I was one of those people who knew I wanted to do "OR" since 11th grade, when I attended a talk on it. I ended up getting my degree in OR (Cornell ORIE, 2003) and then a MEng. in Biological Engineering with applied mathematics bent. I was drawn to OR because of its practicality and generality, it's ability to take an unstructured problem and turn it into a "puzzle" that can actually be solved. To me this is the essence of OR - using mathematics and analytical thinking to structure and solve decision-oriented problems (e.g., applying OR to astrophysics research is probably not a major element...although there are probably a few doing this). I have always considered myself an OR practitioner first. However, I'd like to offer some personal experiences in being an OR person who works almost exclusively with non-OR people.
First, my career has pivoted several times. I started out working for an environmental engineering firm, then an environmental and health science consulting firm, and now an IT firm. I have held the title "Engineer", "Consultant" and now "Data Scientist" and I have a Professional Engineering License and I got my Certified Analytics Professional certification last year. So, I have had to straddle two very different worlds in my 10+ year career.
I am mentioning this because I have had to personally "re-brand" myself several times and so I may be able to offer a unique perspective. In particular, I have followed the OR vs. Analytics vs. Big Data vs. Data Science etc issue for several years, as well as seen the OR/MS community attempt to get some traction with non-specialists. The problem with "Operations Research" is that it conjures up "Tayloristic" images of time-motion studies or manufacturing systems, or it's so nebulous that it draws a complete blank (as I've personally experienced many, many times).
However, whenever I say "Analytics", people start asking me questions that indicate they have a sense of what I can offer. To wit, the following has been asked of me after explaining that I am an "analytics guy":
- Can you help us identify process problems in our wastewater treatment plant?
- What types of information are most important to our customers, based on their database queries?
- How can we select a research portfolio that maximizes expected publication output (normalized to subject, of course) subject to policy-driven constraints/targets on breadth, cost, risk?
- How can we use warranty data on tires to identify production issues?
I have done some(admittedly anecdotal and non-random), sampling of opinions.
When I ask people what they think I can do when I say I do "Analytics", they say...naturally...."Analysis". When I probe deeper, they indicate that they assume its "mathematical", with "statistics and computers". They also seem to agree that I would be someone they would talk to about the goals "optimize", "analyze", "describe", "predict".
Now, when I ask people (e.g., my wife, colleagues, friends), what comes to mind when I say "Operations Research", I get "you study how things are done" or something to that effect.
In other words, OR is too evocative of the study of how things are done and not with the what should be done (prescriptive viewpoint). It also does not seem to convey the degree of analytical sophistication involved.
My most recent position as a "Data Scientist" has given me experience in this new "sexy" (per DJ Patil) job arena. It seems that this umbrella term encompasses an absolutely huge swath of people, and many job ads are asking for some pretty crazy skillsets (hence the term "unicorn" for the hypothetical job candidate who would actually fill the qualifications). Overall, it has a very "machine learning" feel to it, with learning (i.e., model fitting to the rest of the mathematical world!) and prediction being coupled to pretty visualizations. It's all very slick and fun sounding....
However, I wasn't hired for my ability to use Hadoop, or wrangle "big data", or use a particular machine learning package. I was hired because I known how to structure an analysis to answer a question, and how to ask the right questions at the right time to help decision-makers move forward. Again, to me, this is the heart of what OR offers, and Data Science seems to be putting this "old wine" into new bottles, with a heavy dose of computer programming.
Take Aways
To wrap all this up, I offer the following ideas for this forum:
- We need to acknowledge that there are a lot of people out there who use math and data to help organizations make better decisions, besides those who would identify as "OR practitioners".
- By analogy to Engineering: Just as there are Associate Degree-level, junior engineers and highly advanced research engineers (PhD) and academics, there are various levels of sophistication in applying analysis to help decision-makers. We wouldn't deny a junior-level engineer their right to say they are an "engineer" (ignoring the legal sense of it...i.e., PE vs non-PE), and we shouldn't deny a bachelors-level, junior person who uses math and data to help decision-makers the right to be part of our "OR" community, just because they do not work at CPLEX, publish stochastic process papers, or write industrial-scale mathematical programming models.
- People still do not seem to know what OR is or what it can do for them, despite directed marketing efforts. The disconnect between the term "operations research" and what we actually can do is apparently rather large.
- Even in the Data Science community, they use the term "Big Data Analytics" or "Visual Analytics", so even the other "hot" job is using the term Analytics, and in a sense that seems appropriate for what OR folks can do.
- (1)-(4) suggest that, as much as it pains me to say it, we need to consider something more evocative and inclusive than "OR" or even "management science" (hint: it's not data science ;-)
This is why I think that Analytics is a great term for the class of people who use math and data to help people make decisions. It's evocative and brings to the fore the key mindset that all of us have and where we can add value. The details are best left to the resume.
In contrast, I think that Data Science, while also very popular now, is not nearly as good, as I always felt that information theorists probably have the best claim on that term, since they are actually "studying" data/information, per se. The rest of us are using (i.e., analyzing) data and mathematical models for some other purpose. For analytics/OR folks, that purpose is ultimately to support some type of decision.
Therefore, while I personally will always think of myself as being an "OR guy", I like the "big tent" connotation of Analytics as opposed to the obscure or arcane/academic connotation of Operations Research.
That's my two cents. Apologies for the long-ish post.
------------------------------
Michael Beyer, PE CAP
Data Scientist/Analytics Professional/OR Practitioner -- take your pick ;-)
------------------------------