Set x axis length of all subplots to same width on screen

I've converted a simple MATLAB script [1] for wavelet decomposition-based
analysis to Python. I now want to create figures similar to [2] that
feature five subplots in one column, with the 1st and 3rd being generated
using 'subplot.plot' and the others being generated using 'subplot.imshow'.

I want to find a way of scaling the x and y axes of all subplots so that
they're the same size on screen across all subplots. Unfortunately I can't
find any combination of 'aspect' or 'extent' that will let me do this. I
should note that I've added colorbars to all my image plots; the problem
exists regardless of whether these are used.

Has anyone got any suggestions? Apologies if this is a trivial matter; I'm
fairly new to MP.

FYI I'm using matplotlib v1.1.0.

[1]
http://www.mathworks.com/matlabcentral/fileexchange/18409-comparing-time-series-using-semblance-analysis
[2] http://www.mathworks.com/matlabcentral/fx_files/18409/1/goldoil.jpg

Regards,

Will Furnass

···

--
View this message in context: http://old.nabble.com/Set-x-axis-length-of-all-subplots-to-same-width-on-screen-tp33753799p33753799.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

You could try

    fig, axes = plt.subplots(nrows=5, sharex=True, sharey=True)

And then call `axes[0].plot(...)` and `axes[1].imshow(...)`. You could also
unpack the axes array and access them that way (if you prefer):

    ax0, ax1, ax2, ax3, ax4 = axes
    ax0.plot(...)

Best,
-Tony

···

On Thu, Apr 26, 2012 at 12:51 PM, willfurnass <will@...4073...> wrote:

I've converted a simple MATLAB script [1] for wavelet decomposition-based
analysis to Python. I now want to create figures similar to [2] that
feature five subplots in one column, with the 1st and 3rd being generated
using 'subplot.plot' and the others being generated using 'subplot.imshow'.

I want to find a way of scaling the x and y axes of all subplots so that
they're the same size on screen across all subplots. Unfortunately I can't
find any combination of 'aspect' or 'extent' that will let me do this. I
should note that I've added colorbars to all my image plots; the problem
exists regardless of whether these are used.

Has anyone got any suggestions? Apologies if this is a trivial matter; I'm
fairly new to MP.

FYI I'm using matplotlib v1.1.0.

[1]

http://www.mathworks.com/matlabcentral/fileexchange/18409-comparing-time-series-using-semblance-analysis
[2] http://www.mathworks.com/matlabcentral/fx_files/18409/1/goldoil.jpg

Regards,

Will Furnass

You could try

    fig, axes = plt.subplots(nrows=5, sharex=True, sharey=True)

And then call `axes[0].plot(...)` and `axes[1].imshow(...)`. You could also
unpack the axes array and access them that way (if you prefer):

    ax0, ax1, ax2, ax3, ax4 = axes
    ax0.plot(...)

···

On Thu, Apr 26, 2012 at 12:51 PM, willfurnass <will@...4073...> wrote:

I've converted a simple MATLAB script [1] for wavelet decomposition-based
analysis to Python. I now want to create figures similar to [2] that
feature five subplots in one column, with the 1st and 3rd being generated
using 'subplot.plot' and the others being generated using
'subplot.imshow'.

I want to find a way of scaling the x and y axes of all subplots so that
they're the same size on screen across all subplots. Unfortunately I
can't
find any combination of 'aspect' or 'extent' that will let me do this. I
should note that I've added colorbars to all my image plots; the problem
exists regardless of whether these are used.

[1]

http://www.mathworks.com/matlabcentral/fileexchange/18409-comparing-time-series-using-semblance-analysis
[2] http://www.mathworks.com/matlabcentral/fx_files/18409/1/goldoil.jpg

----

Hi Tony and others,

Thanks for the suggestion Tony. It works if I remove my colour bars.
However if the colour bars are enabled (for the images on the 2nd, 4th
and 5th rows) then the image on the 5th and last row is squashed up on
the LHS of the subplot area to make room for all 3 (!) colour bars,
even though all subplots should share an x axis. Any further
suggestions?

Also, I'm still having problems getting tick labels on ax.plot() and
ax.imshow() subplots to line up:

I create one subplot using plot(t_vec, y_vec)
I then create an image using t_vec and y_vec and display it as a second
subplot using imshow(image); the width of the image matrix is len(t_vec)
(not t_vec.max())

I am looking for a sane way of displaying the two subplots above each other
so that they are both the same width in pixels and have x tick labels at
consistent intervals (in pixels) that correspond to multiples of some value
of t.

I have tried using
ax.xaxis.set_major_formatter(ticker.FuncFormatter(xformat)) to scale the
image subplot ticks but can't get this to work in conjunction with the
setting of the tick intervals on both subplot axes using set_xticks. Has
anyone got any suggestions?

Regards,

Will

--
View this message in context: http://old.nabble.com/Set-x-axis-length-of-all-subplots-to-same-width-on-screen-tp33753799p33760459.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

You could try

    fig, axes = plt.subplots(nrows=5, sharex=True, sharey=True)

And then call `axes[0].plot(...)` and `axes[1].imshow(...)`. You could also
unpack the axes array and access them that way (if you prefer):

    ax0, ax1, ax2, ax3, ax4 = axes
    ax0.plot(...)

···

On Thu, Apr 26, 2012 at 12:51 PM, willfurnass <will@...4073...> wrote:

I've converted a simple MATLAB script [1] for wavelet decomposition-based
analysis to Python. I now want to create figures similar to [2] that
feature five subplots in one column, with the 1st and 3rd being generated
using 'subplot.plot' and the others being generated using
'subplot.imshow'.

I want to find a way of scaling the x and y axes of all subplots so that
they're the same size on screen across all subplots. Unfortunately I
can't
find any combination of 'aspect' or 'extent' that will let me do this. I
should note that I've added colorbars to all my image plots; the problem
exists regardless of whether these are used.

[1]

http://www.mathworks.com/matlabcentral/fileexchange/18409-comparing-time-series-using-semblance-analysis
[2] http://www.mathworks.com/matlabcentral/fx_files/18409/1/goldoil.jpg

----

Hi Tony and others,

Thanks for the suggestion Tony. It works if I remove my colour bars.
However if the colour bars are enabled (for the images on the 2nd, 4th
and 5th rows) then the image on the 5th and last row is squashed up on
the LHS of the subplot area to make room for all 3 (!) colour bars,
even though all subplots should share an x axis. Any further
suggestions?

Also, I'm still having problems getting tick labels on ax.plot() and
ax.imshow() subplots to line up:

I create one subplot using plot(t_vec, y_vec)
I then create an image using t_vec and y_vec and display it as a second
subplot using imshow(image); the width of the image matrix is len(t_vec)
(not t_vec.max())

I am looking for a sane way of displaying the two subplots above each other
so that they are both the same width in pixels and have x tick labels at
consistent intervals (in pixels) that correspond to multiples of some value
of t.

I have tried using
ax.xaxis.set_major_formatter(ticker.FuncFormatter(xformat)) to scale the
image subplot ticks but can't get this to work in conjunction with the
setting of the tick intervals on both subplot axes using set_xticks. Has
anyone got any suggestions?

Regards,

Will
--
View this message in context: http://old.nabble.com/Set-x-axis-length-of-all-subplots-to-same-width-on-screen-tp33753799p33760460.html
Sent from the matplotlib - users mailing list archive at Nabble.com.