PacificBasinLgI remember most of my undergraduate study like it was yesterday. The freshman introduction to geology was called Earth’s Dynamic Complex Systems, and then there was the year of Geophysics.

It wasn’t until my senior year that a number of things crystallized into a foundation for life. Only 6 of us registered for Dr. John Klasner’s class on Geophysics which had a reputation for being challenging. Part of that foundation was setting a few “academic personal bests,” including earning that A in Geophysics.

This is a story about studying something complex, applying one field to another, and how applied statistics on big data will drive most anything.

Impact of Geophysics and Dynamic Complex Systems

Performing well in geophysics was seminal in developing effective decision making in my career. In this class, we applied the science of capturing data, developing statistics, running distribution calculations on rolling data sets, removing outliers, visualizing the results, and continually iterating until uncertainty was driven out.

You see, geophysics is the applications of quantitative physics to study what you can’t see in the earth. Unlike statics, the earth is dynamic, always changing on its own timeline

Often, Geophysics  is the interpretation of remote-sensed reflection data to model subsurface stratigraphy, structures, fluids, and the subsurface environment. Being subsurface (unseen) and dynamic, it is constantly subject to interpretation, better algorithms, more big data processing, and so on to improve illumination and reduce uncertainty (or increase certainty) in the resulting subsurface model.

Thinking about this brief and limited description, imagine the value geophysics generates to increase certainty for mining, oil and gas, and the environment.

Shifting Focus

My first job was exploration geologist. It wasn’t until after my first that my interest in business, technology, and other aspects of “big data” became a new interest focused on increasing certainty in decision making. Exploration for Pb-Zn deposits in Southeast Missouri was unsuccessful, but it kept me thinking that we were just not looking within the right certainty window with the limited data we had. From a business point of view, this failure shut our project down.

My interest expanded into modeling as a way to understand complex systems and explore opportunity and the goal of better decision making. This has developed into a new field of study and has become in high demand under other names, like risk management, actuarial science, or analytic combinatorics and even has courses in the new Coursera program from Stanford, University of Michigan, Princeton and University of Pennsylvania, titled Model Thinking.

Any way you look at it, model thinking as input to decision making is the application of statistics in order to identify the most probable parameters that frame the decision. It didn’t matter if it was identifying a location for exploratory drilling for an ore deposit, or deciding whether to invest in a new company acquisition to enter a new market. Both are business decisions.

Advent of Big Data and The Cloud

Recently, the term Big Data has been in prominent use by SAS, SAP, EMC, IBM, Hadoop and other companies as part of their marketing pitch. For me, it’s not that new. Geophysicists using mature software technology regularly stitch dozens, if not hundreds of 100+ TB (terabyte) geophysical survey data sets together into multi-PB (petabyte) modeling information systems to study a small area of the earth. Larger data sets in multiples of bytes, like Exabyte and zettabyte and perhaps zottabyte –sized models are conceived.

The cloud we hear about today from Amazon Web Services, Rack Space and Microsoft Azure, is in many ways just a new consumer form of High Performance Computing, a modern derivative of Supercomputing – democratized to broad set of users through the advent of light clients (smart phones, tablets, and of course, laptops).

Why bring up the cloud? Well, with data proliferating at an exponential rate, the processing power required to extract value cannot live on the desktop. And, once this architectural shift is made, the potential of big data is not only revealed, but revealed in a democratized way that involves both access, and creation in the dynamic system.

Pulling It All Together

In pursuing my MBA at the University of Illinois, I studied statistics and decision making as two separate courses. It became apparent that my journey to develop effective data-driven decision making began years ago with the study of geophysics.

I could go on about how this was an area I worked in guiding Autodesk computer aided design products to democratize large design information models supporting engineering and architectural decisions professionals made to build buildings, roads, and bridges. Or using laser scanners and capture-based modeling to support medical decisions for surgeons in the operating room, prosthetic manufacturing and forensic modeling. And more recently creating a new industry category and defining the space for the first predictive and real-time position and movement analytics systems for optimizing location based services for marketing research and asset tracking among other applications in a broad market segmentation model. But these are separate stories.

There are 3 points of wisdom that entrepreneurs, innovators, and future business people need to take away.

  1. Study something complex. It doesn’t matter what it is. for me, it was: The Earth’s Dynamic System. For many it is Mathematics or some other life science. You need to learn something that you can base future thinking on, and complexity is increasing, not decreasing. Computer science alone, is too static, look for something more “dynamic.”
  2. Decisions are made from data, and data is processed with statistics. Statistics is not hard to understand. Some teachers make it difficult, but learning about probabilities, distributions, confidence intervals and certainty are concepts that everyone can understand. This kind of information will be served on smartphones, tablets or other clients – it’s worth knowing where it came from and how it was derived.
  3. Eric Schmidt of Google once said that we create as much data as man created between the beginning of time to the year 2003, in just two days. He said this in 2010. Today, it’s only takes one day to create that much data. This volume of data is wonderful – as it means we have information we can use to drive out uncertainty, and increase certainty that we are making the right decisions and doing the right thing. It is also the grand-unified check and balance system for society (topic for another post!)


Recently cleaning out archives, I came across my bound volume of Geophysics class notes. Dr. Klasners comments on my projects and term papers reminded me where all this began.

The value of studying earth’s dynamics as a complex system, and simplifying an understanding of the subsurface through the systematic processing of big data (geophysics) has helped me grasp other dynamic complex systems.

Business is a complex system that generates big data, just like the petabyte-sized models of an oil field. Democratizing this has great predictive, performance and forensic value.

While abstract, I’ve given you a small point of light into my story. Take it and ponder how this concept of big data has affected your story, and will affect your future.  Leave a comment.

Image Credit: U.S. National Oceanic and Atmospheric Administration, and World Data Center for Geophysics & Marine Geology, National Geophysical Data Center – via Wikimedia in the public domain.

2 Responses to Learn Geophysics: Make Complex Systems – Simple

  1. Erik Larson says:

    I agree, and I see the same patterns. I started by studying astronautics as a dynamic system, especially human factors, i.e. astronaut + spacecraft. It seemed romantic and exciting because it was hard but in the end it was too esoteric and narrow for my tastes. Then I studied leadership in the Air Force, i.e. just the person without the machine, and that was very satisfying but also somewhat limiting since you can only lead as many people as those who follow. Then I focused on business and products to try to get more scale, and although it is fun it didn’t catch my fancy in the same way it did yours, I think because even though many people understand business it is still esoteric and arcane to many others. To use some system dynamics jargon off the top of my head, problems of the zeroth-order (what problem we are solving), first-order (product, budgets and selling) and second-order (marketing, competition, customer resistance to change) were fun and fairly universal, but big-company problems of the third-order (organizational politics) and fourth-order (political resistance to change) were less so, especially without leadership (i.e. some sort of filter to keep the weaker but still present third and fourth-order dynamics from confounding the zero, first and second-order dynamics). So I started my own company, and focused it on the largest and most universal problem I could think of, applying all of my previous learning as you suggest, and so far it feels like it is working. Pretty cool stuff.

    • Andrew Stein says:

      Erik, thank you for this great insight. It’s amazing how the learning path is more common than one would expect. I especially like the way you tiered the zeroth to fourth order elements.

Leave a Reply

Your email address will not be published. Required fields are marked *