Why every analyst should learn to ship

A notebook that only runs on your laptop is not a deliverable. It's a diary entry.

We celebrate the moment a chart finally looks right, then close the laptop. But the chart was the easy part. The hard, valuable part is everything that makes it survive contact with another human: a fixed random seed, a pinned dependency, a README that says which cell to run first.

Analysis is software, whether you admit it or not

Every analysis is a small program. It has inputs, transformations and outputs. The only question is whether you treat it like one. Version the notebook. Name the columns like you'll read them again in six months. Move the magic numbers into a config cell at the top instead of hiding them halfway down.

If nobody else can run it, it never really happened.

Ship it like an engineer would

Before you present a result, hand the notebook to a teammate and watch them run it from a clean checkout. Every place they get stuck is a bug you were about to ship. Fix those, and you've turned a private hobby into insight other people can build on.

That's the whole difference between an analyst who is interesting and an analyst who is trusted.