On Data Analysis
Forget OLAP cubes and pivot tables! As is the case most of the time, 90% of the insight can be gleaned from very simple data plots. To be sure, the remaining 10% does shed valuable new light, but is also orders of magnitude relatively more tedious to distill.
One of my favorite tools for performing quick sanity checks on data and even for inferring high-level trends is a histogram; it simply partitions your data points into a fixed number of buckets, with each bucket holding points that fall within a given range. The resultant bucket sizes are then available to eye-ball, often plotted as bars whose lengths are proportional to the number of elements in the corresponding buckets.