From the Director’s Desk: Analytics – Evolutionary Buzzword or an Advanced Science?

Analytics is a hot topic these days. Many of my students, who are otherwise from core engineering disciplines, are now exploring statistical and analytical fields. There is now a wave of prospective students enrolling in dedicated analytics programs bidding to ride the tide that is sweeping across corporate America.

This is not a new phenomenon—processes that enable efficiency have historically been called various names like “six sigma” or “data mining”. The fact remains though, that these models were only rudimentary.

Assume you are to determine daily revenue pattern per slot machine on a casino floor. I am by no means an expert at this kind of modelling, but I can fathom the complexities of trying to model such a problem.

Recognizing that slot machines interact with humans—who are inherently different from each other at various levels—the questions, at a systems level, which would help determine revenue earning pattern for that slot machine might include the following:

1. Proximity of the machine to traffic flow in the casino.
2. Age of the software or “theme” on the slot machine: Themes on slot machines have a life-cycle. They are a factor of age and the number of times it has been played.
3. Economic conditions in that region and how that factors into spending patterns.
4. The day of the month: It is well documented that social security payday tends to create more slot machine revenue than other days in the month.
5. What exogenous effects, such as conventions or special meetings are being held that might draw in players.

If all you considered were past revenue figures for the said slot machine, you’d soon realize that you bypassed relevant contextual data needed for a robust model. It, therefore, requires a deep understanding of the system in question, and a willingness to research explanatory variables that impact system behaviors. This, especially so that we aren’t relegated to overly simplistic models that under-explain behaviors.

Engineering managers often possess this insight. Combined with relevant statistical knowledge, they make for the best resources to create and interpret the models that will forecast behavioral patterns and define the future of the business.