VIS: let scatter plots obey mpl color scheme (#3338)

[jorisvandenbossche]

Closes #3338.

Fixes the remaining plot functions that don't follow the matplotlib style (functions based on scatter plots).

Rationale:

• set the color keyword c to default one defined in matplotlibs rcParams if c was not specified by the user (because c='b' is harcoded in matplotlib scatter method).
• the color keyword is not checked because, if specified by the user, it already is given preference by matplotlib over c if both are given.

There is a new PR on scatter based plotting methods (#3473), but I suppose that it can be handled in that PR to ensure this new plotting method also follows this approach?

[jreback]

@cpcloud ?

[cpcloud]

pulling it down now

[cpcloud]

Still get obnoxious blue:

[jorisvandenbossche]

@cpcloud How did you obtain this plot?

If I do (example from the docs):

pd.options.display.mpl_style = 'default'
df = pd.DataFrame(np.random.randn(1000, 4), index=pd.date_range('1/1/2000', periods=1000), columns=list('ABCD'))
df = df.cumsum()
df.plot()

I get the new color scheme, as you also see in the docs: http://pandas.pydata.org/pandas-docs/dev/visualization.html#basic-plotting-plot

[jorisvandenbossche]

@cpcloud Ah, it seems like you posted the plot in this part of the docs: http://pandas.pydata.org/pandas-docs/dev/visualization.html#suppressing-tick-resolution-adjustment

Well, this one is generated by explicitly stating the colors you want (df.A.plot(color='r')), so it logically follows this and not the default color scheme.

Off course, we could remove this color=.. like this:

with pd.plot_params.use('x_compat', True):
    df.A.plot()
    df.B.plot()
    df.C.plot()

to get rid of these colors in the docs, but then the purpose of the example is a little bit lost, because you could do the same with just df[['A', 'B', 'C']].plot(x_compat=True)

[jreback]

@jorisvandenbossche ready to go?

[jreback]

@cpcloud this ok?

[jorisvandenbossche]

@jreback for me this is ready

[cpcloud]

Yep good 2 go