Well, that is theoretically a way to learn. It’ll give you experience for the future when you get vendor “sample code” that should be “turnkey”. 
All kidding aside, thanks for the endorsement. I remember back in the day that O’Reilly was the “go-to” publisher for this sort of thing for that very reason - high standards. But even back then, I wouldn’t wish most of the O’Reilly books on beginners.
I have been doing data analysis for 25 years and I am currently only using Python, but the choice of langue should be based on what those that you are working with are using. If you know one language and the art of data analysis, you can learn the other language quickly. If you are learning data analysis in parallel with data analysis i would suggest python as it seems that most of the newer methods in ML seems to be done in python and a fresh student of data analysis is probably going to want to learn ML as well.
As for the IDE i use spyder, but I am transitioning to sublime text which is a better editor, but not as good for playing around with data. Nothing is perfect, so in the end you have to know several tools anyway.
I agree 100%. The challenge is the person doing the learning in this case isn’t in a job yet, and is trying to prepare a more generally-applicable skillset without targeting a very specific job.
Funny that you say this, because she’s currently working through Python - a language I’ve spent zero time with previously - and I was helping her do debugging last night. It’s amazing how portable general programming principles are, once you’ve learned them.