responsible for any of those processes, the technology is. Nevertheless, when equipped with the aforementioned skills he or she can readily monetize them. ❙ Sean Martin, founder and chief technology officer of Cambridge Semantics, has been on the leading edge of Internet technology innovation since the early 1990s. Prior to founding Cambridge Semantics, a provider of smart data solutions driven by semantic web technology, he spent 15 years with IBM Corporation where he was a founder and the technology visionary for the IBM Advanced Internet Technology group.

Help Promote Analytics Magazine It’s fast and it’s easy! Visit: http://analytics.informs.org/button.html

W W W. I N F O R M S . O R G

pply for 2016 THE DANIEL H. WAGNER PRIZE Excellence in Operations Research

Apply to win this prestigious practice prize that rewards professionals who devise innovative analytical methods, utilize those methods in a veriﬁably successful OR/analytics project, and describe their work in Two-page abstract is due by May 1, 2016.

Daniel H. Wagner

This top INFORMS practice prize spans all O.R. and analytics disciplines and application ﬁelds. Any work presented in an INFORMS section or society practice-oriented competition is eligible as long as the work did not result in a published paper.

The Wagner Prize competition is high-proﬁle, with its own track at the INFORMS Annual Meeting. Presentations are widely distributed via streaming video. Finalist papers are published as a special issue in INFORMS respected practice journal Interfaces. The 2016 competition will be held at the INFORMS Annual Meeting, November 13-16, in Nashville, Tennessee. First-place prize of $1,000 will be awarded at the Edelman Gala, during the April 2017 Conference on Business Analytics and O.R. in Las Vegas, Nevada.

Get inspired by the 2015 Wagner ﬁnalists by watching their presentations at the INFORMS Video Learning Center by scanning the QR code. A special issue of Interfaces will publish the 2015 winning paper along with the other ﬁnalists: “Integrated Planning of Multitype Locomotive Service Facilities under Location, Routing, and Inventory Considerations” “Using Analytics to Enhance Shelf Space Management in a Food Retailer” “Scheduling Crash Tests at Ford Motor Company” “Strategic Redesign of Urban Mail and Parcel Networks at La Poste” WINNING PAPER “Machine Learning Framework for Predicting Vaccine Immunogenicity”

www.informs.org/wagnerprize

PRODUC T O R S E R V I C E ?

Making wise decisions with the Internet of Things Consumer fantasy comes true, but are organizations ready for even bigger, wider and deeper data?

BY TAYFUN KESKIN (left) AND HALUK DEMIRKAN he technology that lets us control our smart thermostats and wireless door cameras is a part of the Internet of Things (IoT) ecosystem. In order to make everyday objects “smart,” we equip these “things” with sensors, processors and wireless communication capabilities. The IoT sounds like a consumer fantasy or a science fiction come true. The convenience of turning off the home lights from miles away or leaving the grocery

A N A LY T I C S - M A G A Z I N E . O R G

purchase to the refrigerator when milk needs to be replenished sounds technologically interesting. However, there is more to the IoT than the technological lifestyle enhancement by using smart devices. The actual potential of IoT lies on the corporate side, enabling organizations to collect and analyze data from sensors on equipment, pipelines, weather stations, meters, delivery trucks, traffic lights, automobiles, healthcare devices and other types of machinery. W W W. I N F O R M S . O R G

A NA L Y T I C S

Photo Courtesy of 123rf.com | Vesna Cvorovic

Cisco predicts the global IoT market will be $14.4 trillion by 2022, with the majority invested in improving customer experiences [1]. And the number of connected devices is projected to grow from 22.9B in 2016 to 50.1B by 2020 [2]. Countless data sources enable people, and systems, to make much more effective and efficient decisions about nearly everything. In addition, they will be able to act on those decisions much faster [3]. For example, cities will be transformed with smart technologies through dynamic routing and signage for both drivers and pedestrians. It could manage public transit and predict the need for government services based on environmental conditions. At the individual level, nanotechnology in our clothing could pair up with our smart phones or charge it with the electricity generated by our body movements [4]). Then there is the booming market for wearable tech, such as the Fitbit or Apple Watch, which has already surpassed $2 billion in sales, with well over 84 million units sold so far. These devices monitor heart rate, sleep patterns, diet, exercise and more, and transmit the collected data to apps and cloud servers. IoT already started to change the way we live and work, but every disruption comes with many opportunities and challenges. The main objective of this article is to examine opportunities and challenges of

The IoT is changing the way we live and work, but whatâ&#x20AC;&#x2122;s next? IoT and to provide a basic roadmap for smart IoT analytics. REAL WORLD INSIGHTS FROM IoT PROJECTS IoT solutions can open a completely new world of data for organizations. For example, stoplights with embedded video sensors can adjust their greens and reds according to traffic and the time of day. That represents a double-win, reducing both congestion and smog, since vehicles idling at red lights burn up to 17 percent of the fuel consumed in urban areas [4]. Understanding IoT-produced data requires more than just creating Hadoop big data and big SQL solutions. Organizations need to understand where is all the data M A Y / J U N E 2 016

INT ERN ET O F T H I N G S

provided by those IoT processors going to be stored and what are the problems around them in order to harness the power of the IoT. They also need wider and deep data analytics tools to make their organizations more efficient and effective. If organizations are planning to get insights about customers, employees, manufacturing facilities and many other things that the IoT promises, they are also going to have to keep all that data somewhere, perform analytics and convert them to meaningful information and knowledge for efficient and effective decisions. Processing large quantities of IoT data in real time will increase as a proportion of workloads of data centers, leaving providers facing new security, capacity and analytics challenges [5]. In the IoT ecosystem, smart-things will connect remote devices and systems to each other and transmit a data stream on a data management platform. The data or even the devices will be incorporated into existing processes to provide information on the location, status, activity and functionality of those systems, as well as information about the people who own and operate them. Big data tends to arrive as a steady stream and at a steady pace, although it can arrive in batches such as test logs that can be processed and passed on straight away. The real value can only be uncovered using analytics. It is rarely used for production 50

A N A LY T I C S - M A G A Z I N E . O R G

purposes. On the other hand, used data needs to be deleted very quickly unless it is needed for compliance reasons. Uncovering the business value of IoT data. Analytics is seen as the key to making investments in IoT technology worthwhile. As we discussed earlier, IoT solutions has potential applications well beyond the consumer space. For example, package delivery trucks, manufacturing systems and electrical grids all typically have sensors to monitor performance. More and more companies are now starting to collect and store data from such sensors. The next step is to analyze the data. Looking for patterns in it could illuminate ways to improve business operations, such as doing more preventive maintenance or designing more efficient delivery routes. Organizations need to start developing business cases on how they can incorporate a mix of structured and unstructured information – think of it as wide and deep, not only big data. For many years, online companies have been tracking consumer’s Web-click data and using them for additional revenue generating functions (e.g., targeted advertising). Now it is time to collect and analyze stream data that comes from “things” to provide better services and products to consumers. In order to make sense out of all that data, companies are going to need skilled analysts. W W W. I N F O R M S . O R G

Table 1: Sample list of challenges and opportunities. Managerial

Technological

• Inter- and intra-organizational workflows. Almost every IoT business model has B2B2C (many Ms and Cs) collaborations that requires additional settings for nondisruptive smooth processes with rigid service level agreements, security, data stream and messaging protocols.

•

Massive scaling of connected devices has the potential to stretch the limits of current information and communication technologies. Will existing internet protocols suffice? Will we be able to store massive amounts of data created by smart things? How will this affect energy sources and sustainability since most smart devices will require a battery or grid-power? All these scaling questions will be a part of technological challenges as IoT technologies become ubiquitous.

• Pricing products and services is a tricky business.

•

Traditionally, IoT-enabled devices have been in closed systems. For example, cars collect data from several sensors within the device but it is not typically transmitted to an open environment. However, as consumers expect smart goods to communicate, openness questions will emerge. More importantly, can we expect a reasonable level of privacy as we increasingly quantify ourselves and store this data on the cloud? For example, should we allow an insurance company to use our eating habits to design our healthcare insurance policy? Alternatively, when a smart device measures our heart rate and blood pressure, do we always want a doctor to access this information? New unified communication interfaces, and an open communication environment seems inevitable in the future. However, privacy policies for any IoT-enabled platform are required in order to solve privacy concerns.

•

Security attacks are typical against smart devices such as computers and servers, and it will not be different for refrigerators and other inanimate objects, as they get smarter. Probably because of the openness expectation, IoT players ignore addressing security concerns as a primary agenda item. To secure a smart device, it not only needs to detect and protect itself from the attack, it also needs to perform repairs to function properly.

•

Consumer privacy… securing the personal data of individuals as the consumer goods they use become increasingly digitized.

•

Human interfaces can make or break a product since acceptance of a new technology depends on it. Successful human interfaces require a good understanding of natural controls, creating software based on system identification models, and incorporating human behaviors such as feedback control into the system. Winner devices in the IoT market needs to have the simplest and easiest-to-use interfaces.

•

Managing and integrating data and information flows between varying types of devices from a wide range of global manufacturers with proprietary data and technologies.

•

Network architecture enabling low latency control.

•

Standardization… Almost all Internet-enabled devices being produced today come with proprietary software that makes it hard for various devices to communicate with each other.

•

Performing analytics and management of the physical objects and low-level events to detect signals and predict impact.

•

Development of algorithms enabling derivation of useful information.

•

Orchestrating those signals and objects to fulfill complex events or end-to-end business processes.

•

Data governance… who owns the data.

•

Bandwidth sharing wireless communication.

•

Battery/scavenger sources enabling power for life.

•

Scaling network size enabling processing of sensor generated data at the level of Brontobytes.

• Most successful players figure out crossside network effects and their market presence strategies.

Photo Courtesy of 123rf.com | ximagination

• Envelopment can be disruptive for a market but survivors determine industry standards and reap its benefits. Different sides of the IoT market will have similar consumer bases.

The Internet of Things combined with smart devices create a long list of managerial and technological challenges and opportunities. A NA L Y T I C S

M A Y / J U N E 2 016

INT ERN ET O F T H I N G S

Choosing IoT analytics solution wisely. Building analytics solutions that can handle the scale of IoT solutions isn’t easy, but the right technology stack makes the challenge less time consuming. These data storage, management and analytic solutions need to be chosen wisely [6]. Basic steps of IoT analytics include: • Number protocols enable the receipt of events (or transactions) from IoT devices. It doesn’t matter whether a device connects to the network using Beacon, Wi-Fi, Bluetooth, cellular or a hardwire connection, only that it can send a message to a broker using a defined protocol (e.g., Message Queue Telemetry Transport, Constrained Application Protocol, XMPP). • Once a message is received by a broker such as Mosquito, you can hand that message to the data hosting and analytics system. A best practice is to store the original source data before performing any transformations. • This unstructured message data can be stored in Hadoop, Hive or Couchbase-type NoSQL document databases, or it can be stored in big SQL databases after transformation. Most of the time, data from devices in their raw form are not directly suited for analytics. Quality and 52

A N A LY T I C S - M A G A Z I N E . O R G

transformation steps need to be followed to clean the data and complete the missing data. • After transformation, this data needs to be stored in a NoSQL or SQL database for analysis. Apache Storm is explicitly designed for handling continuous streams of unstructured data in a scalable fashion, performing any calculation that you can code over a boundless stream of messages. There is an ongoing debate about using Hadoop type of framework to analyze unstructured data or using Big SQL databases for large relational structured data. • After data storage and in-memory metric development, analytic tools like Tableau, BIRT, Pentaho, JasperReports or similar tools can be utilized to create any required reports or visualization. Companies need to consider architectural changes. A Boeing engine generates 40 terabytes of data an hour. With about 29,000 commercial U.S. flights per day, engines generate over 2 zettabytes of data per year. Every day, Intelligent Mechatronic Systems Inc. (IMS) collects 1.6 billion data points from hundreds of thousands of automobiles in the United States and Canada. The cars are equipped with devices that W W W. I N F O R M S . O R G

CERTIFIED ANALYTICS PROFESSIONAL Analyze What CAP Can Do For You ®

www.certiﬁedanalytics.org

INT ERN ET O F T H I N G S

track driving distance, acceleration, fuel use and other information on how the vehicles are being operated – data that IMS uses to support usage-based insurance programs and fleet and traffic management initiatives [5]. In August 2015, after a yearlong project, IMS added an Apache Cassandra NoSQL database along with data integration and analytics tools from Pentaho. This setup lets the analytics team perform finer-grained analysis of customer driving behavior in search of patterns and trends that could help insurers fine-tune their usage-based policies and rates. Traditional relational databases that are hosted in local and remote locations will not be sufficient to host and analyze these bigger data sets [7]. Organizations need to investigate how they can transform to federated data architecture models by utilizing emerging NoSQL unstructured big data management frameworks (e.g., Hadoop) and big SQL structured data management solutions (e.g., Apache HIVE, Cloudera Impala, IBM Big SQL and Pivotal HAWQ). Mining bigger data generated by the Internet of Things. Both academic and industry experts agree that creating sustainable business models in the IoT era requires overcoming interoperability and analytics hurdles ([8], [9]). Leading organizations have a tendency to start a standards-war rather than choosing common or open standards. After a brief period of 54

A N A LY T I C S - M A G A Z I N E . O R G

pandemonium, common standards could naturally emerge, or the IoT industry could achieve peace through platforms that are designed to pool data from multiple sides of the market. Unfortunately, most business experts are aware that pooling data from multiple systems is not enough. Bigger, wider and deeper data comes with challenges similar to what firms faced with “big data” years ago. Perhaps we will need to coin a term for “bigger data” created by smart things. IoT starts with smart devices, but it does not end with them. IoT-generated data will require smarter analyses methods. The difference between high- and low-quality decision-making starts with the path that the decision-maker takes (from the preprocessed data to the knowledge). How a decision-maker mines data defines the level of business intelligence in this bigger data era. Mining bigger data will require an adaptive process that includes classification, clustering, association, time series and outlier analyses [10]. Cognitive computing is another growing area to assist these challenges. To summarize, there is simply too much data to mine, analyze and process, yet there will be very little time to make decisions in the IoT era, which leads us to the next section on ... Data filtering and cloud computing. Consider a fast-moving IoT-connected, W W W. I N F O R M S . O R G

driverless car (or any other IoT-enabled smart device) in the near future. IoT-enabled smart devices not only collect large amounts of data through communication technologies, but also generate and swim in the ocean of data flow. What can go wrong when a driverless car swims in the “data ocean” and makes split-second life or death decisions? Obviously, even the most advanced processors will not have time to analyze all of the information. Perhaps smart devices need to filter data in order to make quasi-smart decisions. Established industries such as telecommunication and finance already use data filtering technologies to make faster heuristic conclusions. However, none of those conclusions directly affects a human’s life, but soon smart infrastructure can face legal accountability challenges [11]. The choice between the optimal vs. the quick decision will have severe social implications. Marriage of the IoT and the cloud has the potential to create value for the IoT industry through platform-level synergies [12]. Technology stack and platforms for the Internet of Things can provide a temporary solution. Typically, technological implementation of a connected product requires multiple software and hardware components in a multi-layer stack. Beyond engineering challenges, it would be naïve to expect IoTenabled data to be stagnant. Most probably, it will grow exponentially. When the amount A NA L Y T I C S

of data to be analyzed reaches the capacity of the processing power of cloud servers, filtering technologies may be reconsidered. THE FUTURE OF THE INTERNET OF THINGS Today, millions of devices expose what they see, hear and otherwise sense to the Internet. What happens in the future is certain to be more amazing than what is happening today. Creating the Next Gen IoT will trigger multiple market tornados, redefine global economies and provide room for many new companies. Software and sensors are doing many things much more efficiently, conveniently and cheaply than humans. We talk to our televisions and they listen, thanks to embedded sensors and voice processing chips that can tap into the cloud for corrections. We drive down the road and sensors gather data from our cell phones to measure the flow of traffic. Our cars have mobile apps to unlock them. Health devices send data back to doctors, and wristwatches let us send our pulse to someone else. Of course, businesses and governments need to consider the ramifications of systems that can sense, reason, act and interact for us. We need to solve the security, privacy, reliability and trust issues inherent in a future world where we are constantly surrounded by connectivity and information. We also need to think about how IoT M A Y / J U N E 2 016

INT ERN ET O F T H I N G S solutions can be developed in a way without vulnerability. For example, what sort of disruption can cause traffic lights not to work or an automobile’s computer to be hacked? We need to consider what happens when tasks currently performed by humans can be automated into near invisibility. In addition, we need to think about what it means to be human when ambient intelligence can satisfy our wants and needs before we express them, or before we even know that we have them. Regardless of the obstacles, IoT is becoming part of everyone’s lives, and it is connecting virtual and physical worlds. While IoT solutions are enabling organizations to create B2B value more globally, optimize operations, create innovative business models and align organizations, interoperability, analytics and security are still very challenging. There are incredible upsides to such a future, but there are also drawbacks. Let us make sure we go there with our eyes wide open and plan for the outcomes we want. ❙ Tayfun Keskin (keskin@uw.edu) is an assistant professor of management information systems at the School of Business, University of Washington Bothell. He is co-chair of the Internet of Things special interest group at the International Society of Service Innovation Professionals (ISSIP). Haluk Demirkan (haluk@uw.edu) is a professor of service innovation and business analytics at the Milgard School of Business, University of Washington-Tacoma, and co-founder and board of director for ISSIP. He is a longtime member of INFORMS.

A N A LY T I C S - M A G A Z I N E . O R G

REFERENCES 1. Bradley, J., Barbier, J. and Handler, D., 2013, “Embracing the Internet of Everything to Capture Your Share of $14.4 Trillion,” white paper published by CISCO. Available at http://www. cisco.com/c/dam/en_us/about/ac79/docs/innov/ IoE_Economy.pdf 2. Wellers, D., 2015, “Is this the future of the Internet of Things?” World Economic Forum, Nov. 27. Available at https://www.weforum.org/ agenda/2015/11/is-this-future-of-the-internet-ofthings 3. Bryzek, J., 2014, “Trillion Sensors Movement in Support of Abundance and Internet of Everything,” SensorsCon 2014, Santa Clara, Calif. Available at https://cseweb.ucsd.edu/ classes/sp14/cse291-b/notes/Janusz_Bryzek_ SensorsCon2014.pdf 4. Scott, A., 2016, “The Future is Smart: 8 ways the Internet of things will change the way we live and work,” The Globe and Mail: Report on Business Magazine. Available at http://www. theglobeandmail.com/report-on-business/robmagazine/the-future-is-smart/article24586994/ 5. Stedman, C. 2015, “IoT data analytics spurred on by big data’s expansion,” TechTarget Search Business Analytics. Available at http:// searchbusinessanalytics.techtarget.com/feature/ IoT-data-analytics-spurred-on-by-big-datasexpansion 6. Rhodes, P., 2015, “Build an IoT analytics solution with big data tools,” InfoWorld, Jan 29. Available at http://www.infoworld.com/article/2876247/ application-development/building-an-iotanalytics-solution-with-big-data-tools.html 7. Demirkan, H. and Dal, B., 2014, “Why Do So Many Analytics Projects Still Fail? Key considerations for deep analytics on big data, learning and insights,” Analytics, July-August, pp. 44-52. 8. Bughin, J., Chui, M., and Manyika, J., 2015, “An executive’s guide to the Internet of Things,” McKinsey Quarterly, August 2010. 9. Keskin, T., Tanrısever, F., and Demirkan, H., 2016, “Sustainable business models for the Internet of Things,” ORMS Today, Vol. 43, No. 1. 10. Chen,F., Deng, P., Wan, J., Zhang, D., Vasilakos, A.V., Rong, X., 2015, “Data Mining for the Internet of Things: Literature Review and Challenges,” International Journal of Distributed Sensor Networks. Available at http://www.hindawi.com/ journals/ijdsn/2015/431047/ 11. Whitmore, A., Agarwal, A., and Xu, L.D., 2015, “The Internet of Things: A survey of topics and trends,” Information Systems Frontiers, Vol. 17, No. 2 (April), pp. 261-274. 12. Botta, A., de Donato, W., Persico, V., and Pescape. A., 2016, “Integration of Cloud Computing and Internet of Things: A Survey,” Future Generation Computer Systems-the International Journal of Science,” Vol. 56, pp. 684-700.

W W W. I N F O R M S . O R G

http://meetings.informs.org/analytics2016/current-sponsors/

B IG DATA A NA LY T I C S

Cloud-supported machine learning services Exploring and comparing the potential of big data analytics on selected cloud providers’ platforms.

BY (L-R) LAKSHMI D. BASKAR, NEIL LOBO, PRAVEEN ANANTH AND PALANIAPPA KRISHNAN

ig data analytics and cloud computing are two of the most prominent technologies of high interest to business organizations and researchers across the globe today. Increasing market penetration of highly affordable sensors, smart gadgets and connected devices have resulted in a continuous, massive quantity of heterogeneous data, i.e., big data [1]. Harnessing its potential has become the key to competitive 58

A N A LY T I C S - M A G A Z I N E . O R G

advantage for data-driven businesses. The complex, fast-streaming “big data” flows from many areas, including social media, the Internet of Things (IoT) and finance as companies seek to gain richer insight in a timely and cost-efficient manner. This phenomenon, characterized by the “3Vs” (volume, velocity and variety), produces information far beyond the processing capability of conventional tools. Building powerful computational infrastructure to W W W. I N F O R M S . O R G

warehouse and make sense of all the data requires large investments and poses practical limitations. Cloud computing offers an attractive alternative. CLOUD COMPUTING: A NEW PARADIGM Cloud computing and Cloud computing and big data are gaining significance and value in small big data are clearly gaining and large enterprises. significance and value in both small and â&#x20AC;˘ Software as a Service (SaaS) has large enterprises [2], [3]. Cloud service applications operating in the cloud as providers such as Google, Microsoft a service to the end users. and Amazon are renting some of their well-managed, massive, worldwide data While there is a certain level of concern centers to developers and companies on privacy and security issues, cloud as a that require large computing power and delivery mechanism has grown in interest storage resources to run their applications. among researchers and enterprises. Clouds enable users to have easy access MACHINE LEARNING IN THE CLOUD to large distributed resources in an ondemand fashion, similar to utilities, thereby Traditional analytical approaches are decreasing the overall cost of system insufficient to analyze big data as the administration and efficient management emphasis is on comprehensive analysis of resources. Cloud delivery models can of highly scalable, unstructured data capbe categorized as: tured in real time. Machine learning (ML), â&#x20AC;˘ Infrastructure as a Service (IaaS), one of the solutions to address this chalthe most basic layer allows users lenge, enables a system to automatically to set up virtualized hardware and learn patterns from data that can be levsoftware resources in the data eraged in future predictions. Extracting center. information can be achieved via superâ&#x20AC;˘ Platform as a Service (PaaS) offers vised, unsupervised or reinforced learndevelopers with the tools and ing. Learning big data with conventional environment to build and deploy their ML algorithms becomes computationally applications in the cloud. expensive and sometimes intractable for A NA L Y T I C S

M A Y / J U N E 2 016

CLOU D COM P U T I N G which cloud computing offers a practical alternative [4], [5], [6]. Given that data and software already reside in a cloud, bringing over the computation logic is a natural progression, thereby reducing overall input/output overhead and monetary cost. CLOUD-HOSTED ML SERVICE By exploring the potential of big data analytics on major cloud providers’ platforms including Azure, AWS, Google and IBM, we can lay out the workflow automation of data science for big data analytics in these clouds as shown in Figure 1. 1. Set business objective or research question. “A business problem well-stated is a problem half solved.” Formulating an appropriate business question in accordance to a company’s goals and market trend is of utmost importance. Case study: Predicting click-throughrate (CTR) is a very essential learning problem in the online advertising industry.

It is an evaluation metric often used for sponsored search advertising and realtime bidding auctions. For our case study, we used a CTR prediction data set obtained from ShareThis social network traffic. ShareThis network traffic captures terabytes of consumer social engagement data across three million publishers and processes the data to make it actionable for businesses. 2. Collect & store. In this process, data scientists decide on the number of features to be gathered for analysis. Based on business requirements, the collected data will be stored in highly scalable data storage that provides a secure way to retrieve information. For this discussion, we cover computing and storage services offered by Amazon, Microsoft and Google. Interested readers could also look at services from cloud providers including IBM SmartCloud, Rackspace and many more. Most cloud providers provide a combination of IaaS and PaaS services.

Figure 1: Data science workflow for big data analytics. 60

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

Amazon Web Services (AWS) is one of the largest providers of highly scalable, cost-effective infrastructure platform. In AWS, Amazon’s Elastic Compute Cloud (EC2) provides users an effortless way to configure the computing capacity using either pre-configured or customizable Amazon Machine Images. Amazon’s Simple Storage Service (S3) offers a simple, secure, inexpensive and scalable data storage. Windows Azure, offered by Microsoft, allows end-users to host, store, scale and run Web applications on a network

A NA L Y T I C S

of Microsoft datacenters. For computing services, Azure supports creating virtual machines similar to Amazon’s EC2. Azure’s storage service provides various forms of persistent storage such as tables, blobs and queues that can be accessed using interfaces. Google Cloud Platform integrates many suites to store and analyze data on Google’s infrastructure. Google Compute Engine allows users to set up virtual machines hosted by Google. Google’s BigQuery, a highly scalable database, can accommodate and retrieve multi-terabytes

M A Y / J U N E 2 016

CLOU D COM P U T I N G of data. Data can be directly imported to BigQuery through Google Cloud Storage, an advanced storage service on Google’s infrastructure. Case study: Our CTR data set is CSV formatted, approx. 8.9 GB in size, with more than 71 million rows and 25 features including click flag (click/nonclick), click behaviors, timestamps, impression details, campaign, ad groups, URL, IP address, device details, user agents, mobile device, verticals, image details, etc. At ShareThis, the data

is stored in BigQuery, a Google Cloudhosted analytical database. 3. Data preprocessing and ML service. Real-world big data is generally unstructured, incomplete, noisy and inconsistent. Data preprocessing is an essential step to improve data quality, resulting in better model building and scoring. Preprocessing data includes the following techniques that are not mutually exclusive: data cleaning, data transformation and data reduction.

Table 1: Data processing and ML steps for AML and Azure ML. Service/ Data limit Attributes (FreeTier) Azure ML 10GB

Data Source

Data Wrangling ML Algorithms

Upload text file locally, HiveQL, SQL databases, Azure Blob Storage, Web URL

• Attribute distribution • Descriptive Statistics • Feature selection, Learning by counts, Normalize • Handles Missing Value • Data Split customizable

Storage: Pay for large data sets

AML

100GB

AWS S3, AWS RDS, Redshift Storage: Pay for large data sets

Data Insights: • Target and Invalid values Distributions • Attributes distribution by data type • Customizable data split • API - feature processing

• Supervised Binary and multiclass classification, Regression, Decision Tree, Forest, Boosted, Bayes, SVM, Neural Network • Unsupervised: K-Means • Anomaly detection • Text Analytics • Binary and multi-class Classification • Regression

A N A LY T I C S - M A G A Z I N E . O R G

Interpretability Versioning & Pricing ROC, AUC, Runs are Accuracy, cached Precision Recall, F1-Score, Model building Confusion $1per hour matrix, Root Mean Squared Application Error integration with API $2 per computing hour

Evaluate Model: AUC, Accuracy, Precision and Recall, F1-Score, Root Mean Squared Error

Other Language Support R, Python, SQL

Runs are cached

No*

Model building: $0.42 per hour

* Uses In-Built Tools

Predictions Batch: $0.10 per 1000 predictions Real-time: $0.0001 per prediction

W W W. I N F O R M S . O R G

GREAT STORIES CREATE LASTING IMPRESSIONS... YOURS SHOULD TOO.

Learn how to become a compelling data storyteller. Register today for ESSENTIAL PRACTICE SKILLS FOR HIGH-IMPACT ANALYTICS PROJECTS June 16−17 | 8:30am-4:00pm University of Denver Denver, Colorado

CLOU D COM P U T I N G Preprocessing is followed by model building and scoring, which involves selecting appropriate algorithms that enable the timely processing of largescale data for new business opportunities. As discussed earlier ML is time-consuming, and for big data analytics, cloud providers offer ML as a service to build models and deploy in production. Amazon Machine Learning (AML) [7] is a platform for users of all skill levels to deploy machine learning for making predictions at scale. Similarly, Microsoft introduced a fully managed ML as a Web service on Azure for model building and deployment. The Azure ML [8] service is primarily built and evaluated in a crossplatform, browser-based development tool, â&#x20AC;&#x153;Azure ML Studioâ&#x20AC;? â&#x20AC;&#x201C; an intuitive, user-friendly, drag-and-drop, collaborative Web interface with zero-installation. We will walk through data processing and ML steps in both AML and Azure ML using a case study and Table. 1. Case study: We used Azure ML and AML console to evaluate their respective services with data stored on their respective storage services. Using our domain expertise in CTR, we performed both manual and automatic methods to eliminate redundant and irrelevant features. R script and Azure ML were used to produce data summaries and

A N A LY T I C S - M A G A Z I N E . O R G

Figure 2a: Performance analysis.

Figure 2b: Computational time analysis. visualizations of feature distribution and perform initial subset selection. Missing data was handled using sample mean/ mode substitutions. Training data was obtained using a 70 percent split on data. After the training data upload, we chose to modify schema, ignore unnecessary columns, etc. Appropriate learning models were selected to build

W W W. I N F O R M S . O R G

the model on a training data set. The model was validated with the 30 percent of remaining data (test data). Results were interpreted using AUC metrics. Figure 3: Brief summary of ML service platform. Experiments and results: For this as a model in both the services. ML serexperiment, we sampled our data set vices provided the performance (AUC) using 10 percent, 25 percent and 50 based on evaluation of the test data set percent of the complete data set. For on the trained model. As shown in Figure example, for the 25 percent balanced 2a and 2b, both Azure and AML offered data set, every fourth row of the data was scalability and robustness for increasing taken. We analyzed ML services on the traffic. AML scored higher AUCs when following dimensions: compared to Azure ML; however, Azure â&#x20AC;˘ Scalability: ability of the system to ML was much faster in building and valiefficiently handle increasing amounts dating the model. of data by making use of additional Other services: As shown in Figure 3, resources with less overhead each platform offers distinguishing advancomplexity. tages over other services. â&#x20AC;˘ Robustness: ability to process large Google Prediction API (application data and build models in real time programming Interface) [9], primarwith low time complexity. ily intended for developers, provides â&#x20AC;˘ Performance of the system: a platform to insert and train data, and usually measured in terms of area generate prediction models through an under the curve (AUC), determines API with JSON (JavaScript object notahow accurate the model predictions tion) responses. Google Prediction API are, with a perfect model scoring an requires upload of a clean data set that AUC of 1 (100 percent). follows certain conventions and supports real-time evaluations via API method Our observed results in the computacalls. The Prediction API provides asyntional time and performance experiments chronous cloud-based training, automatwere based on default parameter setic model selection and tuning, and the tings. Binary classification was selected ability to add training data on the fly.

A NA L Y T I C S

M A Y / J U N E 2 016

CLOU D COM P U T I N G IBM Watson Analytics service [10] offers data insights, data visualizations and predictive ability of the features. Unlike AML and Azure ML, Watson Analytics aims to be a visualization and smart discovery solution available on the cloud, to guide data exploration and create dashboards and infographics. This service, in our opinion, is suited for data exploration, feature analysis and subset selection process. Deployment: Azure ML, AML and Google Prediction API provide deployment of the machine-learning models via API endpoints or as Web services. IBM Watson does not provide creation of machine learning models for deployment. CONCLUSIONS AND FUTURE WORK Today there is a growing interest in cloud delivery, and many organizations are evolving to support cloud services. As cloud-supported big data analytics is not a one-size-fits-all solution, through this survey, we have provided a quantitative review to help businesses and researchers make use of these platforms effectively. Topics for future work include optimizing parameter tuning, deploying in a production setup and analyzing API implementations. ❙ Lakshmi Dhevi Baskar (lakshmibaskar@gmail.com) is a data scientist and Neil Lobo (neillobo@outlook. com) is a software engineer at ShareThis in Palo Alto,

A N A LY T I C S - M A G A Z I N E . O R G

Calif. Praveen Ananth (praveenananthg@gmail.com) is an incubator strategist with eBay Mobile Labs in San Francisco. Palaniappa Krishnan (baba@udel. edu) is an associate professor of applied economics and statistics at the University of Delaware in Newark, Del., and a member of INFORMS.

GLOSSARY OF DEFINITIONS CTR: click-through-rate metric Scalable storage: Handles increasing amounts of data. Persistent data store: Retains data even when the machine is powered off or any failure occurs. API: Application programming interface. ROC: receiver operating characteristic AUC: area under the curve JSON: JavaScript object notation

REFERENCES 1. R. L .Villars, C. W. Olofson, and M. Eastwood, 2011, “Big data: what it is and why you should care,” white paper, IDC. 2. Q. Zhang, L. Cheng, and R. Boutaba, 2010, “Cloud computing: state-of-the-art and research challenges,” Journal of Internet Services and Applications, Vol. 1, No. 1, pp. 7-18, May 2010. 3. D. Talia,2013, “Clouds for scalable big data analytics,” Computer, Vol. 46, No. 5, pp. 98-101, May 2013. 4. I. T. Hashem, I. Yaqoob, N. B. Anuar, S. Mokhtar, A. Gani, and S. U. Khan, 2015, “The rise of big data on cloud computing: Review and open research issues,” Information Systems, Vol. 47, pp. 98-115, January 2015. 5. M. D. Assunção, R. N. Calheiros, S. Bianchi, M.A.S. Netto, and R. Buyya,2015, “Big Data computing and clouds: Trends and future directions,” Journal of Parallel and Distributed Computing, Vol. 79–80, pp. 3-15, May 2015. 6. H. Hu, Y. Wen, T. Chua, and X. Li, 2014, “Toward Scalable Systems for Big Data Analytics: A Technology Tutorial,” IEEE Access, Vol. 2, pp. 652-687, June 2014. 7. https://aws.amazon.com/machine-learning/. 8. https://azure.microsoft.com/. 9. https://cloud.google.com/predictions/. 10. https://community.watsonanalytics.com/.

W W W. I N F O R M S . O R G

To-Do-List Go to the gym Begin to eat healthy

EXPAND YOUR NETWORK!

N I O J S M R O F IN Is the largest association for analytics in the center of your professional network? It should be. • INFORMS Allows You to Network With Your Professional Peers and Others Who Share Your Interests • INFORMS Connect, the New Member-only, Online Community Lets You Network with your colleagues • INFORMS Provides Unsurpassed Networking Opportunities Available in INFORMS Communities and at Meetings • INFORMS Offers Certification for Analytics Professionals • INFORMS Helps You Take Leadership Roles to Help Build your Professional Profile • INFORMS Career Center Provides You with the Industry's Leading Job Board

Join Online Today! http://join.informs.org

CO RPO RATE P RO F I LE

Steelcase Inc. Advanced Analytics team helps global company unlock human promise by creating great work experiences, wherever work happens.

BY TIM MERKLE

or more than 100 years, Steelcase Inc. has helped create great experiences for the world’s leading organizations, across industries, through its family of brands including Steelcase, Coalesse, Designtex, PolyVision and Turnstone. Together, they offer a comprehensive portfolio of architecture, furniture and technology products and services designed to unlock human promise and support social, economic and environmental sustainability. Steelcase is

A N A LY T I C S - M A G A Z I N E . O R G

globally accessible through a network of channels, including more than 800 dealer locations. Steelcase is a global, industryleading and publicly traded company with fiscal 2015 revenue of $3.1 billion. As the global leader in furnishing workplace environments, Steelcase has a unique ability to satisfy the needs of its customers anywhere in the world, wherever they work. The company’s products and services are inspired by more than 100 years of insight gained from serving the world’s leading organizations.

W W W. I N F O R M S . O R G

Steelcase WorkLife Centers around the world offer an array of applications and floor-plate designs (WorkLife Chicago). Headquartered in Grand Rapids, Mich., Steelcase’s global reach provides a broad context for understanding emerging issues and what it means to be a responsible corporate citizen. In a fast-changing world that’s more interdependent every day, Steelcase provides insights, products and services that help people do their best work. Unlocking human promise is the fundamental principle on which the company was founded in 1912, and it remains the focus today. Steelcase began as The Metal Office Furniture Company in Grand Rapids and received its first patent in 1914 for a steel wastebasket – a major innovation at a time when straw wastebaskets were

A NA L Y T I C S

a serious office fire hazard. That led to metal desks, and Steelcase has led the way with product and service innovations ever since. Across borders, time zones and languages, Steelcase’s global network of capabilities – unmatched in its industry – gives the company roots and reach to provide products to its customers all over the world. From environmental leadership to supporting diversity to strengthening communities, and spearheading efforts to improve urban schools, Steelcase invests in helping people realize possibilities. Through its own efforts and collaboration with others, Steelcase is recognized as a company that creates

M A Y / J U N E 2 016

CO RPO RATE P RO F I LE

social, economic and environmentally sustainable value. As a global company, Steelcase is dedicated to creating relationships with diverse businesses to create a better, more sustainable world. BUILDING AN ADVANCED ANALYTICS TEAM In 2013, Steelcase formalized its commitment to become data-driven by establishing a new Advanced Analytics team. The team was organized not just to solve problems, but also to think through the solutions differently, specifically with data. As Steelcase enters its

second century of operations, it strives to learn more from its data and leverage it to its fullest potential. To transform data to insights and drive action effectively, the Advanced Analytics team must approach data with greater care, and think past the initial transactional use of data and realize the greater value on “Day 2” and beyond. New people, processes and technology come together to bring a broad spectrum of methodologies to solve problems large and small. The Advanced Analytics team has partnered with the company’s Business Intelligence Competency Center to provide end-to-end descriptive, predictive

In a protected “incubator” environment, teams are free to test and develop fragile ideas, accelerating iteration and innovation (Innovation Center, Grand Rapids, Mich.). 70

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

and prescriptive analytics to support the varying needs of the organization. The Advanced Analytics team is comprised of analytics professionals with robust experience applying statistics, economics, computational science and operations research (O.R.). As a composite of four sub-groups (Statistical Learning, Computational Intelligence & Machine Learning, Operations Research and Architecture), the Advanced Analytics team is designed for agility; it handles a wide array of problems by varying the composition of the project teams. Early in the team’s history, its executive sponsor drove engagements with key leaders throughout the business to assess initial analytics opportunities. This led to a large list of projects for the new team to undertake. The rapid influx of projects led to the growth of the Advanced Analytics team and a need for project governance. To that end, guidance committees representing every major business function were organized to develop, prioritize and support execution of projects within the Advanced Analytics portfolio.

the trust and confidence in analytics. As an internal consulting group, the team has become a hub of cross-functional information, plugged into the company’s large global network at various levels. The team’s portfolio is a diverse collection of projects from Marketing, Sales, Finance, Sustainability, Procurement, Manufacturing, Logistics, Quality and Information Technology. The team’s analysts often find themselves working with multiple groups simultaneously on very different problems. In order to maneuver the complexity of Steelcase and meet customer expectations, the team relies heavily on its business partners and sponsoring organizations. Learning from the subject matter experts and studying the systems that create, modify and store their data is an important part of the team’s process. The team does not operate unilaterally; rather, it exposes its business partners directly to the new analytical capabilities and methodologies employed by the team. This is critical to the customer experience, trust in solutions and growing a data-driven mindset.

BEYOND PROJECTS: BUILDING PARTNERSHIPS

DELIVERING INSIGHT

Every project the Advanced Analytics team undertakes is an opportunity to increase revenues or deliver cost savings; moreover, it is an opportunity to deepen A NA L Y T I C S

With a portfolio spread across major business functions, the Advanced Analytics team strives to find “one-to-many” opportunities; that is, solutions where the methodology, model or results can be M A Y / J U N E 2 016

CO RPO RATE P RO F I LE

utilized by multiple stakeholders. We also want to help mentor and grow analytics capabilities throughout the global team. Certain areas of Steelcase were more prepared than others to leverage traditional analytics because their people, leaders and/or their data were more progressive and prepared. A few examples of early applications: Optimizing networks: As Steelcase modernizes its industrial system, the team has found multiple opportunities to support network redesign. Efficient inbound and outbound transportation is critical to deliver product and ensure the customer experience is maximized. To this end, Advanced Analytics supported multiple efforts to study and optimize distribution networks. The most recent undertaken by the O.R. group was to rapidly formulate a capacitated optimization model to find optimal distribution center locations. A build-to-order business model resulted in some interesting constraints that not only challenged the team, but helped the team better understand operations in all three operating regions simultaneously. The resulting optimization model was designed to handle discrete inputs, but it could be modified for stochastic inputs in the future. We ran 300+ scenarios to identify the optimal locations for distribution centers in the current and proposed 72

A N A LY T I C S - M A G A Z I N E . O R G

industrial system, thus providing a solid business case for implementation. Right-sizing energy: As a passionate advocate and leading organization in environmental sustainability, Steelcase is very much interested in protecting the environment through reducing our carbon footprint and energy waste. The Advanced Analytics team partnered with the companyâ&#x20AC;&#x2122;s sustainability and procurement groups to study historical electricity consumption of the Grand Rapids facilities. The result yielded a new purchase strategy for electricity blocks. Leveraging optimization, the team not only rightsized purchased electricity blocks to handle Michigan seasonal weather and manufacturing swells, but through mapping the problem we developed a vendor management strategy that yielded even more value. Through the use of data we challenged the vendor to provide us more flexibility shifting from annual block purchases to a more dynamic model. Improved forecasting: Long histories and structured hierarchies were ripe for applying time-series forecasting to improve or provide forecasts to areas of the business previously without. Over the last 24 months, the Advanced Analytics team has kicked off multiple strategic projects to provide higher frequency, automated, hierarchical forecasts across Steelcase. The team has established a robust W W W. I N F O R M S . O R G

analytics ecosystem to provide tens of thousands of statistical forecasts to support various stakeholders. Additionally, we are currently piloting hybrid forecasts, leveraging system dynamic simulation and other non-traditional approaches for areas with less than desired histories. Clearer demand signals and the ability to navigate hierarchies are changing Award-winning Brody WorkLounge is designed for your body, brain and work. the way our business partners operate drive math and science deeper into the with data. core business processes throughout. Near future: As our current partnerThe team is working closely with key ships shift into business implementation, executive sponsors to diversify its portwe keep a weathered eye on the future. folio of solutions and grow descriptive, A few areas will get special focus from predictive and prescriptive analytics cathe O.R. group. We are actively develpabilities throughout an exceptionally oping projects that will apply stochastic talented global workforce. The strength processes, Markov chains, dynamic simof Steelcaseâ&#x20AC;&#x2122;s culture is the passion and ulation, optimization, Industrial Internet of resolve for solving wicked problems; Things and game theory. Leveraging the the Advanced Analytics team is excited depth and breadth of the entire Advanced to offer its business partners new and Analytics team and strong business partexpanding capabilities to translate data ners, the opportunities for complex eninto actionable insight. â?&#x2122; semble solutions are abundant. CONTINUED GROWTH Building on Steelcaseâ&#x20AC;&#x2122;s tradition and commitment to excellence, the Advanced Analytics team continues to A NA L Y T I C S

Tim Merkle is the manager of Advanced Analytics at Steelcase, leading the analytics team and program to improve data-driven capabilities throughout the global enterprise. He is a former Marine Corps officer, combat veteran, current Steelcase representative to the INFORMS Roundtable and a member of INFORMS.

M A Y / J U N E 2 016

CO N FERE N C E P R E V I E W

INFORMS International Conference set for Hawaii The 2016 INFORMS International Conference will take place on June 12-15 at the Hilton Waikoloa Village Resort on the Kohala Coast in Waikoloa, Hawaii. The invited tracks will span the full range of emerging topics from global supply chains to social networks, and all aspects Hilton Waikoloa Village Resort, site of the 2016 International Conference. in between. The informative program is designed to educate attendees on primarily focused on networking with colcurrent advances that are at the cutting leagues and other international attendees. edge of the field anywhere in the world. Conference Chair Saif Benjaafar and Through a series of diverse speakers, the rest of the conference committee will panels, tutorials and structured network- host a welcome reception on Sunday ing, the conference will offer attendees a evening. The Tuesday evening general forum for rich intellectual exchange on a reception will feature an authentic Hawaibroad range of OR/MS applications. ian Luau that is sure to be a feast for all Gang Yu will deliver the plenary talk of your senses. Men and women in ornate on Sunday, June 12. Yu is the co-found- costumes will perform a festival drum er and executive chairman of New Peak dance from the islands of Tahiti. A tradiGroup. Prior to founding New Peak tional Polynesian luau feast will be served Group, he was the co-founder and chair- as well. This event is included in the conman of Yihaodian, a leading e-commerce ference registration fee. Early registration company in China. rates will expire on May 20. In addition to the technical tracks, the For more information or to register, visit program also includes two receptions meetings.informs.org/2016international. â?&#x2122;

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

2016 INTERNATIONAL CONFERENCE HAWAII June 12–15, 2016 Hilton Waikoloa Village

NETWORK AND SHARE YOUR EXPERIENCES ENHANCED WITH AN AUTHENTIC HAWAIIAN LUAU BY THE BEACH! Hawaii 2016 delivers an impressive lineup of keynote and plenary speakers along with 340 contributed and 500 invited papers on emerging topics in tracks ranging from global supply chains to social networks. Listen to an impressive array of speakers presenting their latest research on: • • • • • • •

OR in Medicine Entrepreneurship & Innovation Applied Probability Service Operations Big Data Analytics Inventory Systems Cloud Computing

• • • • • • •

Practice-Focused Operations Global Supply Chains Mechanism Design & Game Theory Business Strategy Robust Optimization Scheduling and Project Management and much more

Take this opportunity to network and collaborate with colleagues across the globe from both academia and industry.

2016 INTERNATIONAL

HAW II

PLENARY SPEAKER: Gang Yu

Executive Chairman & Co-Founder of New Peak Group

Thomas L. Magnanti

Founding President of the Singapore University of Technology and Design

FIVE- M IN U T E A N A LYST

Border walls Before we can talk about the merits of a wall, we should try to understand the size of the problem.

BY HARRISON SCHRAMM

This time, I’m going to take a look at a subject in current events – the proposed construction of an antiimmigration wall along the U.S. border with Mexico. If you have read this column a few times, you know that it’s all about analysis, and seamlessly moves from the frivolous, such as popular television and movies, to the very serious, such as vaccinations and welfare policy. While analysis is the focus, the requirements on both data and rigor are much higher when you are working on current events. This article turned out to be a lot more complicated than I initially expected it to be. For those who haven’t been following, the “Border Wall” is a proposed barrier running the length of the U.S. southern border with Mexico. This is one of the key elements of Donald Trump’s platform running for the Republican nomination for president. At the time of this writing (April 1, 2016), Mr. Trump is the frontrunner for the Republican nomination. For the remainder of this article, we will assume that the goal is to prevent illegal immigration into the United States, although even that is contested [1]. Before we can talk about the merits of a wall, we should try to understand the size of the problem. The

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

difficulty with estimating immigration – and illegal activity in general – is that we have data on those who were apprehended, but we do not have a true estimate of those who successfully crossed. We know how many unsuccessful attempts were made. There are some very interesting techniques to attempt to discern the true size of the problem using partial data [2], but such an attempt is beyond the scope of the current effort. For southern border apprehensions by year, see Figure 1. When I sat down to write this, my plan was to compare the cost of the wall with the cost of current Customs and Border Protection budgets to determine which was more effective. However, after seeing Figure 1, I got derailed; my analytic voice was screaming: Wait a minute – why are apprehensions decreasing? There are at least three reasons why apprehensions at the border might be decreasing. They are: 1. Enforcement efforts have decreased; if you aren’t trying to catch border crossers, you will catch fewer. 2. Border crossers have gotten better; if the people running the border have gotten better at it, you will catch fewer. 3. Enforcement has increased and fewer attempts have been made; if you try harder to catch border crossers and this is known, people A NA L Y T I C S

Figure 1: Southern border apprehensions FY 2000-2016 [3]. The number of immigrants apprehended at the southern border has been decreasing by approximately 75,000 per year (R2 = .82, p = 0). will assess their probability of successfully crossing as low and not attempt it. Evaluating (2) or (3) is beyond the scope of this article, but there is readily available data to help us think about (1). Let’s add the annual CBP budget to our existing graph. Figure 2 does not support the hypothesis that the U.S. government gave up enforcing the border. Another way to look at this problem is to consider

Figure 2: Southwest border apprehensions (blue line) vs. total Customs and Border Protection budget (red line, “then-year” dollars) [4]. M A Y / J U N E 2 016

FIVE- M IN U T E A N A LYST

Figure 3: Southern border apprehensions vs. unemployment rates. Source: United States (Bureau of Labor Statistics [5]) and Mexico (World Bank [6]).

unemployment rates in both Mexico and the United States. See Figure 3. Earlier, in Figure 1, we regressed year vs. apprehensions and concluded that immigration was decreasing, but we did not consider any of the underlying causes. Now, we perform a regression of unemployment rates vs. apprehensions, and the results are interesting. The p-value suggests significance at .003, and the R-square is .49. Why would I prefer a model with a lower R-square? Because it is tied

to a variable of interest. Recall that immigration is a complicated process and what we have performed here is a very rudimentary look. While I typically do not like to present diagnostic charts in a “polished” setting, the predicted vs. fitted chart for this data is informative: I was planning to end this column by considering the historical and proposed costs of the “Border Wall,” but I spent our five minutes together on these charts instead. The historical and proposed costs may be the topic of a future column if it is still part of the national discussion. ❙ Harrison Schramm (harrison.schramm@gmail. com), CAP, PStat, is an operations research professional with CANA Advisors, LLC. He is based out of California’s Central Coast. He has been a member of INFORMS since 2006. REFERENCES 1. For example, see: http://www.rand.org/pubs/ periodicals/rand-review/issues/2012/fall/ leadership/immigration-reform.html and http://thehill.com/blogs/congress-blog/ foreign-policy/203984-illegal-immigrantsbenefit-the-us-economy 2. Shearer, Robert, “Operations Analysis in Iraq: Sifting through the Fog of War,” Military Operations Research, Vol. 16 No. 2, see p. 68. 3. https://www.cbp.gov/sites/default/files/ documents/BP Total Apps%2C%20 Mexico%2C%20OTM%20FY2000-FY2015.pdf “Southwest Border” Retrieved March 25, 2016. 4. https://www.cbp.gov/sites/default/files/ documents/BP Budget History 1990-2015.pdf 5. http://data.bls.gov/timeseries/LNS14000000

Figure 4: U.S. unemployment rate vs. immigration line fit plot. 78

A N A LY T I C S - M A G A Z I N E . O R G

6. http://data.worldbank.org/country/mexico

W W W. I N F O R M S . O R G

2016 Franz Edelman Winner

2016 INFORMS Prize Winner

UPS

General Motors

“UPS On-Road Integrated Optimization and Navigation (ORION) Project”

2016 UPS George D. Smith Prize Winner Heinz College at Carnegie Mellon University

2016 Daniel H. Wagner Prize Winner Georgia Institute of Technology, Emory University, and Centers for Disease Control and Prevention “Machine Learning Framework for Predicting Vaccine Immunogenicity”

Awarded at the 2016 Edelman Gala in Orlando, Florida For more information on the conference, visit http://meetings.informs.org/analytics2016

THIN K IN G A N A LY T I CA LLY

Cell towers Figure 1: How many cell towers are needed? As the head of analytics for a cell phone company, you have been asked to optimize the location of cell towers in a new area where your company wants to provide service. The new area is made up of several neighborhoods. Each neighborhood is represented by a black house icon in the accompanying image (Figure 1). A cell tower can be placed on any square (including squares with or without a neighborhood). Once placed, a cell tower provides service to nine squares (the eight adjacent squares surrounding it and the one it sits on). For example, if you placed a cell tower in B2, it would provide service to A1, B1, C1, A2, B2, C2, A3, B3 and C3. The company recognizes that it may not be worthwhile to cover all neighborhoods, so it has instructed you that it needs to cover only 70 percent of the neighborhoods in the new area. Each cell tower is expensive to construct and maintain, so it is in your best interest to only use the minimum number of cell towers. Question: What is the minimum number of cell towers needed to provide service to at least 70 percent of the neighborhoods? Send your answer to puzzlor@gmail.com by July 15. The winner, chosen randomly from correct answers, will receive a $25 Amazon Gift Card. Past questions and answers can be found at puzzlor.com. â?&#x2122;

BY JOHN TOCZEK

John Toczek is the assistant vice president of Predictive Modeling at Chubb in the Decision Analytics and Predictive Modeling department. He earned his BSc. in chemical engineering at Drexel University (1996) and his MSc. in operations research from Virginia Commonwealth University (2005). He is a member of INFORMS.

A N A LY T I C S - M A G A Z I N E . O R G

W W W. I N F O R M S . O R G

OPTIMIZATION GENERAL ALGEBRAIC MODELING SYSTEM

GAMS-related Courses and Workshops Whether you are new to GAMS or already an experienced user looking to deepen or expand your knowledge in a certain area - take a look at our diverse list of GAMS related courses. Learn advanced, state-of-the-art techniques in a focused and interruption-free setting using the professional's choice in modeling software - GAMS. Domain experts will be teaching the following courses at locations worldwide. Courses in 2016 include: May

Practical General Equilibrium Modeling with GAMS Energy and Environmental CGE Modeling with GAMS

June

GLOBE Model Course: Global General Equilibrium Modeling with GAMS STAGE Model Course: Single Country General Equilibrium Modeling with GAMS Agro-Economic Modeling with GAMS Practical General Equilibrium Modeling with GAMS Energy and Environmental CGE Modeling with GAMS Advanced Techniques in General Equilibrium Modeling with GAMS Overlapping Generation General Equilibrium Modeling with GAMS

August

Basic GAMS Modeling Advanced GAMS Modeling

September

Climate and Trade Policy Analysis with GAMS and MPSGE

November

Modeling and Optimization with GAMS (basic) Modeling and Optimization with GAMS (advanced)

Tak e

be Tu

t ins

youtube.com/ GAMSLessons

at o ur Y ou

ruction

ok lo

Channel

w i

ÂŠ Monkey Business Images / Shutterstock

e id

al v

Online, Online Practical General Equilibrium Modeling Continuous with GAMS Online Advanced Techniques in General Equilibrium Modeling with GAMS