Probably this change would upset John Tukey if he were alive.
I still wonder how you ignore the data since boxplot is there to represent
the data :)
Tell me how to undiff the changes then I will test your idea.
On Wed, Feb 24, 2010 at 12:03 AM, Ariel Rokem aro...@berkeley.edu wrote:
No - not
I found an inconsistency with how boxplots are rendered between version 0.99.1
and the svn head. See attached images. I have never seen a boxplot cross back
on itself like this before. Is this the expected behavior?
Thanks,
-Ben
Ben Axelrod
Robotics Engineer
(800) 641-2676 x737
...@coroware.com]
Sent: Tuesday, February 23, 2010 9:31 AM
To: matplotlib-users@lists.sourceforge.net
Subject: [Matplotlib-users] boxplot bug
I found an inconsistency with how boxplots are rendered between version
0.99.1 and the svn head. See attached images. I have never seen a boxplot
cross back
On Tue, Feb 23, 2010 at 3:02 PM, Ariel Rokem aro...@berkeley.edu wrote:
Hi - more generally, is there any way to control the location of the median
line, the vertical size of the box and the vertical location of the
whiskers?
Thanks - Ariel
Aren't those generically calculated from the
Hi -
yes - but I want something that looks like the generic boxplot, but in which
I can control where the edges of the boxes are placed what the sizes of the
whiskers are. A combination of errorbar and bar, with this appearance, if
you will.
Cheers - Ariel
On Tue, Feb 23, 2010 at 2:49 PM,
On Tue, Feb 23, 2010 at 5:04 PM, Ariel Rokem aro...@berkeley.edu wrote:
Hi -
yes - but I want something that looks like the generic boxplot, but in
which I can control where the edges of the boxes are placed what the sizes
of the whiskers are. A combination of errorbar and bar, with this
No - not the 'widths' kwarg. I want something that *looks* like the boxplot,
but for which I will have control of setting the ranges delimited by the box
and delimited by the whiskers (in the vertical dimension, not the horizontal
dimension). I resorted to hacking something from the existing code
