Sorry to tell the truth (own opinion), Jupyter is not on the table for good data analysis practice and software development.

Apologize that I am not a Jupyter person at all. Every time I saw ‘%matplotlib inline’ in the middle of import in a piece of code, I cannot persuade myself to continue reading.

It’s like meeting an ‘medium’ data scientist without math & stats evidence and good coding practice.

I would strongly suggest using IDEs for reproducibility, debugging (e.g. step-over) and version control purposes. Some of the pros are here. (in this case medium is a good practice)