Author 
Message 
Jens Müller science forum beginner
Joined: 13 Oct 2005
Posts: 1

Posted: Thu Oct 13, 2005 1:11 pm Post subject:
Visualizing 2D data with confidence intervall/standard deviation



I don't know if this is the right newsgroup, I have a problem with
visualizing data.
I have a 2D matrix with numerical values, the two indices represent two
parameters each ranging from 0% to 100%, in 5%steps.
The numerical values themselves derive from running an experiment
multiple times, and I have mean, standard deviation and confidence
intervall for them.
Currently I only plot the mean using R, with a contour plot, an image
plot (different colors depending on the value) and perspective plots
from 4 different perspectives.
Does anyone know a way to integrate the confidence interval or the
deviation, ( in such a way that you can still recognize something ...)? 

Back to top 


Duane Bozarth science forum beginner
Joined: 06 Jul 2005
Posts: 37

Posted: Thu Oct 13, 2005 3:17 pm Post subject:
Re: Visualizing 2D data with confidence intervall/standard deviation



Jens Müller wrote:
Quote: 
I don't know if this is the right newsgroup, I have a problem with
visualizing data.
I have a 2D matrix with numerical values, the two indices represent two
parameters each ranging from 0% to 100%, in 5%steps.
The numerical values themselves derive from running an experiment
multiple times, and I have mean, standard deviation and confidence
intervall for them.
Currently I only plot the mean using R, with a contour plot, an image
plot (different colors depending on the value) and perspective plots
from 4 different perspectives.
Does anyone know a way to integrate the confidence interval or the
deviation, ( in such a way that you can still recognize something ...)?

How (and how successfully) probably depends on the data themselves as
well as the shape of the surface and the magnitude and distribution of
the error estimates.
As is not knowing anything about the visualization tools you're using or
have available.
Anyway, I'd probably start by trying various forms of surface plots w/
the errors superimposed as either additional surfaces in differing
shades or as transparent or by adding the data points and point error
estimates w/ vertical lines connecting them or variations of the above.
Also, it might be possible to learn/visualize a lot by selective
sectional/conditional plotsthat again depends on the smoothness of the
data. 

Back to top 


N. Shamsundar science forum beginner
Joined: 10 Oct 2005
Posts: 40

Posted: Fri Oct 21, 2005 12:31 pm Post subject:
Re: Visualizing 2D data with confidence intervall/standard deviation



Jens Müller wrote:
Quote:  I don't know if this is the right newsgroup, I have a problem with
visualizing data.
I have a 2D matrix with numerical values, the two indices represent two
parameters each ranging from 0% to 100%, in 5%steps.
The numerical values themselves derive from running an experiment
multiple times, and I have mean, standard deviation and confidence
intervall for them.
Currently I only plot the mean using R, with a contour plot, an image
plot (different colors depending on the value) and perspective plots
from 4 different perspectives.
Does anyone know a way to integrate the confidence interval or the
deviation, ( in such a way that you can still recognize something ...)?

Draw a pair of concentric circles with mean diameter proportional to the
mean value that you wish to plot, \delta radii=the confidence interval
width, and center at the location corresponding to the indices.
I'm assuming that the confidence interval has a oneone relationship to
the standard deviation.
N. Shamsundar
University of Houston 

Back to top 


Google


Back to top 



The time now is Sun Apr 21, 2019 12:52 am  All times are GMT

