Analytics Section

Analytics, OR and INFORMS – Where the three meet

A recent survey asked the question of INFORMS members – what is the relationship of operations research and analytics? There was disagreement. Is OR a subset of Analytics, or Analytics a subset of OR, or do they merely intersect? Where are they the same, and where do they differ? This is a point of debate and opportunity for INFORMS. The Analytics Section can help define what is Analytics, and how it should be integrated with INFORMS. Please contribute your thoughts on the subject!

OR and Analytics

The Encyclopedia of OR/MS (Gass and Harrs, ed.) defines OR/MS in brief as “the science of operational processes, decision making and management”. The book proceeds to cover hundreds of topics, from A* Algorithms to zero sum games. Wikipedia provides the following definition of OR: Operations research (also referred to as decision science or management science) is an interdisciplinary mathematical science that focuses on the effective use of technology by organizations. Some of the tools used by operational researchers are statistics, optimization, probability theory, queuing theory, game theory, graph theory, decision analysis, mathematical modeling and simulation.

In his book, Competing on Analytics, Tom Davenport defines “analytical competitor” as an organization that uses analytics extensively and systematically to outthink and out execute the competition.

My read on these definitions is that there is not a big difference between analytics and OR/MS, but a difference in their relative emphases. Both areas discuss the use on application of advanced techniques by organizations. However, OR clearly emphasizes the tools and techniques; analytics emphasizes more the analytical process, the tool application and integration, and their impact on organizational competitiveness and efficiency.

Analytics Section and INFORMS

A better question for the current audience might have been what is the relationship of Analytics (and this Section) and INFORMS?

The INFORMS Vision Statement states, “INFORMS is recognized as the premier organization for advancing the profession, practice, and science of operations research and management science.” And its mission statement discusses dissemination of OR/MS excellence and achievements via journals, magazines and other media, and providing OR/MS education and career opportunities.

On our Section website, we state that: The Analytics Section is focused on promoting the use of data-driven analytics and fact-based decision making in practice. The Section recognizes that analytics is seen as both a complete business problem solving and decision making process, and a broad set of analytical methodologies that enable the creation of business value. The Analytics Section promotes the integration of a wide range of analytical techniques and the end-to-end analytics process. Analytic methodologies include descriptive techniques (what happened), predictive techniques (what will happen), and prescriptive techniques (what should happen).

The Section will support activities that illuminate significant innovations and achievements in specific steps and/or in the execution of the process as a whole, where success is defined by the impact on business decision making. The Analytics Section will leverage INFORMS academic integrity to establish INFORMS as the thought leaders of the Analytics movement, focusing on the predictive and prescriptive space.

In a similar way to the definitional differences of OR and analytics, where INFORMS focuses somewhat on advancing the tools and techniques that in many ways empowers the analytics movement, analytics is about creating a vision, selling an idea, seizing the opportunity, selecting of tools and technologies, managing the implementation process, mapping the data requirements and systems integration, and leading and sustaining the organizational change that allow OR to deliver on its promise. As I developed this sentence, I changed it three times; my point is that my understanding is always evolving.

In some ways, analytics is "what we have always done" at INFORMS - develop tools that enable better decisions, but in other ways, is what perhaps we could do better - advocating, explaining, demonstrating, socializing and delivering the benefits of OR to establish its central place in organizations efforts to make better decisions. Thus, Analytics and INFORMS go hand in glove, and both can benefit from the integration of the two – as I imagine you will agree, as a member of the Analytics Section of INFORMS!

Let’s move forward to define Analytics in a way that meets the needs of current INFORMS members and helps position INFORMS for the future. Please weigh in!

Value Proposition for the Analytics Section

In my assessment, the Analytics Section was created for two reasons: 1) To introduce the INFORMS' membership to this very hot new wave in business, and 2) To invite Analytics professionals into the INFORMS organization, whose skills sets align perfectly with the Analytics movement. Thus, we have the confluence of two very different groups coming together in our Section - INFORMS members wanting to learn more about Analytics, and Analytics professionals who want to learn more about INFORMS. We take the value proposition for each in turn.

Analytics Section for the INFORMS member:

The INFORMS member who joins the Analytics Section is taking advantage of a booming trend in business and becoming identified with practitioner terms for what INFORMS is known for. Members of the Section will learn what all the analytics buzz is about -- and help to shape its evolution. The INFORMS member will broaden their understanding of implementing OR methodologies in practice, and will see examples of the integration of multiple tool sets for addressing practical problems. Coming to understand the process of Analytics, and how OR methodologies fit into that process furthers the cause of seeing OR methodologies successfully implemented despite the organizational, financial, change management issues surrounding analytics projects. Many MIS colleagues would suggest that Analytics is simply the evolution of Business Intelligence; the Analytics Section will help to understand the interplay of business intelligence databases and OR methodologies. Simply, the INFORMS member who joins the Analytics Section will help advance the awareness and use of OR methodologies, which is a core mission of INFORMS.

What other values does the Analytics Section bring to the INFORMS member? Please weigh in!

INFORMS Analytics Section for the Analytics Professional:

The Analytics Section is a gateway for you to a large, diverse organization - INFORMS - that has established itself as the premier organization for the advancement of Operations Research and Management Science methodologies. As such, INFORMS brings to the Analytics professional a host of the most advanced predictive (statistical, simulation) and prescriptive (optimization) methodologies available to tackle the toughest Analytical problems. These leading edge techniques are coupled with the academic integrity on which INFORMS was founded and continues to rely. While its academic base is rock solid, the Analytics professional will find many interest groups, sessions and presentations that focus on the application of analytics, as well as a host of tutorials and software demonstrations to show how to put these techniques into action. The ability to meet OR experts and other analytics professionals and discuss analytics issues will help us forge a greater understanding of how analytics techniques are used - and overcome obstacles for why they are not used - in practice. Meeting, socializing and learning with like-minded individuals is at the heart of our Section.

What other values does the Analytics Section bring to the Analytics professional? Please weigh in!

Thanks,
Mike

Comments

valuable post and very well written.

I will put myself in the Operations Research is a subset of Analytics camp.
Analytics is an interdisciplinary field that brings together Computer Scientists who are concerned with storing and retrieving data; Operations Researchers and others data analysis centric professions like Statisticians to help make sense of the data; and you also need domain expertise in things like supply chain, manufacturing, remote sensing, and so on. Operations Research has always been an interdisciplinary field and so certainly has a claim on Analytics.
I think the categorization of areas of Analytics into descriptive, predictive, and prescriptive is a useful way of thinking about the field. From an Operations Research perspective our modeling tool bag is absolutely applicable. But, we also have something to contribute in making sense of large, high-dimensionality data sets.
I see the expansion of outlier detection techniques or, if you prefer, the detection of anomalous data patterns as an important and emerging aspect of Analytics. Our friends in Computer Science and Statistics certainly have something to contribute here, but so does Operations Research. There is simply too much data and not enough people to extract all of the value. We need automated methods of dealing with the large streams of data on a close to real time basis. This is not new. For example, even back in the days of pneumatic control systems in chemical plants alarms would go off if the temperature on a reactor exceeded some level. This kind of rule based method of detecting outliers can be taken further. But, there is so much more that could be done to make use of big data.
Doing complex and detailed modeling, which can take months, has its place. However, in Analytics this is not always needed. Operations Research can contribute in other ways. For example, most Operations Researchers have heard of Little’s Law (Work-In-Process = Throughput * Cycle Time). Using this and other rules-of-thumb can allow one to do useful quick turnaround kinds of queuing analysis rather than doing a full-blown simulation model.

Paul T. Otis, P.E., Ph.D.

In his post, Mike mentions the INFORMS vision statement. Something else to consider now is this announcement that recently went out to INFORMS members:

Board Sets INFORMS Strategic Direction

The Board of Directors, at its recent meeting, approved two overarching strategic goals to guide INFORMS:

1. INFORMS will be recognized as THE leading association for advanced analytics professionals by advancing the practice, research, methods, and applications of advanced analytics, and identifying and serving analytics professionals with products and services they value.

2. INFORMS will provide up-to-date ONLINE systems that enable and encourage access, collaboration, and effective exchange of information, content, services, and benefits to and among our members, potential members, and those interested in knowing more about OR/MS and analytics.

The INFORMS scienceofbetter.org web site says that, "In a nutshell, operations research is the discipline of applying advanced analytical methods to help make better decisions." While we INFORMS veterans appreciate that OR encompasses more than that and that a longer definition of OR can be helpful in some contexts, in other contexts, it is useful to have that concise, easily understood definition of OR. The short definition of OR is convenient in talking with people who ask, "What is OR?"

Similarly, our society might want to agree on a concise definition of analytics. Perhaps something like, "In a nutshell, analytics is using data to drive decision making." Such a definition would not be as accurate as a longer definition, but it might be helpful in summarizing in easily understood terms the most essential essence of analytics: using data to drive decisions.

On a partially different point...

While OR practitioners have conducted data gathering and analysis for a long time, generally, this has been done as a means for applying predictive and prescriptive models. There are many analytics practitioners who have done only the data work without building models which use the data. These folks gather the data, scrub it, dice/slice it, and consider their job done when they provided management with view(s) of the data. Perhaps the analytics section can help these "data guys" to do what they've been doing in a better (more scientific) way... and also provide a path for them to expand their role beyond that of data gatherers/manipulators.

Upcoming Events

2013 INFORMS Conference on Business Analytics and O.R.
April 7-9, 2013
San Antonio, Texas
Conference Website »

Key Contacts

Michael F. Gorman, Ph.D.
937-229-3382
michael.gorman@udayton.edu

Subscribe to Analytics

Upcoming INFORMS Events