[IPython-User] best practice for notebook use?

Lev Givon lev@columbia....
Tue Jan 28 11:07:57 CST 2014


Received from Fritz Payr on Tue, Jan 28, 2014 at 10:51:45AM EST:
> I am a fan of IPython notebooks for doing and sharing scientific
> work, but I see that in the course of my PhD-studies I have run into
> a number of practical problems using them, e.g.
> 
>  * combining version control and notebooks (I find it hard to save
>    notebooks so they don't get lost between versions or when migrating
>    my PC);
>  * how to write notebooks that are reproducible (i.e. executable) /and/
>    readable;
>  * how to organize
>      o code (3rd party libraries, own classes and modules,...),
>      o data (primary, or intermediate results), and
>      o presentation (in the notebook);
>  * how to publish a notebook and a journal-paper "as a useful combination";
>  * etc.

I find it convenient to store all of my project notebooks in a single git
repository on whatever machine I am running the notebook server so that I can
commit revisions after working on a notebook. Assuming that your notebook's
contents are not confidential, this enables you to immediately publish them to
GitHub and view them via nbviewer.ipython.org. Several examples my colleagues
and I recently developed using this workflow are available at

http://neurokernel.github.io/docs

> Is recommendable documentation of best practices available?
> (If not: Do best practices at least exist as tacit understanding
> amongst experts? Can you point out good examples?)
> Are good sources available outside the IPython-cosmos (seeing that
> the problem may be about presenting scientific results in general)?
> 
> Thanks for any hints, ideas, or references!
> Fritz

NB: ipython correspondence is being consolidated on the ipython-dev
list [1]. You may wish to direct future questions/comments there.

[1] http://mail.scipy.org/pipermail/ipython-user/2013-December/013486.html
-- 
Lev Givon
Bionet Group
http://www.columbia.edu/~lev/
http://lebedov.github.io/



More information about the IPython-User mailing list