As a data educator, I’ve helped plenty of self-styled math- and technophobes find out that data and analytics is for them, and it’s fun. These learners are often pretty surprised by how approachable and humorous many analytics professionals are. I mean, not all of us are out of touch… at least those of us who want to sell books.
Advancing into Analytics is a technical book. And yes, technical books have a reputation to maintain as dense information slogs. At the same time, tech enthusiasts are known to be eccentric and irreverent. O’Reilly (my publisher’s) does a great job at delivering top-notch technical content in a fun and delightful manner. I’d like to think I’m known for the same.
To that end, I’d like to present some of my favorite quotes from Advancing into Analytics, along with a bit of context and relevant passages from the book.
“Data… you never know what’s gonna come through that attachment”
This quote is a paraphrase from the opening of the TV show Pawn Stars, where Rick Harrison says: “You never know what is gonna come through that door.” It opens Chapter 1, Foundations of Exploratory Data Analysis. As data analysts, we often peddle in data as varied as Corey and Chumlee do antiques:
… confronted with a new dataset, you never know what you are going to find. This chapter is about exploring and describing a dataset so that we know what questions to ask of it. The process is referred to as exploratory data analysis, or EDA.
Fortunately, exploratory data analysis (EDA) gives us a structured process to size up whatever dataset makes its way to your workstation. Through a series of descriptive and visual techniques covered in this chapter, you’ll be poised to test relationships in the data later in the book, using Excel, R and Python.
Well, I guess this isn’t so much a quote as a tautology, but it’s an important one. Have you ever considered why variables have that name? I explain it in Chapter 1 of the book:
We call them variables because their values may vary across observations. If every observation we recorded returned the same measurements, there wouldn’t be much to analyze.
In fact, researchers have traditionally struggled to analyze data where over 90% of observations take on the same value. Between larger datasets and increased computing power, this has become less of a constraint, but the ability to vary is still what data’s all about.
VLOOKUP() is the duct tape of Excel”
Chapter 5 of Advancing into Analytics is titled “The Data Analytics Stack.” While the book only covers Excel, R and Python in depth, I felt it was important to situate them in the wider analyst toolkit. In explaining how relational databases work, I like to start off with an example that every seasoned Excel analyst knows well: the mighty
VLOOKUP(). (Yes, I hear you,
INDEX()/MATCH() and now
XLOOKUP() fans. The analogy works for you, too.)
As I explain in the book, “I like to call
VLOOKUP() the duct tape of Excel because of its ability to connect
datasets together.” Now, duct tape is a helpful tool. With skilled hands, it can accomplish a lot. But, you may not want to trust duct tape for every job. In data, we have relational database joins to make more stable, efficient connections between data. As I go on to write, “If
VLOOKUP() is like duct tape, then relational database joins are welders.”
“There’s a package for that!”
This quote comes from Chapter 6, First Steps with R for Excel Users. In this chapter, readers install the R code base. I draw the following analogy to explain the difference between this code base and packages:
Imagine if you weren’t able to download applications on your smartphone. You could make phone calls, browse the internet, and jot notes to yourself—still pretty handy. But the real power of a smartphone comes from its applications, or apps. R ships much like a “factory-default” smartphone: it’s still quite useful, and you could accomplish nearly anything necessary with it if you were forced to. But it’s often more efficient to do the R equivalent of installing an app: installing a package.
Both R and Python feature staggering collections of packages for all sorts of tasks. To paraphrase the famous slogan: “There’s a package for that!” This analogy carries conceptual but also practical value: to use an app, you download it once, but open it each time you need it. The same goes in R and Python… so you’ll be using
import, respectively, a lot.
“It’s your world… the data’s only living in it”
This comes at the end of Chapter 3, Foundations of Inferential Statistics. There are a lot of rules and operations in statistics to remember. It’s tempting to go into autopilot when crunching data, plugging-and-chugging for p-values. This can open the door to incomplete or misleading analysis. As I explain in the book:
Statistics and analytics are powerful tools for making sense of the world, but they’re just that: tools. Without a skilled craftsperson in control, they can be useless at best and harmful at worst. Don’t be content to take the p-value on its face; consider the broader context of how statistics works and the objective you’re aiming to meet (without gaming the results, as you’ve seen is possible). Remember: it’s your world, the data’s only living in it.
Data and analytics work best in perspective — as supporting evidence for what we decide. The data can’t speak for itself. It’s YOUR world… the data’s only living in it!
Don’t rely on anecquotes: Read the book
I’ve just given you a few fun quotes from Advancing into Analytics. But, it’s a pretty small sample: can you generalize about the approachability and zaniness of the whole book based on a few anecdotes… or should I say “anecquotes?” Find out for yourself: read the book.