From Data to Information

Ten thousand rows of numbers is data: true, but useless on its own. Your eye cannot scan ten thousand anything and find a pattern. A program can, though. It can sort the rows, group them into categories, count how often something happens, or plot each value as a point on a graph. Once that happens, the data becomes information: something a person can actually read and act on. A table of every daily temperature for two years is data. A line graph of those temperatures, with a wave that rises every June and falls every January, is information. The shape only appears once the values are arranged in order and plotted, so the eye can catch the repeating curve.

The same trick works on anything sortable. Take 500 test scores and sort them from lowest to highest. Suddenly you can see where scores cluster, whether most students landed in the 70s and 80s, whether there is a long tail of very low scores, whether the distribution is roughly symmetric or skewed. None of that was visible in the original unsorted list, even though every number was already there. Sorting and grouping do not add new facts. They just reveal structure that was always hidden in the raw rows.

Joining two data sets can change the story entirely, because a conclusion drawn from one source only reflects what that source measured. A city looking only at reported bike thefts might conclude wealthy neighborhoods barely have a theft problem. Join in a second data set on reporting rates, and a hidden factor surfaces: those neighborhoods might report thefts more often, not experience fewer of them. This is why Big Idea 2 treats information as something built, not just found: the same underlying bits can support very different conclusions depending on what data gets combined and how.