Working Up

by Dwayne Phillips

Many of today’s data scientists are similar to man of the web designers of the 1990s. I think this is a good thing.

A recent conversation with a colleague helped me to realized something about data science in today’s world. I thought data scientists were computer scientists, engineers, and mathematicians who knew programming and a field of specialty in addition to their STEM background. (I think I read that description somewhere.)

“Oh no,” said my colleague. “Most of them I work with have a liberal arts degree, a feel for numbers and logic, and enough smarts and initiative to have learned how to string together ten lines of Python to call the right packages and do something.”

What these data scientists lacked was a feel for science, repeatable experiments, rational thought, and such. They were basically parroting things they saw online.

Hmmm, that sounds familiar. In the 1990s in the days of Web 1.0 we had “web designers.” A successful web designer I knew had a bachelor’s degree in English. He had an appreciation for art and what looked good on the screen. He could read and write. He had enough smarts and initiative to learn HTML and a little about cascading style sheets and the like.

Then the dot com boom crashed in the late 1990s and he went to grad school to work on a Masters of Fine Arts.

Will we have a data science boom crash ka-bang or something and all our current stuff crumble to the abyss? I don’t know. I hope not.

I liked the idea in the 1990s of liberal arts majors working in the tech field. They brought a lot with them to the rooms full of techies. They made us and the industry better.

I like the idea of liberal arts majors being data scientists. They bring a lot with them to the ZoomerTeams meetings of full of techies. They make us and the industry better.

They also demonstrate the idea of “democratization.” I hate the term, but like the idea. We have built tools that people can use. These liberal artists are smart. They can learn these tools and use them (often better than us techies who built the tools). Sure, they take missteps along the way and have experiments that aren’t repeatable and don’t know what configuration management is (come to think of it, most STEMmers don’t know what configuration management is either, but that is the topic for another day).

Still, these tools bring more people into the room, and we are all better because of that.