[Matplotlib-users] [Question] Possible bug in plotting large NumPy array

[Paul Hobson]

Dipu,

http://matplotlib.org/ has both a gallery if images (click the image to see
the source code) and a broad, organized list of examples.

I'd start there.

On Fri, Aug 19, 2016 at 7:58 AM, Dipankar “Dipu” Ganguly <dipugee at gmail.com>
wrote:

> Hi:
>
> I am trying to learn the basics of plotting in matplotlib. My immediate
> need is for displaying grayscale images with titles, axis labels & ticks
> and  within-image annotations. Where can I get good tutorials/demos for
> such things?
>
> Thanks.
>
> Dipu
>
>
> Dipankar “Dipu” Ganguly
> dipugee at gmail.com
> Cell: 408-203-8814
>
>
>
>
>
>
>
> On Aug 19, 2016, at 4:31 AM, <juhaszp95 at gmail.com> <juhaszp95 at gmail.com>
> wrote:
>
> Hi Jens,
>
> Thanks very much! This indeed works. I thought PyPlot automatically
> detects if the data is along which axes, but fair, expecting it in a
> certain way is more consistent.
>
> Many thanks for your help once again!
>
> Best wishes,
> Péter
>
> *From:* Jens Nielsen [mailto:jenshnielsen at gmail.com
> <jenshnielsen at gmail.com>]
> *Sent:* Friday, August 19, 2016 12:32 PM
> *To:* juhaszp95 at gmail.com; matplotlib-users at python.org
> *Subject:* Re: [Matplotlib-users] [Question] Possible bug in plotting
> large NumPy array
>
> The problem is the way you are slicing your data
>
> by doing y[:1] you are creating a 1 by 400 array. When you try to plot
> that you will get 400 individual line plots each with only one point. A
> line plot with one point is invisible since there in no other points to
> draw the lines between. You will not notice that with scatter since it
> defaults to drawing a point for each data point. If you add a marker you
> can see whats going on.
> I.e. do
> ax1.plot(x, y[:1], 'o')
> in the first example and you will notice that the points color cycle as a
> new plot is created.
>
> If you really want to slice this way you have to ensure that your data is
> along the first dimension.  I.e. you can do `ax1.plot(x.transpose(),
> y[:1].transpose())` which plots 400 by 1 arrays
>
> Hope that helps
> Jens
>
>
>
> On Fri, 19 Aug 2016 at 11:10 <juhaszp95 at gmail.com> wrote:
>
> Dear Matplotlib users,
>
> I am Péter Juhász and I started experimenting with this demo:
> http://matplotlib.org/examples/pylab_examples/subplots_demo.html.
>
> What I then tried to achieve is to let *y* be a NumPy array of 2 rows,
> the first holding sin(x**2) values, the second holding something else (I
> set cos(x**2)), and the make the same plots as in the original demo only by
> using the first row of *y*. Unfortunately, I was quite surprised to see
> that pyplot/matplotlib only succeeds with this plotting if *scatter plots*
>  or *markers* are used *only*, instead of the usual plotting with lines.
> On the other hand, with some tricks in the NumPy array (some reshapes and
> slicing) I was able to make pyplot/matplotlib work for both scatter plots
> and the usual simple plots. This however possibly takes much more time, so
> for big datasets does not seem feasible and anyway I do not see an obvious
> reason why a dataset should only work with scatter plots but not the usual,
> simple plots.
>
> Please find attached my code below and I would appreciate any
> help/explanation, or it is indeed a bug, then if it could be raised to the
> developers’ attention.
>
> My original trial, working only with scatter plots (see for yourself!
> maybe it’s only my installation what is going wrong?):
>
> import matplotlib.pyplot as plt
> import numpy as np
>
> # Simple data to display in various forms
> x = np.linspace(0, 2 * np.pi, 400).reshape(1, 400)
> print(x.shape)
> y = np.vstack((np.sin(x ** 2), np.cos(x ** 2)))
> print(y[:1].shape)
>
> plt.close(*'all'*)
>
> # Three *subplots* sharing both x/y axes
> f, (ax1, ax2, ax3) = plt.subplots(3, sharex=True, sharey=True)
> ax1.plot(x, y[:1]) # This is not showing!
> ax1.set_title(*'Sharing both axes'*)
> ax2.scatter(x, y[:1])
> ax3.scatter(x, 2 * y[:1] ** 2 - 1, color=*'r'*)
> # Fine-tune figure; make *subplots* close to each other and hide x ticks
> for
> # all but bottom plot.
> f.subplots_adjust(hspace=0)
> plt.setp([a.get_xticklabels() for a in f.axes[:-1]], visible=False)
>
> # row and column sharing
> f, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, sharex=*'col'*, sharey=
> *'row'*)
> ax1.plot(x, y[:1]) # This is not showing!
> ax1.set_title(*'Sharing x per column, y per row'*)
> ax2.scatter(x, y[:1])
> ax3.scatter(x, 2 * y[:1] ** 2 - 1, color=*'r'*)
> ax4.plot(x, 2 * y[:1] ** 2 - 1, color=*'r'*) # This is not showing!
>
> plt.show()
>
>
> The working solution:
>
> import matplotlib.pyplot as plt
> import numpy as np
>
> # Simple data to display in various forms
> x = np.linspace(0, 2 * np.pi, 400).reshape(400, 1)
> print(x.shape)
> y = np.vstack((np.sin(x ** 2), np.cos(x ** 2))).reshape(400, 2)
> print(y[:, 0].shape)
>
> plt.close(*'all'*)
>
> # Three *subplots* sharing both x/y axes
> f, (ax1, ax2, ax3) = plt.subplots(3, sharex=True, sharey=True)
> ax1.plot(x, y[:, 0])
> ax1.set_title(*'Sharing both axes'*)
> ax2.scatter(x, y[:, 0])
> ax3.scatter(x, 2 * y[:, 0] ** 2 - 1, color=*'r'*)
> # Fine-tune figure; make *subplots* close to each other and hide x ticks
> for
> # all but bottom plot.
> f.subplots_adjust(hspace=0)
> plt.setp([a.get_xticklabels() for a in f.axes[:-1]], visible=False)
>
> # row and column sharing
> f, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, sharex=*'col'*, sharey=
> *'row'*)
> ax1.plot(x, y[:, 0])
> ax1.set_title(*'Sharing x per column, y per row'*)
> ax2.scatter(x, y[:, 0])
> ax3.scatter(x, 2 * y[:, 0] ** 2 - 1, color=*'r'*)
> ax4.plot(x, 2 * y[:, 0] ** 2 - 1, color=*'r'*)
>
> plt.show()
>
> I did not modify anything else in the code respective to the original demo
> than what I mentioned. I hope you will be able to help.
>
> Kind regards,
> Péter
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users at python.org
> https://mail.python.org/mailman/listinfo/matplotlib-users
>
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users at python.org
> https://mail.python.org/mailman/listinfo/matplotlib-users
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users at python.org
> https://mail.python.org/mailman/listinfo/matplotlib-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160819/fc506df8/attachment-0001.html>