Best Software for Visualizing Data

[vociferous]

Without consideration for cost, but with consideration for features and ease of use, what software do you think is best for producing graphical visualizations of data points, curve fitting, et cetera)?

Which ones have the easiest learning curve?

Which ones do you use, which ones have you used, and which ones do you prefer?

Mathematica, Matlab, Excel, GNUPlot, SPSS?

 

[elect_eng]

Without consideration for cost, but with consideration for features and ease of use, what software do you think is best for producing graphical visualizations of data points, curve fitting, et cetera)?

Which ones have the easiest learning curve?

Which ones do you use, which ones have you used, and which ones do you prefer?

Mathematica, Matlab, Excel, GNUPlot, SPSS?

Matlab is quite good for this, but you need to do some programming to get good results. This means it is good for detailed and long term work, but not as good for getting quick results in a hurry.

I hate using Excel for plotting and visualization.

I only use Mathematica (and similar environments) for symbolic calculations and quick plotting, but I have noticed that graphing is simpler in these environments compared to Matlab. I'm curious to hear what others say about this.

I used GNUPlot years ago and thought it was good, but can't remember enough to say if it's better or worse than others.

I never used SPSS, but years ago I used another statistical package (I forgot the name now) for a project that required it. It had amazing features for visualizing data. I suspect this might be the best choice, although it's just a guess.

 

[mgb_phys]

For lots of volume/image data IDL is probably better than those, or if you want to do some programming look at the free vtk.

 

[story645]

I never used SPSS, but years ago I used another statistical package (I forgot the name now) for a project that required it. It had amazing features for visualizing data.

Maybe SAS, which seems to be similar to SPSS but more flexible/powerful. I use SPSS. It's pretty easy to learn, but limited. It's an awesome package for the social sciences and similar, but not very good for dynamic physics systems. I use plain old python+libraries for most of my data visualization needs and it's pretty serviceable. Different tools for different things and all that.

 

[Dr Transport]

I have essentially convinced my group at work to use Matlab, or more correctly, use me and a couple of the younger scientists/engineers do all the data plotting for presentations. The ability to quickly generate theoretical predictions, overlayed with experimental data has gotten us more work than we can handle so far this year. The fact that using Matlab to script the analysis and theory has gotten upper managements attention.

 

[vociferous]

I dabbled in Matlab before (mostly doing programming, not data visualization). Right now I am struggling to get Mathematica to do what I want it to. Some of the features are pretty obvious, others are not, and the documentation related to the various plotting features seems to often be poorly documented. Also, importing data into Mathematica is more of a hassle than it should be.

Is there any software out there you guys use (other than Excel, which is more designed with business in mind than data analysis) that allows complex data visualization through a GUI, or is most scientific plotting software similar to Matlab and Mathematica, where everything has to be specified in the argument that creates the plot?

 

[story645]

Is there any software out there you guys use (other than Excel, which is more designed with business in mind than data analysis) that allows complex data visualization through a GUI, or is most scientific plotting software similar to Matlab and Mathematica, where everything has to be specified in the argument that creates the plot?

openDX and MayaVi, sort of.
Also, maybe Sage?

 

[antiņš]

I'm just a student, but for my lab reports I use http://www.r-project.org/" . R is used by statisticians in world (and my university too), I always can ask someone for help. R makes beuatiful graphics, and the best — R isn't just statistical application, it's programming language too, so you can do everything, what you want.

P.S. My English is terrible, I'm sorry.

 

[ibmichuco]

This is highly dependent on what you want to do. If
you want some simple 2d, contour, density plot
straight out of a data file with slight format modifications
such as legends, symbols, etc. then I would think
that any of the packages you mentioned should do
a decent job.

On the other hand, if you want, for example, map
the contour onto a sphere where the color scaled
from Cadet Blue (RGB[95;159;159]) to Royal Blue
(RGB[65;105;225]) then I would definitely suggest
Mathematica or Matlab.

I am only familiar with Mathematica, Gnuplot and
Matlab. Of the three, Gnuplot is the easiest one
to use, plus it is great for batch job if you are
comfortable with script file (I suspect that the other
two could too).

Another thing to look for is the support from the user's
group. Chances are that someone already did or even wrote
a package on what you want to do. My experience with
the Mathematica and Gnuplot people has been great.
Don't know much about the Matlab group, but a casual
check at the comp.soft-sys.matlab shows quite a few
unanswered queries.

Cheers,

Michuco

 

[suvorav]

OriginLab

 

[kbaumen]

Gnuplot usually. If I just need to make a nice graph without any numbers, I will probably resort to TikZ.

I've looked at some examples and Matplotlib looks well nice, however, I haven't really had time to actually try it out.