Among a host of other achievements, the multi-talented David Byrne wrote a great book. You may know Byrne as the cofounder and creative force behind the American rock band Talking Heads. His most recent book, How Music Works, is a brilliant and insightful look at “how music works, or doesn’t work,” as a phenomenon that is inextricably part of its environment.
“How Music Works” begins with the revelation of Byrne’s acknowledging his “extremely slow-dawning insight about creation,” to wit: “…[C]ontext largely determines what is written, painted, sculpted, sung, or performed.” He adds that, consciously or unconsciously, we inevitably create in reverse, in a sense working backward to produce work that fits the context into which it is being born. Bach, for example, wrote and played music in the early 1700s for live performances in a reverberant church in Leipzig where he was “Kapellmeister,” or music director, and the music he wrote then sounded particularly good in that venue.
Mozart, writing just a bit later than Bach, was composing for smaller, more intimate rooms where every detail of the music could be heard. That is, at least when members of the audience weren’t making a lot of noise with their dancing. Mozart’s music contained that detail, and was even written for the bigger orchestras that were needed to play above the sound of dancing feet. Context inextricably shapes the creative process, your creative process.
All of that got me thinking about data science (of course). Big data requires technical skill, no question – just ask anyone trying to hire qualified people for a data science team, where the list of math, statistics and computer science tools that need to be in the toolbox is startlingly if not impossibly long. But there is no small amount of art in a successful big data strategy, a generous measure of creativity that will provide insight where none of those tools will succeed on their own. Exploring massive datasets for actionable intelligence is essentially a creative process. And as a creative process, it is shaped by its context.
Which leads us to some simple questions, some food for thought:
- Could it be that your compute environment is hindering your ability to gain insight from your data?
- Are your data scientists only asking the questions they know (or even just think) they can get answers to, because of the limited performance of their IT systems?
- Would your organization benefit from the capacity to ask tougher, more probing questions, and from being able to ask them again and again in the course of a day, knowing that your compute environment is up to the task of supporting that kind of iterative exploration?
The answer to those questions might very well be “yes.” If you think so, Cray has just the kind of performant technology you need to provide the right context for discovery.