Analyst’s Pendulum (Systems)
A Concept From General Systems Theory That All Analysts Should Understand
The field of analytics is, at-best, multidisciplinary. Often, it is simply the wild west of self-taught, self-validated, and self-absorbed silliness. The one time basis of all science and education is now a flea market of unusual techniques and off-the-shelf software solutions in a bottle. And so the pendulum swings…
The pendulum is a great metaphor and symbol. It swings from one side to the other. It has been used to measure and to divine. It in many ways speaks to the systems, structures, and balances that control the field of data science and analytics. So what better brand for a new series, especially one focused on two-sided concepts in analytics taken from areas like systems theory and other disciplines of science?
There are no absolutes… well, maybe just one.
Open vs Closed Systems
Before we get too far, let’s shoot the elephant. I am well aware that very few things are truly binary. Systems types may or may not be one of them. I am also aware that there are no absolutes… in an open system. A closed or isolated system is another thing entirely.
So if you want to argue that there are in fact closed, open, and isolated systems — I will not protest. I will tell you all things are relative… at least within a closed system. A system is either closed or open depending on whether external influences are able to act upon it. You can assign whatever vocabulary you’d like to the relative variants. In the end, you either have a closed system or an open one.
Now, you won’t find too many people in the halls of corporate analytics debating open and closed systems. Fewer still would mention General System Theory or Ludwig von Bertalanffy, I mention them only for those who wish to dig deeper. But that lack of interest does not indicate a lack of importance. Knowing whether a system is opened or closed matters.
Closed systems mirror the laboratory. They allow for absolutes, relative to the system they are contained in. They are a bastion for the scientific method. They are also really great places for concepts like hypothesis testing, experimental design, and most of your classic sciences. They have a certain simplicity and purity. And in most areas of DSA they are hard to come by.
The open system is another story entirely, although so many pseudo-scientists seem to forget it. Open systems destroy absolutes. They destroy controls. They are complicated, noisy, and governed by feedback loops. Open systems require probabilities, assumptions, and triangulation. All living systems and the vast majority of DSA areas of study are open.
So take a lesson from General Systems Theory, recognize the type of system you are analyzing. It will greatly determine the methods, models, and science that you should employ. Analytics is a multidisciplinary science. The key term is discipline. Understanding the structure and constraints of what you are analyzing will only swing the pendulum in your favor. Thanks for reading!
For more series like this:
