Why does Excel 2007's nonlinear solver crap out on k-means unlike the newer 3 versions of Excel?
But my favorite question:
Why in the world is your face on the cover of your book? and upside down to boot?
But this wasn't done arbitrarily. No, the decision behind it was intentionally made by myself, the cover designer, and the publisher.
The upside-down face is meant to set expectations for the reader. This book has more than a point-of-view -- it has a strong, yet somewhat kooky, narrative presence.
Strong narrative presence ==> my face.
Some fairly strange content (for example, a playlist of mid-90s club hits) ==> turn it upside down.
The cover is a reflection of the content.
Data Smart is an experiment in narratively heavy-handed, self-conscious technical writing. And so this question about the cover is actually part of a larger discussion on technical books:
Should you write the kind of technical book that would lead to such a cover?
Point of View versus Narrative Presence
Most, dare I say all, technical books have a point of view. They're written by humans, so the mere decision of what to put in and what to leave out of a text is pointed. Hell, even if the author leaves nothing out of a book then that act of non-editing is pointed.
Take for instance Shifrin's 1995 book: Abstract Algebra: A Geometric Approach. I love this book. It's tough, it's dry as a bone, but the fact that the author chooses to explain things like group theory using geometric language is none-the-less a point of view.
Similarly, in Data Smart, I teach optimization early on and then build on optimization to teach other techniques. A lot of other data science books ignore optimization entirely, simply letting optimization techniques embedded deep in packages silently do the heavy liftying. The fact that I think optimization is important is my point of view.
That said, having a strong narrative presence is about more than just having a point of view. It's about the reader never being able to forget that the author is talking to them. Instead of letting dry words speak plainly for themselves, I am, as the author, constantly jumping in with silly similes (how a Big M constraint is like a dead squirrel), odd data (empire waisted dress sales are a good pregnancy predictor unless we're in Sense and Sensibility), and dogmatic statements about minor aesthetics (Excel's 3d pie charts are the spawn of Satan).
You see this type of authorial presence in movies all the time. Take for instance, the concept of a "Spielberg visual gag," such as the T-rex in the rearview mirror or the librarian stamp scene in Last Crusade:
So why do it?
100% Clarity is Overrated
A lot of people feel that a technical book should minimize the author's voice and lay out information in a straight-forward, tight way. I get it.
But I thought about the way I learned math back in college. I read, I worked, and I smoked. I smoked a lot when I studied.
The smoke break was ritual for me when studying tough concepts. It was a Selah that I inserted between reading proofs in my topology textbook. Would it have been more efficient to not take smoke breaks? If I weren't human, sure. But as a person, I could only read about the hairy ball theorem for so long before I needed a break.
And that's what all the stupidity in my book is about. Sure, it inherently slows down the pace and inserts jokes where a clearer sentence would have sufficed. But it's a welcome pause for many readers. A chance to remind the reader that I'm still there with them working through the material. Take a breath, have a laugh, and let's keep moving.
And who is that reader?
The reader here is important. The audience for this technical book was non-technical. I specifically wanted to teach data science to those folks I felt like were largely ignored by the manifold R and Python based O'Reilly texts floating around. I wanted folks with more of an MBA background to find a foothold in data science.
And those people probably aren't used to dense technical material. At the same time, I didn't want to dumb the text down. There are plenty of "about books" on data science already (Moneyball books as my tech editor calls them). So the only way to talk to these folks without dumbing down subject matter is to make it very clear, use tools they're familiar with, pace the content well, and make it approachable.
Should more technical books have faces on the cover?
It depends on the audience, but generally, I think yes.
Recently, we've heard plenty from the media about how the best persuaders, the best educators, the best communicators are good story tellers. Stories activate more of the brain.
Should technical books capitalize on this? I don't see why not. Storytelling, even if it's in the form of approachable, real-world datasets, referential humor, or anecdotes from the writer's own practical use of the techniques, can go a long way to keeping the reader engaged and tracking with you.
I hope I'm proved right on this point. Because putting your face on a book is going a bit further than just plopping your name on front. If people don't like Data Smart, I will forever be the jackass with his face on the front of a 400 page turd. It's a risk, but one that I hope will change the way people think about teaching and learning complex technical concepts.