Raw weblog data gets organized into summarized, usable database tables via a couple of data processing steps. Sometimes intermediate tables are used to help calculate sums, groups, and values at the level we are interested in. For example, variables recorded at very granular time intervals might later be summed or averaged in longer groups of time. Storing what happened in the past hour or day is intuitive, whereas its hard to think of when you would need to know the exact millisecond a user clicked on something. You end up with tables that have a structure closer to what you might remember from the datasets you saw in the Intro to Stats class you aced, and these can be regularly used by developers or analysts without much hassle/manipulation.
Sideboob Exposed: Raw, Summary, and Usable Web Data aka: Why we have a whole page dedicated to this phenomenon.