Move graph position within plot (matplotlib)

Question

I am trying to create a horizontal bar graph with a table in matplotlib. I am almost there, but fail to align the table with the graph. What I got so far:

import matplotlib
matplotlib.use('Agg')
import numpy as np
import matplotlib.pyplot as plt


# Example data
appsol = ['llolLl', 'nnM', 'lllld bbbblnl', 'x2x', 'foobar', 'EXZ', 'Flups', 'Flaps', 'Foobar Barfooment', 'ABC', 'FABAS', 'common', 'AQT', 'Faberjak', 'simsalsa', 'LESS', 'Wermut']
y_pos = np.arange(len(appsol)) - .3
y_pos_2 = np.arange(len(appsol)) - .1
y_pos_3 = np.arange(len(appsol)) + .1
y_pos_4 = np.arange(len(appsol)) + .3
num_tickets = [4, 4,3,2,6,7,8,1,4,4,3,2,6,7,8,1,9]
num_tickets_2 = [7,6,5,4,3,4,2,1,2,4,1,0,3,0,2,1,0]
num_tickets_3 = [1,2,1,1,1,2,2,3,1,1,2,1,3,1,1,2,3]
num_tickets_4 = [8,7,6,2,13,6,8,9,7,6,5,4,3,6,8,9,12]

bar_width = .2

fig = plt.figure(figsize=(20,20))
ax = fig.add_subplot(111)

# correct yticks
plt.yticks(y_pos_2, appsol)

plt.barh(y_pos, num_tickets, bar_width,  align='center', alpha=0.4, color='r')
plt.barh(y_pos_2, num_tickets_2, bar_width,  align='center', alpha=0.4, color='b')
plt.barh(y_pos_3, num_tickets_3, bar_width,  align='center', alpha=0.4, color='y')
plt.barh(y_pos_4, num_tickets_4, bar_width,  align='center', alpha=0.4, color='g')
plt.yticks(y_pos, appsol)
plt.xlabel('Numbers')
plt.title('Horizontal Bar Chart with table')

# Table

empty_labels = ['' for a in appsol ]
plt.yticks(y_pos_2, empty_labels)
plt.tick_params(\
    axis='y',          # changes apply to the x-axis
    which='both',      # both major and minor ticks are affected
    left='off',      # ticks along the bottom edge are off
    right='off',         # ticks along the top edge are off
    labelbottom='off')

# Adjust layout to make room for the table:
plt.subplots_adjust(left=0.4, bottom=0.2)

cell_text = []
i = len(num_tickets) - 1
for j in num_tickets:
    cell_text.append([num_tickets[i], num_tickets_2[i], num_tickets_3[i], num_tickets_4[i]])
    i -= 1

row_lables = appsol
column_labels = ['So Huge\nMice', 'Elephants', 'Reptiles', 'Germs']

the_table = ax.table(cellText=cell_text,
        rowLabels=row_lables,
        colLabels=column_labels,
        loc='left')
the_table.set_fontsize(15)
the_table.scale(.5,5.0)
plt.savefig('barh_graph.png')
plt.close('all')

This produces enter image description here almost what I want, except that the table rows are not aligned with the bars of the graph. So I need a way to either move the graph down half a table row, or the table up half a table row. How can I do this?

Are the items even aligned at all? 'Wermut' seems to be at the top of the barplot but at the bottom of the table. — Rutger Kassies
– Rutger Kassies, Commented May 9, 2014 at 12:12
You are right - that was messed up. Thanks for pointing this out. Hurray for code review ;) It's corrected now. — Isaac
– Isaac, Commented May 9, 2014 at 12:35
I am glad you sorted out your problem, however it is best to post the minimal amount of code required to replicate your problem (and most mpl problems can be reproduced in <20 lines), you probably would have found your own problem. — tacaswell
– tacaswell, Commented May 9, 2014 at 13:23
Basic question: Why do you change your default backend to 'Agg', using matplotlib.use('Agg')? Secondary question: Do you have any idea why plt.show() might not work with 'Agg' backend (as in my case)? — Apostolos
– Apostolos, Commented Jun 30, 2018 at 21:53
I think I need the Agg backend to be able to save plots as png files. I need the pngs to include them in a pdf report. I think the plt.show() is not "working" with the Agg backend because its designed to produces files, instead of showing them on the screen. hth — Isaac
– Isaac, Commented Jul 2, 2018 at 6:07

Rutger Kassies · Accepted Answer · 2014-05-09 14:30:03Z

5

I think its easier if you create two subplots and add the table to the left subplot, instead of deriving a new subplot from the single subplot you have now. If you add the table to an existing subplot, you can use the bbox to stretch it from 0 to 1 (so fully) in the y-direction. Since the table has a header, setting the ylim of the right plot to (0, n_items), will make both align properly, and adding a slight offset because the bars are also given an offset of 0.4 (offset of the outer bar + half a barwidth). This should work automatically if the number of elements changes.

bar_width = .2

fig, axs = plt.subplots(1,2, figsize=(12,6))
fig.subplots_adjust(wspace=0, top=1, right=1, left=0, bottom=0)

axs[1].barh(y_pos[::-1], num_tickets[::-1], bar_width,  align='center', alpha=0.4, color='r')
axs[1].barh(y_pos_2[::-1], num_tickets_2[::-1], bar_width,  align='center', alpha=0.4, color='b')
axs[1].barh(y_pos_3[::-1], num_tickets_3[::-1], bar_width,  align='center', alpha=0.4, color='y')
axs[1].barh(y_pos_4[::-1], num_tickets_4[::-1], bar_width,  align='center', alpha=0.4, color='g')


axs[1].set_yticks([])
axs[1].set_xlabel('Numbers')
axs[1].set_title('Horizontal Bar Chart with table')
axs[1].set_ylim(0 - .4, (len(appsol)) + .4)

cell_text = list(zip(num_tickets, num_tickets_2, num_tickets_3, num_tickets_4))

row_lables = appsol
column_labels = ['So Huge\nMice', 'Elephants', 'Reptiles', 'Germs']

axs[0].axis('off')

the_table = axs[0].table(cellText=cell_text,
                     rowLabels=row_lables,
                     colLabels=column_labels,
                     bbox=[0.4, 0.0, 0.6, 1.0])

the_table.set_fontsize(15)

plt.savefig('barh_graph.png')
plt.close('all')

enter image description here

If you look closely you might notice that the bars are ranging from the top of the table to the bottom, you could tweak them a bit to make them start at the center of the upper and lower cell of the table.

edited May 9, 2014 at 14:30

answered May 9, 2014 at 14:17

Rutger Kassies

65k17 gold badges119 silver badges102 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Isaac Over a year ago

Cool, that is what I really wanted. It's not only more flexible, but it also looks nicer.

Isaac Over a year ago

I actually had to use plt.savefig('barh_graph.png', bbox_inches='tight') to get it right - copy pasting your version gave me a plot that filled the whole figure, so no boundary was shown, no xlegend or title.

Rutger Kassies Over a year ago

You can narrow the area which the subplots cover by setting there edges more inward to the figure, like: top=0.9, right=0.9, left=0.1, bottom=0.1, or even more, depending on what else you want to show in your figure.

Isaac · Accepted Answer · 2014-05-09 13:18:13Z

I found a workaround: I adjust the y-boundaries of the graph:

ax.set_ylim([-.5,17.5])

Of course these numbers only work with these data dimensions. So still, if someone has a better soltution I would like to see it. For completness, here is the enhanced version (also changed font size and added a legend):

import matplotlib
matplotlib.use('Agg')
import numpy as np
import matplotlib.pyplot as plt

# set font and size
font = {'family' : 'Bitstream Vera Sans',
        'size'   : 18}
matplotlib.rc('font', **font)

# Example data
appsol = ['llolLl', 'nnM', 'lllld bbbblnl', 'x2x', 'foobar', 'EXZ', 'Flups', 'Flaps', 'Foobar Barfooment', 'ABC', 'FABAS', 'common', 'AQT', 'Faberjak', 'simsalsa', 'LESS', 'Wermut']
y_pos = np.arange(len(appsol)) - .3
y_pos_2 = np.arange(len(appsol)) - .1
y_pos_3 = np.arange(len(appsol)) + .1
y_pos_4 = np.arange(len(appsol)) + .3
num_tickets = [4, 4,3,2,6,7,8,1,4,4,3,2,6,7,8,1,9]
num_tickets_2 = [7,6,5,4,3,4,2,1,2,4,1,0,3,0,2,1,0]
num_tickets_3 = [1,2,1,1,1,2,2,3,1,1,2,1,3,1,1,2,3]
num_tickets_4 = [8,7,6,2,13,6,8,9,7,6,5,4,3,6,8,9,12]

fig = plt.figure(figsize=(20,20))
ax = fig.add_subplot(111)

# correct yticks
plt.yticks(y_pos_2, appsol)
# this aligns table and graph!!!
ax.set_ylim([-.5,17.5])

labels = ['So Huge Mice', 'Elephants', 'Reptiles', 'Germs']
bar_width = .2

plt.barh(y_pos, num_tickets, bar_width,  align='center', alpha=0.4, color='r', label=labels[0])
plt.barh(y_pos_2, num_tickets_2, bar_width,  align='center', alpha=0.4, color='b', label=labels[1])
plt.barh(y_pos_3, num_tickets_3, bar_width,  align='center', alpha=0.4, color='y', label=labels[2])
plt.barh(y_pos_4, num_tickets_4, bar_width,  align='center', alpha=0.4, color='g', label=labels[3])
plt.yticks(y_pos, appsol)
plt.xlabel('Numbers')
plt.title('Horizontal Bar Chart with table')

# Legend
plt.legend(loc='lower center', shadow=True)
num_plots = 4
x_legend = 0.3
y_legend = -0.1
ax.legend(loc='lower center',
    bbox_to_anchor=(x_legend, y_legend),
    ncol=num_plots, # we want the legend on just one line
    shadow=True)
# Table

empty_labels = ['' for a in appsol ]
plt.yticks(y_pos_2, empty_labels)
plt.tick_params(\
    axis='y',          # changes apply to the x-axis
    which='both',      # both major and minor ticks are affected
    left='off',      # ticks along the bottom edge are off
    right='off',         # ticks along the top edge are off
    labelbottom='off')

# Adjust layout to make room for the table:
plt.subplots_adjust(left=0.4, bottom=0.1)

cell_text = []
i = len(num_tickets) - 1
for j in num_tickets:
    cell_text.append([num_tickets[i], num_tickets_2[i], num_tickets_3[i], num_tickets_4[i]])
    i -= 1

row_lables = appsol
column_labels = ['So Huge\nMice', 'Elephants', 'Reptiles', 'Germs']

the_table = ax.table(cellText=cell_text,
        rowLabels=row_lables,
        colLabels=column_labels,
        loc='left')
the_table.set_fontsize(18)
the_table.scale(.5,5.34)

plt.savefig('barh_graph.png')
plt.close('all')

And this is what it looks like: enter image description here

Collectives™ on Stack Overflow

Move graph position within plot (matplotlib)

2 Answers 2

3 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related