making matplotlib graphs look like R by default?

Question

Is there a way to make matplotlib behave identically to R, or almost like R, in terms of plotting defaults? For example R treats its axes pretty differently from

Answer 1

# # # # # #

EDIT 10/14/2013: For information, ggplot has now been implemented for python (built on matplotlib).

See this blog or go directly to the github page of the project for more information and examples.

# # # # # #

To my knowledge, there is no built-in solution in matplotlib that will directly give to your figures a similar look than the ones made with R.

Some packages, like mpltools, adds support for stylesheets using Matplotlib’s rc-parameters, and can help you to obtain a ggplot look (see the ggplot style for an example).

However, since everything can be tweaked in matplotlib, it might be easier for you to directly develop your own functions to achieve exactly what you want. As an example, below is a snippet that will allow you to easily customize the axes of any matplotlib plot.

def customaxis(ax, c_left='k', c_bottom='k', c_right='none', c_top='none',
               lw=3, size=20, pad=8):

    for c_spine, spine in zip([c_left, c_bottom, c_right, c_top],
                              ['left', 'bottom', 'right', 'top']):
        if c_spine != 'none':
            ax.spines[spine].set_color(c_spine)
            ax.spines[spine].set_linewidth(lw)
        else:
            ax.spines[spine].set_color('none')
    if (c_bottom == 'none') & (c_top == 'none'): # no bottom and no top
        ax.xaxis.set_ticks_position('none')
    elif (c_bottom != 'none') & (c_top != 'none'): # bottom and top
        ax.tick_params(axis='x', direction='out', width=lw, length=7,
                      color=c_bottom, labelsize=size, pad=pad)
    elif (c_bottom != 'none') & (c_top == 'none'): # bottom but not top
        ax.xaxis.set_ticks_position('bottom')
        ax.tick_params(axis='x', direction='out', width=lw, length=7,
                       color=c_bottom, labelsize=size, pad=pad)
    elif (c_bottom == 'none') & (c_top != 'none'): # no bottom but top
        ax.xaxis.set_ticks_position('top')
        ax.tick_params(axis='x', direction='out', width=lw, length=7,
                       color=c_top, labelsize=size, pad=pad)
    if (c_left == 'none') & (c_right == 'none'): # no left and no right
        ax.yaxis.set_ticks_position('none')
    elif (c_left != 'none') & (c_right != 'none'): # left and right
        ax.tick_params(axis='y', direction='out', width=lw, length=7,
                       color=c_left, labelsize=size, pad=pad)
    elif (c_left != 'none') & (c_right == 'none'): # left but not right
        ax.yaxis.set_ticks_position('left')
        ax.tick_params(axis='y', direction='out', width=lw, length=7,
                       color=c_left, labelsize=size, pad=pad)
    elif (c_left == 'none') & (c_right != 'none'): # no left but right
        ax.yaxis.set_ticks_position('right')
        ax.tick_params(axis='y', direction='out', width=lw, length=7,
                       color=c_right, labelsize=size, pad=pad)

EDIT: for non touching spines, see the function below which induces a 10 pts displacement of the spines (taken from this example on the matplotlib website).

def adjust_spines(ax,spines):
    for loc, spine in ax.spines.items():
        if loc in spines:
            spine.set_position(('outward',10)) # outward by 10 points
            spine.set_smart_bounds(True)
        else:
            spine.set_color('none') # don't draw spine

For example, the code and the two plots below show you the default output from matplotib (on the left), and the output when the functions are called (on the right):

import numpy as np
import matplotlib.pyplot as plt

fig,(ax1,ax2) = plt.subplots(figsize=(8,5), ncols=2)
ax1.plot(np.random.rand(20), np.random.rand(20), 'ok')
ax2.plot(np.random.rand(20), np.random.rand(20), 'ok')

customaxis(ax2) # remove top and right spines, ticks out
adjust_spines(ax2, ['left', 'bottom']) # non touching spines

plt.show()

Of course, it will take time for you to figure out which parameters have to be tweaked in matplotlib to make your plots look exactly like the R ones, but I am not sure there are other options right now.