It’s the Data, Dolts

Image may contain Drawing Art Doodle and Face
Illustration by Luci Gutiérrez

The whistle-blower Christopher Wylie held a roomful of British lawmakers rapt for three and a half hours on Tuesday, like a pink-haired, nose-ringed oracle sent from the future to explain data. . . .

At times he seemed slightly pitying of British officials who are investigating data mining, saying they did not have enough resources and lacked “a robust technical background.”

“I have had to explain and re-explain and re-explain and re-explain, you know, how relational databases work, what is an eigenvector, what is dimensionality reduction,” he said.

—“Cambridge Analytica Whistle-Blower Contends Data-Mining Swung Brexit Vote,” in the Times.

Explanation No. 1:

I will try to go slowly, so I don’t lose anybody. I’ve talked about relational databases already, but if you need a refresher please go to http://relationaldb.@^+66-7777.fda.lihgiw. Next, I will explain eigenvectors. These are just vectors, and we all know what vectors are—they’re things that go someplace, right? So you take regular vectors and make them eigen, and you get eigenvectors. Now let’s tackle dimensionality reduction. That simply means that you take a certain dimensionality and then you reduce it. So—are we good? Good!

Explanation No. 2:

Apparently, some of you with less robust tech backgrounds did not completely follow me the first time, so let’s review. Relational databases—I mean, come on, how hard can that be? They’re databases that relate to one another. Use your heads. Now, eigenvectors—we all use eigenvectors in everyday life, they’re practically self-explanatory. You take an ordinary vector and simply “eigen” it—please. Dimensionality reduction is so easy I’m not even going to explain it this time. Look at the diagram and work it out for yourselves.

Explanation No. 3:

I can’t believe you are eminent British lawmakers and yet can’t seem to grasp such basic stuff. Bring me my whiteboard. O.K., from the top, once again—relational databases are so ridiculously easy that even a two-year-old can understand them, and, in fact, many actual two-year-olds do understand them and are working for our company right now, making great big lovely heaps of money mining data and swinging elections, all while still in their nappies! Simply put, a relational database is a database that’s related either to another database or to Camilla Parker Bowles. (No, I’m kidding, guys.)

And why in the world can’t you understand eigenvectors? Look back over what I said in my first explanation. What do you mean, you weren’t taking notes? Do you think I’m telling you all this just to hear myself talk?

I want all of you to put your hands out on your desks. Yes, right now. You, too, Mr. Chief Solicitor! Now I’m going to rap you all on the knuckles with my phone. Yes, of course it hurts! It’s supposed to hurt. If it doesn’t hurt, you won’t remember. I advise you all to take notes the next time.

I’m doing some deep breathing before I proceed to dimensionality reduction, iteration three. O.K., let me give you an example that I’m sure will be obvious even to brilliant Solons of the bar like yourselves. Have you ever been to a 3-D movie? I’m sure your mommies and daddies took you when you were little. Remember you wore those funny glasses? Well, the “D” stands for “dimension.” Now, if we want to reduce the dimensionality, we simply take the glasses off. And what dimensionality do we see the movie in then? Two-dimensionality! Very good. That is what we mean by dimensionality reduction, and exactly the same principle applies to these very large computers you seem to have absolutely no clue about.

Explanation No. 4:

I’ve had it up to here with you people! I believe you actually know less than you did when we began. Somehow, everything I’ve tried to teach you has only made you stupider. I am sad, and ashamed, and very, very disappointed in you. Do I need visual aids? Must I dress up as an eigenvector? Must I bake relational-database cupcakes? What is wrong with you? Deep down, do you not want to know what a relational database is? Is the problem cultural? Is it my nose ring? Would you understand better if I took my nose ring out? What if I put some extra nose rings in? I am at my wit’s end. Wait—what if you put in nose rings?

The next day.

Wonderful! You all look splendid, and I’ve gone over the quizzes I gave you, post-nose-rings, and you got every question right. So that was the problem all along—the nose rings! Now, if you want to become really clever, we’ll do the hair. ♦