I tamed the algorithm

Like a dragon lurking in the shadows, the dreaded algorithm hides. It's this mysterious shadowy figure, not unlike the shapeshifting boogeyman thwarting you at every turn. 

It at least seems like that, doesn't it? I promise you. The algorithm can be vanquished. It is not as mysterious as it may seem.


I am the dragon tamer

Okay, I’m being a little bit dramatic with the whole dragon tamer thing, but it doesn’t change the fact that I did, in fact, conquer the algorithm. I'm trying to make this as un-nerdy and readable as possible. If you have any questions about anything here, let me know.

Let’s start off by giving you a little bit of context.


My doctorate is in an overly pretentiously named and no longer existing program. I just say that I'm a psychometrician (a psychologist statistician) with some knowledge in clinical psychology. One thing I know about is psychometrics, which is the field of how you study and measure people.

Before I even wrote my proposal for my dissertation, I researched different kinds of data mining techniques, so I have a working understanding of how these types of techniques work


Then, when I wrote my proposal, I had to include math proofs on the algorithm I was using and the statistical analysis technique I was using. I also had to understand it enough to explain it to my committee members, some of whom knew nothing about data mining.

Point is that I became familiar with different kinds of ways of analyzing various kinds of data.


Now-a-days when I look at a bookseller's recommendations (for instance), I like playing a guessing game of, what kind of data did the bookseller collect and what kind of analysis did the bookseller's computer use? Based on what I've read, my guesses have been on the money. Yeah. I'm a nerd.


My dissertation was the first project in the department to use data science (like data mining methods) as the focus. I used two different statistical methods on the same data to see how it affected the results, and it did. One method couldn’t find the needle in the haystack. The other could. So yeah, I found out interesting stuff. I also showed that using data science was a good idea for dissertation projects. More importantly, I got my doctorate. Yay me.

So take away. I tamed the algorithm dragon. I had it for dinner. She's really nice once you get to know her. It got me a doctorate. Yay!

If you'd like to know more about how algorithms work in normal people English (or NPE), just drop me a line. I'd be happy to dispel the mysteries. As a former educator, I do like explaining things. It's an itch I don't get to scratch often enough.


An Algorithm

An algorithm is a step-by-step instruction to a computer on how to analyze data and what to do with the analysis afterwards. If you ignore the mathy looking parts and squint your eyes, they kind of look like recipes.

A bookseller, for instance, will have more than one separate algorithm running at the same time. "Other people bought books like this" will be a different algorithm than "other people also looked at this" because the data and the analysis of that data will be different.

Image by Artis Kančs