January 5, 2011
I've been staring at lines of data lately and looking for trends. No, not economic trends like the above, but abnormal values or sustained low or high values of hardware that might explain equipment failures. Often the values from a test track each other, much like the large dips and jumps in the DOW Jones Industrial Average can be seen mirrored in gas prices. Both are no doubt related to some third value of consumer demand or consumer confidence or perhaps even industrial production or health of the economy depending on how far you want to extrapolate. Both are charging up right now, leading many to wonder if there isn't a bubble in the stock market. But gas prices we tend to expect will continue to rise as the economy recovers. We might even see rising gas prices lead to rising inflation which in general would not be good for the stock market.
But it's difficult to make any predictions on data you can't see. I've talked before about the problems of data extrapolation which I think are especially true when something pertains to the economy. And yet here I sit, strolling through a massive amount of data looking for trends or trying to predict when failures can occur. Even when it might be a binary failure that no data can predict, I must look for trends nonetheless.