How to change the positions of the objects

Question

As you can see on the image below, my entire plot (whether it's the heatmaps, the title or the colorbar) is “tighten” in the top of the figure/window and there is a large blank space in the bottom

Is there a way to lower it entirely AND to reduce the blank space in the bottom?

I've already tried to use the fig.tight_layout() and to change the values of the figsize argument of the plt.figurefunction but the space in the bottom stays the same (proportionally)...

Here is the piece of code:

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import AxesGrid



def graphs():
    fig = plt.figure(figsize=(13.5, 6.8))
    Grid = AxesGrid(fig, 111, nrows_ncols=(1,3), axes_pad=0.2, share_all=False, label_mode='L', cbar_location="bottom", cbar_mode="single")

    Matrices = [np.random.randint(-100,10,(10,10)) for i in range(3)]
    M, m = np.max(Matrices), np.min(Matrices)

    for values, ax in zip(Matrices, Grid):

        ax.tick_params(top = True, bottom = False, labeltop = True, labelbottom = False)
        ax.xaxis.set_label_position('top')

        heatmap = ax.imshow(values, vmin=m, vmax=M, cmap='plasma')
        ax.set_xticks(np.arange(0,10), labels=list(np.arange(1,11)), fontsize=6)
        ax.set_xlabel('Arrival', fontweight='bold', fontsize=8)
        plt.setp(ax.get_xticklabels(), rotation=35, ha="center", rotation_mode=None)

        for l in range(len(values)):
            for c in range(len(values[l])):
                text = ax.text(c, l, values[l,c], ha="center", va="center", color='k', fontsize=6)

        ax.set_title('Gaps', fontweight='bold', fontsize=9)

    Grid[0].set_yticks(np.arange(0,10), labels=list(np.arange(1,11)), fontsize=6)
    Grid[0].set_ylabel('Depart', fontweight='bold', fontsize=8)

    for cax in Grid.cbar_axes:
        cax.remove()

    cbar = Grid[0].figure.colorbar(heatmap, ax = Grid, fraction=0.2, aspect=50, location='bottom', label='Gaps')
    plt.suptitle('Visualisation of different gaps', fontsize=15, fontweight='bold', y=0.97)
    plt.show()

Check: https://stackoverflow.com/questions/74267842/how-to-save-figure-in-matplotlib-ajusted-to-the-object-size/74269297#74269297 — Joao_PS, Jun 02 '23 at 01:17
@Joao_PS I've tried the two options (```plt.figure``` and ```set_layout_engine```) but both of them don't solve my issue and I get an error : ```UserWarning: There are no gridspecs with layoutgrids. Possibly did not call parent GridSpec with the "figure" keyword plt.savefig('Test.png', dpi=800, transparent=True)``` — Alistair Lucet, Jun 02 '23 at 05:35
When I execute your code I get `AttributeError: 'ImageGrid' object has no attribute 'get_figure'`, — gboffi, Jun 02 '23 at 07:01
@gboffi I'm running my code on Python 3.11.3 with matplotlib 3.7.1 and I don't get any error message and I don't use ```get_figure``` attribute in my code — Alistair Lucet, Jun 02 '23 at 10:46

gboffi · Accepted Answer · 2023-06-02T20:05:38.020

You don't trust Matplotlib capabilities.

To obtain the above figure, I have

left alone the contents of grille.cbar_axes
used grille.cbar_axes to place the colorbar
left the suptitle's position to be decided by Matplotlib
used plt.tight_layout() to have all the spaces left to a minimum

def graphs():
    fig = plt.figure(figsize=(13.5, 6.8), dpi=60) # low dpi to fit the window on my screen
    grille = AxesGrid(fig, 111, nrows_ncols=(1,3),
                      axes_pad=0.2, share_all=False, label_mode='L',
                      cbar_location="bottom", cbar_mode="single")

    Matrices = [np.random.randint(-100,10,(10,10)) for i in range(3)]
    M, m = np.max(Matrices), np.min(Matrices)

    for values, ax in zip(Matrices, grille):

        ax.tick_params(top = True, bottom = False, labeltop = True, labelbottom = False)
        ax.xaxis.set_label_position('top')

        heatmap = ax.imshow(values, vmin=m, vmax=M, cmap='plasma')
        ax.set_xticks(np.arange(0,10), labels=list(np.arange(1,11)), fontsize=6)/
        ax.set_xlabel('Arrival', fontweight='bold', fontsize=8)
        plt.setp(ax.get_xticklabels(), rotation=35, ha="center", rotation_mode=None)

        for l in range(len(values)):
            for c in range(len(values[l])):
                text = ax.text(c, l, values[l,c], ha="center", va="center", color='k', fontsize=6)

        ax.set_title('Gaps', fontweight='bold', fontsize=9)

    grille[0].set_yticks(np.arange(0,10), labels=list(np.arange(1,11)), fontsize=6)
    grille[0].set_ylabel('Depart', fontweight='bold', fontsize=8)

    cbar = plt.colorbar(heatmap, cax=grille.cbar_axes[0], orientation='horizontal', label='Gaps')
    plt.suptitle('Visualisation of different gaps', fontsize=15, fontweight='bold')
    plt.tight_layout()
    plt.show()
graphs()

Addendum

You can change the "thickness" and the spacing of the colorbar using appropriate keyword arguments to AxesGrid; e.g., in the following I've used cbar_size="2%", cbar_pad=0.05
Re white space, the whole of titles, plots and colorbar when packed has a well defined aspect ratio, additional white space is added vertically or horizontally according to the aspect ratio of the figure — here it is what happens when figsize=(13.5, 5)

It helps with the space in the bottom even if it returns a warning: ```UserWarning: This figure includes Axes that are not compatible with tight_layout, so results might be incorrect. plt.tight_layout()``` but I can't change the width of the color bar with the ```aspect``` parameters — Alistair Lucet, Jun 02 '23 at 12:49
My code may not be the most effective as I “build” it step by step, adding details one by one — Alistair Lucet, Jun 02 '23 at 12:54
It;s just a warning, "results may be incorrect" — but they are correct. — gboffi, Jun 02 '23 at 14:12
Where should '''cbar_size''' and '''cbar_pad''' be placed as they are not argument of '''plt.colorbar''' — Alistair Lucet, Jun 02 '23 at 19:45
@AlistairLucet W.r.t. usage of `cbar_*=...`, please see my recent edit. — gboffi, Jun 02 '23 at 20:07

How to change the positions of the objects

1 Answers1

Addendum