matplotlib-users Mailing List for matplotlib (Page 7)

Hi guys,
If I do the following, no plot shows:
ply5@xroa-dt-20:~> python
Python 2.6.9 (unknown, Apr 7 2015, 08:28:12)
[GCC 4.3.4 [gcc-4_3-branch revision 152973]] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import pylab as plt
>>> plt.figure()
<matplotlib.figure.Figure object at 0x16e5310>
>>> plt.hist([1.0,2,0])
(array([ 1., 0., 0., 0., 0., 1., 0., 0., 0., 1.]), array([ 0. ,
0.2, 0.4, 0.6, 0.8, 1. , 1.2, 1.4, 1.6, 1.8, 2. ]), <a list of 10
Patch objects>)
>>> plt.show
<function show at 0x1730b90>
>>>
I'm using SUSE Linux Enterprise 11 SP3 64-bit with python-matplotlib
v1.3.1-70.11.
Anyone have any ideas? I'm a bit of a matplotlib beginner so any advice is
extremely welcome!
Thanks,
Paul

The `get_axes` method (
https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/artist.py#L213
).
See
http://matplotlib.org/api/api_changes.html#prevent-moving-artists-between-axes-property-ify-artist-axes-deprecate-artist-get-set-axes
for documentation.
Tom
On Tue, Nov 24, 2015 at 11:43 AM Mads Ipsen <mad...@gm...> wrote:
> Hi,
>
> I get the following warning in some code I wrote for matplotlib-1.2.
>
>
> python2.7/site-packages/matplotlib-1.5.0-py2.7-linux-x86_64.egg/matplotlib/artist.py:221:
> MatplotlibDeprecationWarning: This has been deprecated in mpl 1.5,
> please use the
> axes property. A removal date has not been set.
> warnings.warn(_get_axes_msg, mplDeprecation, stacklevel=1)
>
> Clearly, I should use 'the axes property', but as a replacement for?
>
> The code base is relatively large, so it's not clear to me what triggers
> the warning.
>
> Best regards,
>
> Mads
>
> --
> +---------------------------------------------------------------------+
> | Mads Ipsen |
> +----------------------------------+----------------------------------+
> | Overgaden Oven Vandet 106, 4.tv | phone: +45-29716388 |
> | DK-1415 København K | email: mad...@gm... |
> | Denmark | map : https://goo.gl/maps/oQ6y6 |
> +----------------------------------+----------------------------------+
>
>
> ------------------------------------------------------------------------------
> Go from Idea to Many App Stores Faster with Intel(R) XDK
> Give your users amazing mobile app experiences with Intel(R) XDK.
> Use one codebase in this all-in-one HTML5 development environment.
> Design, debug & build mobile apps & 2D/3D high-impact games for multiple
> OSs.
> http://pubads.g.doubleclick.net/gampad/clk?id=254741551&iu=/4140
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi,
I get the following warning in some code I wrote for matplotlib-1.2.
python2.7/site-packages/matplotlib-1.5.0-py2.7-linux-x86_64.egg/matplotlib/artist.py:221: 
MatplotlibDeprecationWarning: This has been deprecated in mpl 1.5, 
please use the
axes property. A removal date has not been set.
 warnings.warn(_get_axes_msg, mplDeprecation, stacklevel=1)
Clearly, I should use 'the axes property', but as a replacement for?
The code base is relatively large, so it's not clear to me what triggers 
the warning.
Best regards,
Mads
-- 
+---------------------------------------------------------------------+
| Mads Ipsen |
+----------------------------------+----------------------------------+
| Overgaden Oven Vandet 106, 4.tv | phone: +45-29716388 |
| DK-1415 København K | email: mad...@gm... |
| Denmark | map : https://goo.gl/maps/oQ6y6 |
+----------------------------------+----------------------------------+

The way we do the alpha blending, the output value is (alpha * v1) +
((alpha-1) * v2). All of the artists are compsited down on top of a white
background so the compositing is not commutative.
(a * .5) + (.5 * (b * .5 + .5)) =/= (b * .5) + (.5 * (a * .5 + .5))
On Mon, Nov 23, 2015 at 1:55 PM Sterling Smith <sm...@fu...>
wrote:
> Maybe the issue is with the colormap not having an alpha? Does this
>
> http://stackoverflow.com/questions/10127284/overlay-imshow-plots-in-matplotlib
> help?
>
> Otherwise, you might file a bug at
> https://github.com/matplotlib/matplotlib/issues/new
>
> -Sterling
>
> On Nov 20, 2015, at 4:46PM, Brian Merchant <bhm...@gm...> wrote:
>
> > Hi all,
> >
> > In order to get circles such that their coloring is radially symmetric,
> with center being the darkest, and exponential decay in color as one moves
> farther away from the center along the radius, I used imshow with clip_path
> using Circle patches.
> >
> > Here's a toy script that overlaps two such circles:
> https://gist.github.com/bmer/7063cc2dd09f1b80a252
> >
> > As you can see if you run the script (or, if you follow this link:
> http://i.imgur.com/H9jEAZ3.png), even though the alpha is set at 0.5,
> there doesn't seem to be proper "color mixing" occurring (we should see a
> result that is symmetric along the x-axis).
> >
> > Why is that, and what could I do to fix this issue?
> >
> > Kind regards,
> > Brian
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
> ------------------------------------------------------------------------------
> Go from Idea to Many App Stores Faster with Intel(R) XDK
> Give your users amazing mobile app experiences with Intel(R) XDK.
> Use one codebase in this all-in-one HTML5 development environment.
> Design, debug & build mobile apps & 2D/3D high-impact games for multiple
> OSs.
> http://pubads.g.doubleclick.net/gampad/clk?id=254741551&iu=/4140
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Maybe the issue is with the colormap not having an alpha? Does this
http://stackoverflow.com/questions/10127284/overlay-imshow-plots-in-matplotlib
help?
Otherwise, you might file a bug at 
https://github.com/matplotlib/matplotlib/issues/new
-Sterling
On Nov 20, 2015, at 4:46PM, Brian Merchant <bhm...@gm...> wrote:
> Hi all, 
> 
> In order to get circles such that their coloring is radially symmetric, with center being the darkest, and exponential decay in color as one moves farther away from the center along the radius, I used imshow with clip_path using Circle patches. 
> 
> Here's a toy script that overlaps two such circles: https://gist.github.com/bmer/7063cc2dd09f1b80a252
> 
> As you can see if you run the script (or, if you follow this link: http://i.imgur.com/H9jEAZ3.png), even though the alpha is set at 0.5, there doesn't seem to be proper "color mixing" occurring (we should see a result that is symmetric along the x-axis). 
> 
> Why is that, and what could I do to fix this issue? 
> 
> Kind regards, 
> Brian
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi all,
In order to get circles such that their coloring is radially symmetric,
with center being the darkest, and exponential decay in color as one moves
farther away from the center along the radius, I used imshow with clip_path
using Circle patches.
Here's a toy script that overlaps two such circles:
https://gist.github.com/bmer/7063cc2dd09f1b80a252
As you can see if you run the script (or, if you follow this link:
http://i.imgur.com/H9jEAZ3.png), even though the alpha is set at 0.5, there
doesn't seem to be proper "color mixing" occurring (we should see a result
that is symmetric along the x-axis).
Why is that, and what could I do to fix this issue?
Kind regards,
Brian

Hello,
after looking at the code I found that there are two settings that are
not documented (they are not in the template matplotlibrc) that control
the legend frame: "legend.edgecolor" and "legend.facecolor".
The first is enough for my needs, however I think it would be nice to
add the possibility to control the frame linewidth. I have a small patch
that add a "legend.linewidth" setting that defaults to "inherit",
meaning that the legend frame has the same width of the axes. The
setting name is not the most intuitive, but it is coherent with the
existing "legend.edgecolor" and "legend.facecolor".
It would be better to replace "legend.frmaeon" with a boolean setting
"legend.frame and control the frame with "legend.frame.edgecolor",
"legend.frame.facecolor" and "legend.frame.linewidth". But this would
require deprecating the old settings. I can prepare a patch if this is
desired.
Cheers,
Daniele
On 13/11/15 17:08, Benjamin Root wrote:
> Indeed, it looks like there isn't a very good way to control all of the
> properties of the frame portion of a legend. This could certainly use
> some improvements, partly in allowing a dictionary of property values to
> be passed in `plt.legend()` (there is already a dictionary of font
> properties), but also to have some rcParams that could be made
> available, too. Such improvements are always welcome!
> 
> Cheers!
> Ben Root
> 
> On Fri, Nov 13, 2015 at 7:57 AM, Daniele Nicolodi <da...@gr...
> <mailto:da...@gr...>> wrote:
> 
> Hello,
> 
> there is a way to control the edgecolor and the linewidth of the frame
> drawn around the legend? I set the axes linewidth to 0.5 but the legend
> frame linewidth is set to 1.0 and it does not look nice. Also, most of
> the time I don't want the frame edge to be drawn at all.
> 
> Always doing:
> 
> l = plt.legend()
> l.get_frame().set_edgecolor('none')
> 
> is boring.
> 
> If the setting are not there, would a patch adding a setting to style
> the legend frame be considered? Additionally, would it make sense to
> default the legend frame linewidth to the axes linewidth?
> 
> Cheers,
> Daniele
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> <mailto:Mat...@li...>
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 

Hello,
I use matplotlib 1.5.0 via macports on MacOSX 10.10.5 and I'm testing
the TkAgg backend (the MacOSX backend has some annoying bugs).
I notice the following strange behavior in a IPython 4.0.0 console:
1. run the following lines:
In [1]: %matplotlib tk
In [2]: import matplotlib.pyplot as plt
In [3]: plt.figure()
2. focus the figure window (with command-tab, for example)
3. press command-w to close the window, the window closes
4. focus the terminal window (with command-tab, for example)
5. focus the Python app which still exist (with command-tab, for
example), the figure window re-appears (!!) but it is not functional
6. cycle the focus between the windows of the Python app (with
command-`), the figure window will be interactive again
7. try again to close the window (clicking the 'x' or again with
command-w), the window disappears
8. back to 4
I haven't found a way to get rid of the window. The MacOSX backe works
correctly in this respect.
Can someone else on MacOSX confirm the bug?
Where should I start debugging?
Thanks. Best,
Daniele

I needed an apng viewer plug-in for Chrome
https://chrome.google.com/webstore/detail/apng/ehkepjiconegkhpodgoaeamnpckdbblp?hl=en
-Sterling
 
On Nov 13, 2015, at 1:31PM, Warren Weckesser <war...@gm...> wrote:
> Matplotlib users,
> 
> I just put the package "numpngw" up on pypi: https://pypi.python.org/pypi/numpngw
> The development version is on github: https://github.com/WarrenWeckesser/numpngw
> 
> The reason this might be of interest to maplotlib users is the class "numpngw.AnimatedPNGWriter". This class can be used as the "writer" argument of the "save" method of the Animation class. So if you've ever wanted to use matplotlib to create an animated PNG (and who hasn't?), now you can!
> 
> If you go to the github page, scroll down to Example 8 to see an example of how to save an animation as an animated PNG. You'll need a browser that supports animated PNG to actually see the animation. Firefox does, Safari doesn't, and I haven't checked any others.
> 
> If you use it and find problems or have suggestions for improvements, email me or create an issue on github.
> 
> Warren
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Matplotlib users,
I just put the package "numpngw" up on pypi:
https://pypi.python.org/pypi/numpngw
The development version is on github:
https://github.com/WarrenWeckesser/numpngw
The reason this might be of interest to maplotlib users is the class
"numpngw.AnimatedPNGWriter". This class can be used as the "writer"
argument of the "save" method of the Animation class. So if you've ever
wanted to use matplotlib to create an animated PNG (and who hasn't?), now
you can!
If you go to the github page, scroll down to Example 8 to see an example of
how to save an animation as an animated PNG. You'll need a browser that
supports animated PNG to actually see the animation. Firefox does, Safari
doesn't, and I haven't checked any others.
If you use it and find problems or have suggestions for improvements, email
me or create an issue on github.
Warren

Indeed, it looks like there isn't a very good way to control all of the
properties of the frame portion of a legend. This could certainly use some
improvements, partly in allowing a dictionary of property values to be
passed in `plt.legend()` (there is already a dictionary of font
properties), but also to have some rcParams that could be made available,
too. Such improvements are always welcome!
Cheers!
Ben Root
On Fri, Nov 13, 2015 at 7:57 AM, Daniele Nicolodi <da...@gr...>
wrote:
> Hello,
>
> there is a way to control the edgecolor and the linewidth of the frame
> drawn around the legend? I set the axes linewidth to 0.5 but the legend
> frame linewidth is set to 1.0 and it does not look nice. Also, most of
> the time I don't want the frame edge to be drawn at all.
>
> Always doing:
>
> l = plt.legend()
> l.get_frame().set_edgecolor('none')
>
> is boring.
>
> If the setting are not there, would a patch adding a setting to style
> the legend frame be considered? Additionally, would it make sense to
> default the legend frame linewidth to the axes linewidth?
>
> Cheers,
> Daniele
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hello,
there is a way to control the edgecolor and the linewidth of the frame
drawn around the legend? I set the axes linewidth to 0.5 but the legend
frame linewidth is set to 1.0 and it does not look nice. Also, most of
the time I don't want the frame edge to be drawn at all.
Always doing:
 l = plt.legend()
 l.get_frame().set_edgecolor('none')
is boring.
If the setting are not there, would a patch adding a setting to style
the legend frame be considered? Additionally, would it make sense to
default the legend frame linewidth to the axes linewidth?
Cheers,
Daniele

I want to plot contours on top of an image, using twinx and twiny for the
contour coordinates. I don't want to show the x- and yticklabels for the
contour coordinates. I find I can remove the xticklabels, but not the
yticklabels. There is mention of this problem here:
http://stackoverflow.com/questions/12358890/matplotlib-using-twinx-and-twiny-together-like-twinxy
but I wonder if there is now a solution?
The lines below illustrate the problem.
import matplotlib.pyplot as plt
import numpy as np
x1 = np.arange(10)
y1 = np.arange(20)
data1 = np.arange(x1.size * y1.size).reshape(y1.size, x1.size)
x2 = np.arange(10, 40)
y2 = np.arange(-20, 10)
data2 = np.sin(np.arange(x2.size * y2.size).reshape(x2.size, x2.size))
fig = plt.figure()
ax = fig.add_subplot(111)
ax.pcolormesh(x1, y1, data1)
ax2 = ax.twinx().twiny()
ax2.contour(x2, y2, data2, colors='k')
plt.setp(ax2.get_xticklabels(), visible=False)
# Neither of the lines below work
#plt.setp(ax2.get_yticklabels(), visible=False)
plt.setp(ax2.axes.yaxis.get_ticklabels(), visible=False)
plt.show()
Thanks!

If that fails that is a bug in PyQt/Qt or in your anaconda installation and
not a Matplotlib bug. You could try reinstall QT and PyQt.
As a workaround you can tell Matplotlib to use a different backend
http://matplotlib.org/faq/usage_faq.html#what-is-a-backend
BTW the matplotlib mailing list has changed to mat...@py....
best
Jens
fre. 30. okt. 2015 kl. 12.43 skrev Jonno <jon...@gm...>:
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> ImportError: DLL load failed: The specified module could not be found.
>
> On Fri, Oct 30, 2015 at 2:26 AM, Jens Nielsen <jen...@gm...>
> wrote:
>
>> It sounds like your PyQt package is broken.
>>
>> What happens if you do:
>>
>> from PyQt4 import QtCore, QtGui
>>
>> in a python shell
>>
>> /Jens
>>
>> fre. 30. okt. 2015 kl. 03.06 skrev Jonno <jon...@gm...>:
>>
>>> Not sure where to post this.
>>>
>>> I have a fresh Anaconda Win64 python 2.7.10 install which I then updated
>>> using conda update --all.
>>>
>>> If it try to:
>>> from pylab import *
>>> I get the following:
>>>
>>>
>>> File "~\Anaconda\lib\site-packages\matplotlib\backends\qt_compat.py",
>>> line 91, in <module>
>>> from PyQt4 import QtCore, QtGui
>>> ImportError: DLL load failed: The specified module could not be found.
>>>
>>>
>>> I have the following installed:
>>> qt: 4.8.7
>>> pyqt 4.11.4
>>> matplotlib 1.4.3
>>>
>>> Should I open an issue on matplotlib github?
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>
>>
>

Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
ImportError: DLL load failed: The specified module could not be found.
On Fri, Oct 30, 2015 at 2:26 AM, Jens Nielsen <jen...@gm...>
wrote:
> It sounds like your PyQt package is broken.
>
> What happens if you do:
>
> from PyQt4 import QtCore, QtGui
>
> in a python shell
>
> /Jens
>
> fre. 30. okt. 2015 kl. 03.06 skrev Jonno <jon...@gm...>:
>
>> Not sure where to post this.
>>
>> I have a fresh Anaconda Win64 python 2.7.10 install which I then updated
>> using conda update --all.
>>
>> If it try to:
>> from pylab import *
>> I get the following:
>>
>>
>> File "~\Anaconda\lib\site-packages\matplotlib\backends\qt_compat.py",
>> line 91, in <module>
>> from PyQt4 import QtCore, QtGui
>> ImportError: DLL load failed: The specified module could not be found.
>>
>>
>> I have the following installed:
>> qt: 4.8.7
>> pyqt 4.11.4
>> matplotlib 1.4.3
>>
>> Should I open an issue on matplotlib github?
>>
>>
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>

It sounds like your PyQt package is broken.
What happens if you do:
from PyQt4 import QtCore, QtGui
in a python shell
/Jens
fre. 30. okt. 2015 kl. 03.06 skrev Jonno <jon...@gm...>:
> Not sure where to post this.
>
> I have a fresh Anaconda Win64 python 2.7.10 install which I then updated
> using conda update --all.
>
> If it try to:
> from pylab import *
> I get the following:
>
>
> File "~\Anaconda\lib\site-packages\matplotlib\backends\qt_compat.py", line
> 91, in <module>
> from PyQt4 import QtCore, QtGui
> ImportError: DLL load failed: The specified module could not be found.
>
>
> I have the following installed:
> qt: 4.8.7
> pyqt 4.11.4
> matplotlib 1.4.3
>
> Should I open an issue on matplotlib github?
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Not sure where to post this.
I have a fresh Anaconda Win64 python 2.7.10 install which I then updated
using conda update --all.
If it try to:
from pylab import *
I get the following:
File "~\Anaconda\lib\site-packages\matplotlib\backends\qt_compat.py", line
91, in <module>
 from PyQt4 import QtCore, QtGui
ImportError: DLL load failed: The specified module could not be found.
I have the following installed:
qt: 4.8.7
pyqt 4.11.4
matplotlib 1.4.3
Should I open an issue on matplotlib github?

Wonderful! thank you!
On Fri, Oct 30, 2015 at 4:51 AM, Joshua Klein <mob...@gm...> wrote:
> My mistake, I thought you were using a DataFrame, not a Series. Instead do
> this
>
> colors = ['r' if row > 0 else 'b' for i, row in meantempanomaly.iteritems()]
> meantempanomaly.plot(kind='bar', color=colors)
>
> ​
>
> On Thu, Oct 29, 2015 at 5:43 AM, questions anon <que...@gm...>
> wrote:
>
>> Thanks for taking the time to respond
>>
>> I am receiving the error:
>> AttributeError: 'Series' object has no attribute 'iterrow'
>>
>> I will look into this further.
>> thank you
>>
>>
>> On Thu, Oct 29, 2015 at 2:31 PM, Joshua Klein <mob...@gm...>
>> wrote:
>>
>>> The pandas plot function doesn’t take colors as it does ‘x’ or ‘y’, but
>>> it lets you pass color information just as you would with raw matplotlib
>>> code, which means you can pass it a sequence of colors which match the
>>> length of your sequence of drawn observations.
>>>
>>> # compute color codes using a ternary expression in a list comprehension over the DataFrame
>>> colors = ['r' if row.anomaly > 0 else 'b' for i, row in meantempanomaly.iterrows()]
>>> meantempanomaly.plot(kind='bar', color=colors)
>>>
>>> ​
>>>
>>> On Wed, Oct 28, 2015 at 9:54 PM, questions anon <
>>> que...@gm...> wrote:
>>>
>>>> I have calculated annual temperature anomaly and I would like to plot
>>>> as a bar plot with all values positive make red and all values negative
>>>> make blue
>>>>
>>>> I am using pandas and the time series data in this example are called
>>>> 'anomaly'
>>>>
>>>> mybarplot=anomaly.plot(kind='bar')
>>>>
>>>> the data look like this:
>>>>
>>>> time
>>>> 2003年01月01日 -0.370800
>>>> 2004年01月01日 -0.498199
>>>> 2005年01月01日 0.246118
>>>> 2006年01月01日 -0.313321
>>>> 2007年01月01日 0.585050
>>>> 2008年01月01日 -0.227976
>>>> 2009年01月01日 0.439337
>>>> 2010年01月01日 0.135607
>>>> 2011年01月01日 0.106105
>>>> 2012年01月01日 -0.102002
>>>> Freq: AS-JAN, dtype: float64
>>>>
>>>> is there some simple way of writing:
>>>> meantempanomaly.plot(kind='bar', anomaly>0:'r', anomaly<0:'b' )
>>>>
>>>> Any feedback will be greatly appreciated
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> Matplotlib-users mailing list
>>>> Mat...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>>
>>>>
>>>
>>
>

My mistake, I thought you were using a DataFrame, not a Series. Instead do
this
colors = ['r' if row > 0 else 'b' for i, row in meantempanomaly.iteritems()]
meantempanomaly.plot(kind='bar', color=colors)
​
On Thu, Oct 29, 2015 at 5:43 AM, questions anon <que...@gm...>
wrote:
> Thanks for taking the time to respond
>
> I am receiving the error:
> AttributeError: 'Series' object has no attribute 'iterrow'
>
> I will look into this further.
> thank you
>
>
> On Thu, Oct 29, 2015 at 2:31 PM, Joshua Klein <mob...@gm...>
> wrote:
>
>> The pandas plot function doesn’t take colors as it does ‘x’ or ‘y’, but
>> it lets you pass color information just as you would with raw matplotlib
>> code, which means you can pass it a sequence of colors which match the
>> length of your sequence of drawn observations.
>>
>> # compute color codes using a ternary expression in a list comprehension over the DataFrame
>> colors = ['r' if row.anomaly > 0 else 'b' for i, row in meantempanomaly.iterrows()]
>> meantempanomaly.plot(kind='bar', color=colors)
>>
>> ​
>>
>> On Wed, Oct 28, 2015 at 9:54 PM, questions anon <que...@gm...
>> > wrote:
>>
>>> I have calculated annual temperature anomaly and I would like to plot as
>>> a bar plot with all values positive make red and all values negative make
>>> blue
>>>
>>> I am using pandas and the time series data in this example are called
>>> 'anomaly'
>>>
>>> mybarplot=anomaly.plot(kind='bar')
>>>
>>> the data look like this:
>>>
>>> time
>>> 2003年01月01日 -0.370800
>>> 2004年01月01日 -0.498199
>>> 2005年01月01日 0.246118
>>> 2006年01月01日 -0.313321
>>> 2007年01月01日 0.585050
>>> 2008年01月01日 -0.227976
>>> 2009年01月01日 0.439337
>>> 2010年01月01日 0.135607
>>> 2011年01月01日 0.106105
>>> 2012年01月01日 -0.102002
>>> Freq: AS-JAN, dtype: float64
>>>
>>> is there some simple way of writing:
>>> meantempanomaly.plot(kind='bar', anomaly>0:'r', anomaly<0:'b' )
>>>
>>> Any feedback will be greatly appreciated
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>
>>>
>>
>

An axes can only belong to one figure at a time. And I also don't think I
have ever seen anyone try and transfer an axes from one figure to another.
You *might* have luck with inset locators from axes_grid:
http://matplotlib.org/examples/axes_grid/inset_locator_demo.html
Cheers!
Ben Root
On Thu, Oct 29, 2015 at 12:07 PM, Alejandro Weinstein <
ale...@gm...> wrote:
> Hi,
>
> I have a previously draw plot that I want to place as an inset in
> another figure. I've tried with passing the previously drawn axes as
> the `axes` parameter to the `add_axes` method of the figure, and also
> tried using the `set_axes` method of the new axes, without success: I
> get the new inset axes, but without the previously drawn plot, in both
> cases.
>
> The following code shows both approaches:
>
> # Passing the inset axes as a parameter to add_axes
> fig_in, ax_in = plt.subplots()
> ax_in.plot([1,2,3])
> fig, ax = plt.subplots()
> ax.plot([1,2,1])
> fig.add_axes([0.72, 0.72, 0.16, 0.16], axes=ax_in)
>
> # Using set_axes
> fig_in, ax_in = plt.subplots()
> ax_in.plot([1,2,3])
> fig, ax = plt.subplots()
> ax.plot([1,2,1])
> ax_new = fig.add_axes([0.72, 0.72, 0.16, 0.16])
> ax_new.set_axes(ax_in)
>
> Any help with this will be appreciated.
>
> Regards,
> Alejandro
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi,
I have a previously draw plot that I want to place as an inset in
another figure. I've tried with passing the previously drawn axes as
the `axes` parameter to the `add_axes` method of the figure, and also
tried using the `set_axes` method of the new axes, without success: I
get the new inset axes, but without the previously drawn plot, in both
cases.
The following code shows both approaches:
 # Passing the inset axes as a parameter to add_axes
 fig_in, ax_in = plt.subplots()
 ax_in.plot([1,2,3])
 fig, ax = plt.subplots()
 ax.plot([1,2,1])
 fig.add_axes([0.72, 0.72, 0.16, 0.16], axes=ax_in)
 # Using set_axes
 fig_in, ax_in = plt.subplots()
 ax_in.plot([1,2,3])
 fig, ax = plt.subplots()
 ax.plot([1,2,1])
 ax_new = fig.add_axes([0.72, 0.72, 0.16, 0.16])
 ax_new.set_axes(ax_in)
Any help with this will be appreciated.
Regards,
Alejandro

Thanks for taking the time to respond
I am receiving the error:
AttributeError: 'Series' object has no attribute 'iterrow'
I will look into this further.
thank you
On Thu, Oct 29, 2015 at 2:31 PM, Joshua Klein <mob...@gm...> wrote:
> The pandas plot function doesn’t take colors as it does ‘x’ or ‘y’, but it
> lets you pass color information just as you would with raw matplotlib code,
> which means you can pass it a sequence of colors which match the length of
> your sequence of drawn observations.
>
> # compute color codes using a ternary expression in a list comprehension over the DataFrame
> colors = ['r' if row.anomaly > 0 else 'b' for i, row in meantempanomaly.iterrows()]
> meantempanomaly.plot(kind='bar', color=colors)
>
> ​
>
> On Wed, Oct 28, 2015 at 9:54 PM, questions anon <que...@gm...>
> wrote:
>
>> I have calculated annual temperature anomaly and I would like to plot as
>> a bar plot with all values positive make red and all values negative make
>> blue
>>
>> I am using pandas and the time series data in this example are called
>> 'anomaly'
>>
>> mybarplot=anomaly.plot(kind='bar')
>>
>> the data look like this:
>>
>> time
>> 2003年01月01日 -0.370800
>> 2004年01月01日 -0.498199
>> 2005年01月01日 0.246118
>> 2006年01月01日 -0.313321
>> 2007年01月01日 0.585050
>> 2008年01月01日 -0.227976
>> 2009年01月01日 0.439337
>> 2010年01月01日 0.135607
>> 2011年01月01日 0.106105
>> 2012年01月01日 -0.102002
>> Freq: AS-JAN, dtype: float64
>>
>> is there some simple way of writing:
>> meantempanomaly.plot(kind='bar', anomaly>0:'r', anomaly<0:'b' )
>>
>> Any feedback will be greatly appreciated
>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>

The pandas plot function doesn’t take colors as it does ‘x’ or ‘y’, but it
lets you pass color information just as you would with raw matplotlib code,
which means you can pass it a sequence of colors which match the length of
your sequence of drawn observations.
# compute color codes using a ternary expression in a list
comprehension over the DataFrame
colors = ['r' if row.anomaly > 0 else 'b' for i, row in
meantempanomaly.iterrows()]
meantempanomaly.plot(kind='bar', color=colors)
​
On Wed, Oct 28, 2015 at 9:54 PM, questions anon <que...@gm...>
wrote:
> I have calculated annual temperature anomaly and I would like to plot as a
> bar plot with all values positive make red and all values negative make blue
>
> I am using pandas and the time series data in this example are called
> 'anomaly'
>
> mybarplot=anomaly.plot(kind='bar')
>
> the data look like this:
>
> time
> 2003年01月01日 -0.370800
> 2004年01月01日 -0.498199
> 2005年01月01日 0.246118
> 2006年01月01日 -0.313321
> 2007年01月01日 0.585050
> 2008年01月01日 -0.227976
> 2009年01月01日 0.439337
> 2010年01月01日 0.135607
> 2011年01月01日 0.106105
> 2012年01月01日 -0.102002
> Freq: AS-JAN, dtype: float64
>
> is there some simple way of writing:
> meantempanomaly.plot(kind='bar', anomaly>0:'r', anomaly<0:'b' )
>
> Any feedback will be greatly appreciated
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

I have calculated annual temperature anomaly and I would like to plot as a
bar plot with all values positive make red and all values negative make blue
I am using pandas and the time series data in this example are called
'anomaly'
mybarplot=anomaly.plot(kind='bar')
the data look like this:
time
2003年01月01日 -0.370800
2004年01月01日 -0.498199
2005年01月01日 0.246118
2006年01月01日 -0.313321
2007年01月01日 0.585050
2008年01月01日 -0.227976
2009年01月01日 0.439337
2010年01月01日 0.135607
2011年01月01日 0.106105
2012年01月01日 -0.102002
Freq: AS-JAN, dtype: float64
is there some simple way of writing:
meantempanomaly.plot(kind='bar', anomaly>0:'r', anomaly<0:'b' )
Any feedback will be greatly appreciated

Wow, that's fantastic Ben. Thanks so much for finding that, it's just 
what I need!
Regards,
- Phil
Benjamin Root wrote:
> Looks like someone else figured out a creative solution using quiver: 
> http://stackoverflow.com/questions/19918502/sawtooth-line-style-in-matplotlib
>
> Here it is (slightly cleaned up):
>
> |import matplotlib.pyplotas plt
> import numpyas np
>
> x= np.linspace(0, 2*np.pi, 100)
> y= np.sin(x)
>
> dx= np.diff(x)
> dy= np.diff(y)
>
> x2= np.linspace(0, 2*np.pi, 10)
> y2= np.sin(x2)
>
> dx= np.zeros_like(x2) + 1e-12
> dy= np.sin(x2+dx) - y2
>
> length= np.hypot(dx,dy)
> dx/= length
> dy/= length
>
> fig, ax= plt.subplots()
> ax.set_aspect("equal")
> ax.plot(x, y, lw=4)
>
> size= 20
> ax.quiver(x2, y2, -dy, dx, headaxislength=size, headlength=size, headwidth=size, color="blue")
> plt.margins(0.2)|
>
> I don't know yet how to get rounded heads, though. Now I am looking to 
> see how the text box styles of "sawtooth" and "roundtooth" are handled 
> in the code to see if that could be exploited, instead.
>
> Cheers!
> Ben Root
>
>
>
> On Fri, Oct 16, 2015 at 10:24 AM, Benjamin Root <ben...@gm... 
> <mailto:ben...@gm...>> wrote:
>
> Hmmm, this is actually an interesting problem. I am also a
> meteorologist, so this is interesting to me.
>
> I haven't figured it out yet, but here are my thoughts:
>
> 1) There are the "^" triangle markers as well as "2" tri_up
> markers:
> http://nbviewer.ipython.org/github/WeatherGod/AnatomyOfMatplotlib/blob/master/AnatomyOfMatplotlib-Part3-HowToSpeakMPL.ipynb#Markers
> 2) The markevery property should be set to a float value to have
> the markers spaced out evenly along the line regardless of aspect
> ratios and zooming (note, this assumes that the line is defined
> with many vertices to give a smooth appearance).
>
> Problem:
> Using markers and markevery in a Line2D object has an inherent
> limitation: all of the markers will be drawn in the same
> orientation. So, we can't orient the markers along the normal of
> the line.
> Also, there is no pre-defined marker for half-circles, so this
> approach wouldn't work well for warm-fronts/dry-lines/etc.
>
> I'll have to see if a PolygonCollection + Line2D might be the
> right approach here...
>
> Ben Root
>
>
>
>
> On Fri, Oct 16, 2015 at 7:22 AM, Phil Cummins
> <phi...@an... <mailto:phi...@an...>> wrote:
>
> Hi,
>
> I would like to plot "toothed" curves using basemap. These are
> curves with triangles on one side, that are used to plot
> pressure fronts in meteorology or thrust faults in geology.
> You need to be able to say which side of the curve the
> triangles should appear on. Does anyone know whether such
> curves can be plotted using mtplotlib/basemap?
>
> Thanks,
>
> - Phil
>
> Australian National University
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> <mailto:Mat...@li...>
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
-- 
Phil Cummins
Prof. Natural Hazards
Research School of Earth Sciences
Australian National University