[PYTHON] Data visualization method using matplotlib (2)

Following on from Yesterday, I will explain the functions of matplotlib.

Colors, markers, linetypes

import numpy as np
from pandas import *
from pylab import *
import matplotlib.pyplot as plt
from matplotlib import font_manager
from numpy.random import randn

prop = matplotlib.font_manager.FontProperties(fname="/usr/share/fonts/truetype/fonts-japanese-gothic.ttf")

r = randn(30).cumsum()

#Specify color, linetype, marker
#Black, dashed line, marker is o
plt.plot(r, color='k', linestyle='dashed', marker='o')

plt.show()
plt.savefig("image.png ")

Lines are connected by a straight line by default. You can change this with the drawstyle option.

#Explicit RGB value
plt.plot(r, color='#ff0000', linestyle='dashed', marker='o', label='dashed')
#Change drawstyle
plt.plot(r, color='#0000ff', drawstyle='steps-post', label='steps-post')
#Add a legend
plt.legend(loc='best')

plt.show()
plt.savefig("image2.png ")

Scale, label

plt.xlim () and plt.xticks () return the current value when called with no arguments. Parameters can be set by specifying a value as an argument to this.

#Check the current value with no arguments
print( plt.xlim() )
# => (0.0, 30.0)
print( plt.xticks() )
# => (array([  0.,   5.,  10.,  15.,  20.,  25.,  30.]), <a list of 7 Text xticklabel objects>)

#Set a new value
plt.xlim([0, 40])
plt.xticks([0,4,8,12,16,20,24,28,32,36,40])

print( plt.ylim() )
# => (-7.0, 3.0)
print( plt.yticks() )
# => (array([-8., -6., -4., -2.,  0.,  2.,  4.]), <a list of 7 Text yticklabel objects>)

plt.ylim([-10, 10])
plt.yticks([-10,-8,-6,-4,-2,0,2,4,6,8,10])

plt.show()
plt.savefig("image3.png ")

Axis customization

Consider the following random walk plot.

fig = plt.figure()
ax = fig.add_subplot(1,1,1)

r = randn(1000).cumsum()
ax.plot(r)

plt.show()
plt.savefig("image4.png ")

Let's customize the scale and label of this. For example, the X-axis is ranked by 250, the characters are tilted 30 degrees, and they are displayed in Japanese.

ax.set_xticks([0, 250, 500, 750, 1000])
ax.set_xticklabels(['A', 'B', 'C', 'D', 'E'], rotation=30, fontsize='small')
ax.set_title('Test matplotlib plot', fontproperties=prop)
ax.set_xlabel('Rank', fontproperties=prop)

plt.show()
plt.savefig("image5.png ")

Add legend

An easy way to identify the plotted data is to label it and display it in a legend. It will be easier to understand if you separate the colors and line types for each data.

fig = plt.figure()
ax = fig.add_subplot(1,1,1)

def randn1000():
    return randn(1000).cumsum()

ax.plot(randn1000(), 'k', label='one')
ax.plot(randn1000(), 'b--', label='two')
ax.plot(randn1000(), 'r.', label='three')
ax.plot(randn1000(), 'g+', label='four')
ax.plot(randn1000(), 'b*', label='five')

plt.ylim([-100, 100])

ax.legend(loc='best')

plt.show()
plt.savefig("image6.png ")

Options when saving a file

You can specify image file options with plt.savefig ().

This time around, pandas hasn't appeared yet, so so far we're talking purely about matplotlib. From the next time, I will plot in combination with pandas.

reference

Matplotlib usage notes http://www.geocities.jp/showa_yojyo/note/python-matplotlib.html

Introduction to data analysis with Python-Data processing using NumPy and pandas http://www.oreilly.co.jp/books/9784873116556/