| Author |
 Topic  |
|
|
cdrozdowski111
USA
247 Posts |
 Posted - 11/21/2013 : 07:50:55 AM
|
OriginPro 9.1 SR0 64-bit/32-bit, Win7 64-bit, running in VMware Fusion 5.0.3
I am working with data and a polar plot. I have a lot of data clustered very near the center of the plot with other points rather farther out on the plot (see a very basic example below). Can anyone suggest an elegant way to deal with the clustered data such that the points in the center are more clearly spaced?
I'm wondering if there is a way to plot this in 3-D while still maintaining the polar nature of the relationship. For example, to "draw up" the center into the 3rd dimension to stretch it out.
 |
Edited by - cdrozdowski111 on 11/21/2013 07:52:08 AM |
|
|
Castiel
343 Posts |
Posted - 11/21/2013 : 08:45:23 AM
|
What about this?
c ȥ fuc c Τ c/
cʥ ' ̩ / թf
`
|
 |
|
|
cdrozdowski111
USA
247 Posts |
Posted - 11/21/2013 : 09:27:44 AM
|
Thanks for the suggestion.
But what I need to be able to do is to better elucidate the values along the radial axes- especially those that are very close together. It needs to be obvious the difference between 0.1 and 0.2. |
|
|
|
cdrozdowski111
USA
247 Posts |
Posted - 11/21/2013 : 09:32:16 AM
|
| BTW- it doesn't have to be an actual polar plot. But it needs to maintain the polar relationship visually. |
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 11/21/2013 : 11:35:06 AM
|
Hi,
> It needs to be obvious the difference between 0.1 and 0.2.
You can try log10 scale in radius axis like below.
How about this?
--Hideo Fujii
OriginLab
P.S. You need to turn OFF the "Center at Zero" check box at the "Radial Axes: Scale" branch in the Axis dialog box (as there is no 0 in the log scale.) |
Edited by - Hideo Fujii on 11/21/2013 12:41:41 PM |
|
|
|
cdrozdowski111
USA
247 Posts |
Posted - 11/21/2013 : 2:27:55 PM
|
Actually my real data is all of the same magnitude. I didn't post it because it is a bit hard to visually parse. The values I listed above were made up simply to show how the points were crowded at the center.
Here is a real sample of my data. It would be plotted such that the values go from highest (inside) towards the lowest (outside).
0 235.131045575
30 235.138545018
60 235.138545076
90 235.138545315
120 235.138545041
150 235.138545202
180 235.133833075
210 235.138545315
240 235.138545076
270 235.138545018
300 235.131045575
|
|
|
|
lkb0221
China
497 Posts |
Posted - 11/21/2013 : 2:53:53 PM
|
Hi,
I think what Hideo suggested (center at non-zero) should work. below is what I got using the actual data you posted:
Zheng
OriginLab |
Edited by - lkb0221 on 11/21/2013 2:54:48 PM |
|
|
|
cdrozdowski111
USA
247 Posts |
Posted - 11/21/2013 : 3:20:08 PM
|
Yes, I was able to plot that myself before I posted the original question.
The thing is that a polar plot such as the one you posted does nothing to effectively communicate the relationships between the points that are very close to each other in value.
In the case of my particular application, the important thing is it's visual relationship to the other points. By plotting it the way it is, the relationship is unclear.
I was simply wondering if there was a way to better visually communicate the relationships.
|
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 11/21/2013 : 5:18:20 PM
|
Hi,
How about the use of axis breaks, such as below??
Unfortunately the axis break currently hasn't been supported in the polar chart ...
But, you can mimic doing so by making 2 layers.
--Hideo |
Edited by - Hideo Fujii on 11/21/2013 5:23:59 PM |
|
|
|
lkb0221
China
497 Posts |
Posted - 11/21/2013 : 5:39:51 PM
|
Hi,
How about use scatter with transparency, colormapped by X value (angular)? Like below:
Of cause using 3D can separate the overlapped points in the center, but the Z-value will be meaningless, and only the X-Y plane projection of the graph makes sense. And currently there seems no way to show polar axis in 3D plane projection.
So can you describe a little more about your requirements with details? Or can you draw a ideal simple example by hand so we can understand what type of graph you want.
Thanks,
Zheng |
|
|
|
cdrozdowski111
USA
247 Posts |
Posted - 11/22/2013 : 06:00:36 AM
|
If I want to take this off-line so that I can explain the story behind it in detail, where should I send an email and to whom should I "Attn" it?
|
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 11/22/2013 : 09:43:44 AM
|
You can send a mail to tech@originlab.com with attachments and the explanation with ATTN. Hideo .
--Hideo Fujii |
Edited by - Hideo Fujii on 11/22/2013 09:44:46 AM |
|
|
|
couturier
France
291 Posts |
Posted - 10/11/2016 : 12:00:11 PM
|
quote:
Originally posted by Castiel
What about this?
�c�� �ȥ� �f�u�c �c�� �Τ� �c/ ����
�c�ʥ� ���� �� �� ���' �̩��� /�� �թ��f
�� �� ���� �` �� ��
��
Can someone explain how did you do to have radial axis that way, shifted away from center ?
Thanks |
|
|
|
snowli
USA
1379 Posts |
Posted - 10/13/2016 : 11:15:45 AM
|
Our TS said,
"We looked into this yesterday, we could not do it. Echo said there is only a workaround. Plot the line with y value+3, then set from=3, set tick label formula to x-3.
I add a jira ORG-15452 for it. If we could allow the center value to values not zero, then it will be easy to plot such graph.
"
So not sure how Castiel plotted it.
Thanks, Snow |
|
|
|
couturier
France
291 Posts |
Posted - 10/13/2016 : 11:24:35 AM
|
quote:
Echo said there is only a workaround. Plot the line with y value+3, then set from=3, set tick label formula to x-3
Thanks Echo and Snow.
This is how I did, but I was wondering if there was something more straightforward. |
|
|
|
Castiel
343 Posts |
Posted - 10/13/2016 : 1:02:19 PM
|
quote:
Originally posted by couturier
quote:
Originally posted by Castiel
What about this?
�c�� �ȥ� �f�u�c �c�� �Τ� �c/ ����
�c�ʥ� ���� �� �� ���' �̩��� /�� �թ��f
�� �� ���� �` �� ��
��
Can someone explain how did you do to have radial axis that way, shifted away from center ?
Thanks
IIRC, simply clear Center at Zero (the exact term I cannot recall though); scaling from a negative value, say -2; first major tick at 0 and major tick labels from dataset like [0, 1, 2, 3, 4, 5, 6].
I'm working on Mac now and cannot confirm if that works.
c ȥ fuc c Τ c/
cʥ ' ̩ / թf
`
|
|
|
|
couturier
France
291 Posts |
Posted - 10/13/2016 : 2:38:42 PM
|
Thanks Castiel for your answer.
This is how I tried at first but couldn't make it work, i.e. there's still a black line joining the upper and lower scale.
quote:
I'm working on Mac now and cannot confirm if that works.
Do you mean you no longer use Origin ?
Too bad. I will certainly miss your wise advices. |
|
|
| |
Topic |
|
Help with Polar Plot