matplotlib-users Mailing List for matplotlib (Page 2)

On Tue, Jul 21, 2009 at 18:00, Sandro Tosi<mat...@gm...> wrote:
> for example, let's say I want to plot this series
>
> s1 = 2,3,6,3,1
> s2 = 1,2,2,4,1
> s3 = 4,1,0,3,7
>
> what's the format of data to pass to hist() ? by row? by column?

Just to clarify: from the above lists of values, I'm expecting to draw
something like:

          |
  s3      | s3[1]  s3[2]  s3[3]  ...
  s2  ->  | s2[1]  s2[2]  s2[3]  ...
  s1      | s1[1]  s1[2]  s1[3]  ...
        --+---------------------------
          |

or

                |  ...    ...    ...
                | s1[3]  s2[3]  s3[3]
                | s1[2]  s2[2]  s3[2]
 s1 s2 s3   ->  | s1[1]  s2[1]  s3[1]
              --+----------------------
                |

Either are fine (once I get how to do, I'll adapt data to that),
simply I don't know how to do that :)

Cheers,
-- 
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi

davide lasagna wrote:
> Thanks Eric for pointing out this. However i've found that psd and
> specgram plot 10*log10*(Pxx/max(Pxx)), so i'll have a range from 0 db
> down. 
Aha, I wasn't looking closely enough!

> Anyway, what do you mean by "should not be axes method"?
That was a side remark about what the Axes class should do or not do. I 
would prefer that it stick to plotting.  For example, at present it does 
not have a method for empirical orthogonal functions, or for 
wavelets--and that's good, it shouldn't.  Numerical code (usually in 
numpy or scipy) should do the calculations, and then the result should 
be plotted using Axes methods or pyplot functions for line plotting, 
contouring, making images, etc.  So getting back to the specgram, I 
would prefer that it be given as an example--a script combining the 
calculation with the plotting--rather than having calculation and 
plotting done in an Axes method.
> Do you mean using the array output of specgram and psd and plotting that
> result by myself?

The calculations are being done by functions in the matplotlib.mlab 
module, so yes, I suggest using those functions directly (not the pyplot 
functions that wrap the Axes methods that call the mlab functions and 
then plot the result) and then doing your own scaling and plotting.  The 
Axes specgram method might guide you in the latter (giving an example of 
how to use imshow, for example), but if you write your own function you 
can make it do exactly what you want instead of accepting the scaling 
choice imposed by the Axes method.

Eric

> 
> Cheers
> 
> Davide
> On Tue, 2009-07-21 at 09:12 -1000, Eric Firing wrote:
>> davide lasagna wrote:
>>> Hello everybody,
>>> this is my first post in this list. 
>>>
>>> I'm plotting a spectrogram with
>>>
>>> Pxx, freqs, bins, im = specgram(y, nfft=256, f_sampling=12000)
>>>
>>> and i want to add a colorbar with
>>>
>>> colorbar()
>>>
>>>
>>> The problem is that the color scale seems to be wrong with respect to
>>> the data in Pxx, i.e. Pxx is of the order of 1e-2 while in the colorbar
>>> i have tick values spanning from -20 to -180. What is the problem??
>> specgram and psd plot power in decibels: 10*log10(Pxx).
>>
>> Because they involve heavy computation and application-specific choices, 
>> like the decibel scale, they really should not be axes methods.
>>
>> You can make a function plot the spectrogram however you like by looking 
>> at the source code in the Axes.specgram method (the pyplot specgram 
>> function is just a thin wrapper for this method), which is using 
>> mlab.specgram to do the calculation and is then just scaling and plotting.
>>
>> Eric
>>
>>
>>> Thanks in advance!
>>>
>>>
>>> Davide
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Enter the BlackBerry Developer Challenge  
>>> This is your chance to win up to $100,000 in prizes! For a limited time, 
>>> vendors submitting new applications to BlackBerry App World(TM) will have
>>> the opportunity to enter the BlackBerry Developer Challenge. See full prize  
>>> details at: http://p.sf.net/sfu/Challenge
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

2009/7/21 John Hunter <jd...@gm...>

>
> > I have the suspect that matplotlib windows support simply doesn't exist:)
> > I'll try prepackaged windows distributions like EPD or Python(x,y) and
> hope
> > this will solve this issue...
>
>
> I'm pretty sure they do not package mpl with the gtk backend, but
> Christoph Gohlke has a version that does
>
> http://www.lfd.uci.edu/~gohlke/#pythonlibs
>
Oh, finally...we have found a package with gtk support. Thanks. I will try
it. Even if I've already downloaded 240 MB of EDP and I'd like to try it as
well.
Thank you very much to you two!

Thanks Eric for pointing out this. However i've found that psd and
specgram plot 10*log10*(Pxx/max(Pxx)), so i'll have a range from 0 db
down. 
Anyway, what do you mean by "should not be axes method"?
Do you mean using the array output of specgram and psd and plotting that
result by myself?

Cheers

Davide
On Tue, 2009-07-21 at 09:12 -1000, Eric Firing wrote:
> davide lasagna wrote:
> > Hello everybody,
> > this is my first post in this list. 
> > 
> > I'm plotting a spectrogram with
> > 
> > Pxx, freqs, bins, im = specgram(y, nfft=256, f_sampling=12000)
> > 
> > and i want to add a colorbar with
> > 
> > colorbar()
> > 
> > 
> > The problem is that the color scale seems to be wrong with respect to
> > the data in Pxx, i.e. Pxx is of the order of 1e-2 while in the colorbar
> > i have tick values spanning from -20 to -180. What is the problem??
> 
> specgram and psd plot power in decibels: 10*log10(Pxx).
> 
> Because they involve heavy computation and application-specific choices, 
> like the decibel scale, they really should not be axes methods.
> 
> You can make a function plot the spectrogram however you like by looking 
> at the source code in the Axes.specgram method (the pyplot specgram 
> function is just a thin wrapper for this method), which is using 
> mlab.specgram to do the calculation and is then just scaling and plotting.
> 
> Eric
> 
> 
> > 
> > Thanks in advance!
> > 
> > 
> > Davide
> > 
> > 
> > ------------------------------------------------------------------------------
> > Enter the BlackBerry Developer Challenge  
> > This is your chance to win up to $100,000 in prizes! For a limited time, 
> > vendors submitting new applications to BlackBerry App World(TM) will have
> > the opportunity to enter the BlackBerry Developer Challenge. See full prize  
> > details at: http://p.sf.net/sfu/Challenge
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 

Hi,

I'm worried that I'm doing something stupid, but can't quite spot it.

testBarCharts() :- X axis in integers. Works fine.
testBarChartsDTMonths() :- X axis in datetimes, 1 month between data
points. Works fine
testBarChartsDTHours() :- X axis in datetimes, 1 hour between data
points. Bars seem to extend as far to the right as possible. When
printing the list of patches returned from plotting, the widths are
reported as 0.8, though.

Am I screwing up or is this a matplotlib bug? Using
matplotlib-0.98.5.3.win32-py2.6.exe and
numpy-1.3.0-win32-superpack-python2.6.exe

Many thanks,

Nick



import matplotlib.pyplot as plt
import matplotlib.dates
import datetime

# This works fine
def testBarCharts():
    xs = [10,11,12,13]
    ys = [100,300,200,600]

    fig = plt.figure(figsize=(20,12)) # dims in inches
    ax1 = fig.add_subplot(111)

    ax1.set_ylabel("foo")

    rects = ax1.bar(xs,ys,color=(1,0,0))
    # debugging
    print xs
    print ys
    for r in rects:
        print r.get_xy(),r.get_width(),"x",r.get_height()

    fig.savefig("foo.png")
    plt.close(fig)



# FIXME: This is all wonky
def testBarChartsDTHours():
    xs = [ \
                        datetime.datetime(2009,6,29,10),
                        datetime.datetime(2009,6,29,11),
                        datetime.datetime(2009,6,29,12),
                        datetime.datetime(2009,6,29,13) ]
    ys = [100,300,200,600]

    fig = plt.figure(figsize=(20,12)) # dims in inches
    ax1 = fig.add_subplot(111)

    ax1.set_ylabel("foo")

    ax1.xaxis_date()

    min_x = min(xs)
    max_x = max(xs)+datetime.timedelta(0,2*60*60)

    rects = ax1.bar(xs,ys,color=(1,0,0))
    # XXX note that all the rectangle dimensions look sane
    print xs
    print ys
    for r in rects:
        print r.get_xy(),r.get_width(),"x",r.get_height()

    ax1.set_xbound(min_x,max_x)

    fig.savefig("foo1.png")
    plt.close(fig)


# Works
def testBarChartsDTMonths():
    xs = [ \
                        datetime.datetime(2009,6,1),
                        datetime.datetime(2009,7,1),
                        datetime.datetime(2009,8,1),
                        datetime.datetime(2009,9,1) ]
    ys = [100,300,200,600]

    fig = plt.figure(figsize=(20,12)) # dims in inches
    ax1 = fig.add_subplot(111)

    ax1.set_ylabel("foo")

    ax1.xaxis_date()

    min_x = min(xs)
    max_x = max(xs)+datetime.timedelta(1)

    rects = ax1.bar(xs,ys,color=(1,0,0))
    # debugging
    print xs
    print ys
    for r in rects:
        print r.get_xy(),r.get_width(),"x",r.get_height()

    ax1.set_xbound(min_x,max_x)

    fig.savefig("foo2.png")
    plt.close(fig)



if __name__=='__main__':
    testBarCharts()
    testBarChartsDTMonths()
    testBarChartsDTHours()