Why Do This?

Do what? Writing. So What?

a quasi-anti-market motivation

Technically, I’m a photographer. In April 2016 I began taking a class called Foundations of Data Science online. Here I learned to collect, visualize, and analyze data. Deciding to take the class was the culmination of several months of independent research on the algorithms that actively shape many parts of our daily lives.

Technically, I take pictures like this for a living. (Kate Spade Valentine’s Day Gift Guide, campaign 2017)

Which parts of our daily lives? The news, content, and image feeds of Instagram & Facebook. But also the advertising that clicks me through to Amazon, and the content “curation” of Netflix or Spotify. I was pretty dissatisfied not knowing how these huge companies delivered content to me — something about realizing my dependence on these near-monopolies, and how they had the power to shape my daily mood and outlook.

Customers who bought this, also bought that: Association Rule Mining a darknet market listings database, from my first (unfinished) R project. Looks like spaghetti, but glad to get a feel for some of the workings behind Amazon’s recommendation engine. Dataset source: gwern.net

Fast forward — a year later and I have enough fluency in R to scrape data, visualize it, and perform some basic analyses. I still have to look up confidence intervals and check my regular expressions (and discover more ggplot stats and geoms). But I feel comfortable enough to perform robust regressions over ordinary least squares when the outliers get unruly, or switch over to random forests when the number of variables becomes overwhelming. One day I might even get to Bayesian methods.

OK, but “Why?”

So after a year of steadily training myself to be as objective as possible in statistical analysis — I realized that nobody cares. Correlation? Causation? Anomaly? ‘So what?’ became the harder question to answer, I learned¹.

Get mozz’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

Turns out most of my peers from art school don’t really like math; the idea of it alone can disgust some of them. I’ll blame educational methods in another rant. Maybe it’s natural to dislike what you don’t know. It seems more interesting to be curious about those kinds of things, though.

An early foray into explaining the obvious: charting darknet market sizes after the demise of Silk Road 2. Data originally from gwern.net

So, Why? “So What?”

If the question is “So what?”, then learning to write seems a way to the answer. And pulling InDesign into the vis workflow (todo: learn Illustrator). I bet I could find at least 10 articles citing the importance of writing and communication coming out of the world of data science. I bet these articles exist in other scientific fields and self-help manuals, too. I could read 10 more of these articles, or hit Publish on this one, so —

Hopefully that’s a partial answer to the question in the title.

:)

Notes

¹I should’ve known that — been trying to answer ‘So what?’ for years. It’s a great question that pervades all fields.