[Matplotlib-users] physical size of plot

vincent.adrien at gmail.com vincent.adrien at gmail.com
Wed Mar 8 10:15:44 EST 2017

Hello Jean-Philippe,

you can use the `figsize` keyword argument when creating a figure to set
its size in inches (1 in = 2.54 cm). You can also set the DPI value
directly with the `dpi` argument. Or you can wait until saving the
figure: the keyword argument is the same.

For example

import matplotlib.pyplot as plt

# Set the dimensions that you want in inches,
# here let's say a quarter of a A4 sheet.
a4_sheet = (21.0, 29.7)  # in cm
my_size = [0.5*val/2.54 for val in a4_sheet]  # converted in inches

fig, ax = plt.subplots(figsize=my_size)  # one could add 'dpi=300'

ax.plot([0, 1, 2], [1, 2, 0])  # dummy plot

# Saving with a specific dot-per-inch (dpi) value, here 300
# Given my_size, the final picture should be ~ 1240x1754
fig.savefig('my_figure_in_300_dpi.png', dpi=300)


You may find more details in the online documentation:

I hope this helps.



PS: the figure size that you report is a bit surprising. The 2.0
defaults are expected to produce 640×480 pixel-pictures (if I am not wrong).

On 08/03/2017 15:19, Jean-Philippe Grivet wrote:
> Hello,
> I use python(3.6)+numpy+matplotlib(2.0) (Anaconda distribution under
> Win7) to draw simple figures and I meet the following problem. All the
> figures that I plot come out the same size (432x288 px). This is OK for
> simple curves but leads to obscure plots in more complicated cases.
> I notice that all figures in the pyplot tutorial have size 550x450 as
> they do also in the Matplotlib examples, while many figures in the
> gallery are 165x135.
> How canI force the size of a picture in order to display it full screen
> or to cover most of anA4 sheet of paper ?
> Thank you for your help
> JP Grivet
> ---
> L'absence de virus dans ce courrier électronique a été vérifiée par le
> logiciel antivirus Avast.
> https://www.avast.com/antivirus
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users at python.org
> https://mail.python.org/mailman/listinfo/matplotlib-users

More information about the Matplotlib-users mailing list