On another note, I admire how you avoided the (too) easy plays on words here around eating / taking a bite of the apple. "Google is eating Apple's lunch as we speak." "Microsoft is about to take a bite of their pie as well."
251 karma · joined April 1, 2011
On another note, I admire how you avoided the (too) easy plays on words here around eating / taking a bite of the apple. "Google is eating Apple's lunch as we speak." "Microsoft is about to take a bite of their pie as well."
It probably does not really need to be repeated, but I guess it cannot hurt. Data being big rarely causes it to be much more interesting. There can be much more actionable information in a few hundred records than in billions, depending on context, proper experimental design, and so on.
Also, an often overlooked point is that precision of statistical estimators (e.g. width of confidence intervals) typically scales with the square root of sample size; yet the cost of processing an additional observation is typically constant (e.g. with distributed systems like hadoop). So in many cases there are decreasing returns to data size.