Pandas doc: add plots to docstrings and reference to in what's new

Enhanced time series plotting; adaptation of scikits.timeseries matplotlib-based plotting code

Let's add sample plots to this.

Continuing discussion on doc, I suggest to take a similar approach like statsmodels:

http://statsmodels.sourceforge.net/devel/generated/statsmodels.graphics.functional.rainbowplot.html#statsmodels.graphics.functional.rainbowplot

The plot in the docs is generated from the docstring the Sphinx source: http://statsmodels.sourceforge.net/devel/_sources/generated/statsmodels.graphics.functional.rainbowplot.txt the docstring: https://github.com/statsmodels/statsmodels/blob/master/statsmodels/graphics/functional.py#L208 So

.. ipython:: python

   ts = Series(randn(1000), index=date_range('1/1/2000', periods=1000))
   ts = ts.cumsum()

   @savefig series_plot_basic.png width=4.5in
   ts.plot()

could be moved to a docstrin in cases where it makes sense.

Can we not reuse the test code that may already be in the code base? https://github.com/pydata/pandas/blob/master/pandas/tests/test_graphics.py#L39

The major problem is that the test writing set-up in pandas is highly modular while docstrings à la statsmodels are more verbose.

But, taking a second look at the docstring above, the tests could actually be used to produce doc plots:

instead

.. plot:: plots/graphics_functional_rainbowplot.py

use

.. plot:: plots/file_with_plot_examples.py

In file file_with_plot_examples.py you may for the mean time import the test functions.

On the other hand, such plot files: https://github.com/statsmodels/statsmodels/blob/master/docs/source/plots/graphics_functional_rainbowplot.py coudl be easily contributed by pandas users for more cases that core devs are probably thinging of currently.

Comment From: timmie

I hope my brainstroming above did not confuse too much. It's my first familiarisation with pandas docs and trying to push ideas around...

Comment From: changhiskhan

The statsmodels docs you linked to are API docs, which are very useful and something we definitely would like to have. However, it is different from the pandas documentation which is more of an in-depth guide/tutorial. So it's not a question of which approach to choose at all.

On Fri, Jun 1, 2012 at 3:41 AM, timmie < reply@reply.github.com

wrote:

Enhanced time series plotting; adaptation of scikits.timeseries matplotlib-based plotting code

Let's add sample plots to this.

Continuing discussion on doc, I suggest to take a similar approach like statsmodels:

http://statsmodels.sourceforge.net/devel/generated/statsmodels.graphics.functional.rainbowplot.html#statsmodels.graphics.functional.rainbowplot

The plot in the docs is generated from the docstring the Sphinx source:

http://statsmodels.sourceforge.net/devel/_sources/generated/statsmodels.graphics.functional.rainbowplot.txt the docstring:

https://github.com/statsmodels/statsmodels/blob/master/statsmodels/graphics/functional.py#L208

Can we not reuse the test code that may already be in the code base?

Reply to this email directly or view it on GitHub: https://github.com/pydata/pandas/issues/1369