May 16, 2012, 1:17 PM —
Image credit: Reuters
Those of us in the analytics field often get excited about the real-time performance of our technology. It's a gratifying challenge to successfully build systems that can react intelligently and instantaneously to massive data feeds pouring in at sub-second speed. It's a significant business need that commands our attention.
On the opposite end of the spectrum are business-driven analytics that must stand the test of time over long periods -- analytics that are judged not by how well they do in millisecond performance, but the accuracy of which are measured in years, even decades. By that I mean projects that use analytic tools to predict what's likely to happen in, say, fifty years -- not fifty nanoseconds.
Consider the new Boeing 787 Dreamliner or the Airbus 380. Both planes have projected 50-year lifespans. Analytics underpins those projections by helping engineers model everything from degradation patterns in materials and electronic components to the countervailing refurbishing capabilities in maintenance programs. The amount of data to factor in is enormous, as is its importance, because the longevity of the plane affects everything from the safety of passengers to the price buyers might be willing to pay.
Long-haul analytics is not just for aerospace companies. Energy companies, for example, confront projects with design criteria measured in years. Wind farms have projected lifespans of 25 years while coal-fired power plants can last even longer. China today is building coal-burning electric stations with 75-year horizons. And in the United States today there are more than 300 coal-fired plants older than 60 years and 10 that are up to 90 years old and still in operation. Although energy companies were not using computers and analytic tools to help them build plants back in the 1920s, they do so today to get the most efficient, clean, and durable electric power plants possible.
So, while most of us will naturally be focused on the benefits real-time analytics brings to business, we should not forget the value of long-haul analytics. Without it, our future would be much less predictable.