Object oriented data science

I got a fun question from one of you:

"This stuff you're saying about OOP is interesting... but I'm a data scientist. How can it help me?"

This question is delicious.

As someone who formerly wore my "data science" hat full-time (I still wear it part-time), OOP is magic. Mastering it permanently "leveled up" all the data processing, analysis and inference I was doing.

The reason:

Data science has two faces. It's often exploratory, at the early stage. You have to hit a lot of different directions, all at once.

But often once your exploration hits its groove... you're in a rush to ship it. The quick-n-dirty prototype gets "promoted" to final product, typically before you've realize it.

With all these different experiments you're slinging at the wall, to see what sticks... well, there's a lot of complexity.

And it's tough. Very quickly, you find yourself slogging through the stinky sticky mud of disconnected scripts, scattered notebooks and disorganized code.

And what's the one thing that can make this easier?

That's right. Object-oriented programming.

LISTEN: If you're a data scientist and haven't studied OOP yet... I'm telling you. IT. WILL. CHANGE. YOUR. LIFE.

I'm not even kidding.

A small investment of OOP knowledge goes farther than you can imagine. Take it from someone who's been right where you are, right now.