"Knowledge\nNow knowledge is where the real action is. As shown in Figure 1.1, knowledge consists of generalizable truths. Here’s what I mean. Information is about specific individuals or occurrences. When we say “Chandra is a female bank teller, and earns $48,000 a year,” or “Austin is a male bank teller, and earns $69,000 a year,” we have in our information repository some individual facts. They can be looked up and consulted when necessary, as you’ll learn in the first part of this book. But if we say “women make less money than men do, even at the same jobs,” we’re in a different realm entirely. We have now generalized from specific facts to more wide-reaching tendencies. In the language of our discipline, we’ve moved from information to knowledge. Properly gleaning knowledge from information is a trickier business than interpreting individual data points. There are established rules, some of them mathematical, for determining when an apparent pattern is actually reliable, what kinds of relationships can be detected with data, whether a relationship is causal, and so forth. We’ll build some important foundations with this kind of reasoning in this Crystal Ball volume and its follow-on companion. For now, I only want to make the point that knowledge – as opposed to mere information – opens up a whole new world of understanding. No longer is the world limited to a chaotic collection of individual observations: we can now begin to understand the general ways in which the world works...and perhaps even to change them.\n\nWisdom\nWisdom is the gold standard. It represents what we do with our knowledge. Let’s say we indeed determine that on average men are paid higher than women in our country, even for the same jobs. What do we do with that realization? Is it okay? Do we want to try and fix it, and if so, how? With laws? Education? Government subsidies? Revolution? You’ll remember my definition of Data Science on p. 2: deriving knowledge from data. This implies that the “wisdom” level of the hierarchy is really outside the discipline, and belongs to other disciplines instead. And that’s partially true: in some sense, the data scientist’s job stops when the deep truths about the real world are ferreted out and illustrated, leaving it to CEOs, directors, and other policy makers to act on them. But the data scientist is often involved here too, for a simple reason: a decision maker wants to know what’s likely to happen if a particular policy is implemented. Most non-trivial interventions will have results that are hard to predict in advance, as well as unintended side effects. One set of tools in the data scientist’s toolkit is for making principled, calculated predictions about such things, as well as quantifying the level of uncertainty in the predictions. Sometimes, the technique of simulation is used – carrying out experiments on virtual societies or systems to see the likely aggregate effects of different interventions. It’s like having a high-dimensional, multi-faceted crystal ball that lets you play out various scenarios to their logical conclusions. Starting with the rough and tumble real world and helping produce wise decisions about how humankind can deal with it all: that’s the grand promise of the data science enterprise. And those are the mighty waters you’re about to dip your toes in! I hope you’ll find it as exhilarating as I do."