Then they're reading on some blog about protobuf, or hdf5, or arrow or whatever, and want to use it - but using either from Python needs installed C libraries. On linux/macos you'll need to dig into your package managers and then hope things are compatible, on Windows it's a complete pain. In conda, they can just 'conda install h5py' or whatever, and get proper hdf5 installed without having to figure out the nitty gritty details.
Well, already I can tell you're disconnected from what "most noobs" are actually like. Presumably you think "most noobs" are data scientists, which just isn't the case in my experience.
It is the case in my personal experience though. I don't think I'm disconnected from what 'noobs' are as I've taught a lot of them through my line of work.
Assume you're a noob and are exploring Python package universe, fire a new private tab and google 'most popular python packages'; in my case the first 3 links are:
- https://www.activestate.com/blog/top-10-must-have-python-pac...
- https://www.ubuntupit.com/best-python-libraries-and- packages-for-beginners/
- https://www.edureka.co/blog/python-libraries/
All of which list numpy, pandas and the rest e.g. tensorflow and friends.
That's the generalisation part I mentioned. The distinction you're making is non-trival for a wee noobie. Better they get the ground work in and expand to where they need to go.