I am delighted to bring you this guest post from Jim Sterne, an international consultant who focuses on measuring the value of the Web as a medium for creating and strengthening customer relationships. He has written eight books on using the Internet for marketing, is the founding president and current chairman of the Digital Analytics Association, produces the eMetrics Summit and sits on Anametrix’s Board of Advisors.
Today's Analogy: Sand
The Scientist ascertains and catalogs the nature of sand. Each grain is unique. Each has a distinctive shape, weight, color and molecular structure. Like snowflakes, no two are alike. The scientist identifies different types of sand and how they might have gotten that way.
Sand grains magnified 110-250 times reveal each grain is unique. Photo copyright Dr. Gary Greenberg.[/caption]
The Data Scientist thinks about Big Sand. Which way is the dune traveling? What can we deduce about its macro movement by the size, shape and direction of the ripples? What new algorithms can we write to help anticipate the flow of sand in different wind or weather?
The sand dunes of Vietnam (source: http://www.travelblog.org/)
The Data Artist takes a sample of the sand and creates a model that does not represent the sand, but the human side of the equation. An artist must understand the raw material well enough to know its limitations and its strengths, and then use that material to create something that communicates to others.
No model is an exact representation of the original. That's what prompted George Box to comment that all models are wrong, but some are useful.
A data model is a representation of the way your business and your customers interact. If the model is good enough, it can be used to see where improvements can be made and to predict outcomes.
These models can get quite complex, to the point of being impractical and unusable; much like an Excel spreadsheet that has too many formulas created by too many people. As George Box also said: overelaboration and over parameterization is often the mark of mediocrity.
Data models also have a time-value boundary. They begin to degrade as soon as they are completed. How companies and customers interact is influenced by a wide variety of forces:
Proximity to payday
Therefore, models must be constantly updated to keep them current.
This is where the Data Artist runs into trouble with his tools. One can build a perfectly lovely sand castle with one's bare hands, but given a shovel, a pail, a trowel and some sticks, the model can get better and better.
We're at a new age in data analytics tools that allow for more options when it comes to data sculpting. The tools are getting more flexible and easier to flex.
Now that the Data Artist can stop spending so much time on making the model work and can spend more time making new models, there is a real opportunity to create and deliver new insights that can stir the imagination of the business decision maker - the consumer of the art.
“Art is not what you see, but what you make others see.”
― Edgar Degas
“Creativity takes courage. ”
― Henri Matisse
“The role of the artist is to ask questions, not answer them.”
― Anton Chekhov