Testing the template on MacOS

[m-wierzba]

I have tested the template on MacOS and it runs pretty much flawlessly. :)

There is one line that needs minor adjustment, I couldn't avoid creating a backup file with sed and needed to run this call instead:
sed -i'.bak' 's/muted/Blues/g' code/mk_figuresnstats.py

But let me use this opportunity to introduce several other minor changes to the README:

• I added a broader overview of how the manuscript is actually built at the beginning of the README (this is a copy-paste of what you already have in the .tex file, but I think it nicely clarifies the general idea of this reproducible paper template).
• I reorganised the software requirements a bit: I would skip introducing the requirements.txt until it's needed. Otherwise, people may try to install seaborn, sklearn and DataLad API themselves only to realise that they only needed to create a virtual environment as explained later on. On the other hand it might be worthwhile to mention the virtualenv as a requirement (especially assuming that not all readers are fluent Python users).
• I tried to stress the difference between datalad get and datalad get -n more, becase this can easily be missed by a beginner.
• I would encourage people to actually READ the manuscript once it's created. :)

Also, I had the general feeling that a lot of details on the basics of DataLad is provided, but not so much on other elements (e.g. LaTeX, Makefiles). But this is understandable, it's a DataLad tutorial after all.