Programs for visualization

From GHER

(Difference between revisions)
Jump to: navigation, search
(New page: A common question is which tools can be used to visualize the output of models and data analysis software. * matlab produce figures of good quality in combination with m_map. But matlab i...)
Line 5: Line 5:
* gnuplot is free software and available on numerous platforms (Windows, Linux, Mac). It supports many output formats, among others eps and png. However, some user find the figures of gnuplot not very pretty. To my knowlegde, it cannot read NetCDF files directly.
* gnuplot is free software and available on numerous platforms (Windows, Linux, Mac). It supports many output formats, among others eps and png. However, some user find the figures of gnuplot not very pretty. To my knowlegde, it cannot read NetCDF files directly.
-
* octave (www.octave.org): The scripting language is almost entirely compatible with matlab. In most cases, the code can run unchanged except for figures. But the
+
* octave (www.octave.org): The scripting language is almost entirely compatible with matlab. In most cases, the code can run unchanged except for figures. But the downside, is that the visualization of octave (based on gnuplot) is quite crude compared to matlab.
-
downside, is that the visualization of octave (based on gnuplot) is quite crude compared to matlab.
+
There are several add-ons to octave to fill this gap (octplot, octaviz, oplot-gl, yapso).
There are several add-ons to octave to fill this gap (octplot, octaviz, oplot-gl, yapso).
Octave can be extended using packages available at http://octave.sourceforge.net/. In
Octave can be extended using packages available at http://octave.sourceforge.net/. In

Revision as of 08:27, 30 September 2008

A common question is which tools can be used to visualize the output of models and data analysis software.

  • matlab produce figures of good quality in combination with m_map. But matlab is a proprietary software and expensive.
  • gnuplot is free software and available on numerous platforms (Windows, Linux, Mac). It supports many output formats, among others eps and png. However, some user find the figures of gnuplot not very pretty. To my knowlegde, it cannot read NetCDF files directly.
  • octave (www.octave.org): The scripting language is almost entirely compatible with matlab. In most cases, the code can run unchanged except for figures. But the downside, is that the visualization of octave (based on gnuplot) is quite crude compared to matlab.

There are several add-ons to octave to fill this gap (octplot, octaviz, oplot-gl, yapso). Octave can be extended using packages available at http://octave.sourceforge.net/. In particular you can add NetCDF support (octcdf), computation of a limited number of eigenvectors and eigenvalues with eigs and svds (arpack), optimal interpolation (optiminterp),... Installing octave from source is not trivial for Linux novices. Seeking help in your IT department can save you a lot of frustrations. I recommend you to use the version 3.0.0 or later.

  • python + matplotlib. It doesn't aim to be compatible with matlab, but its syntax is quite close.

The quality of graphics is excellent. A large number of add-ons is also available. For interactive use, ipython is very useful.

  • ferret can read NetCDF files and has good visualization capabilities. It may take some time to understand the ferret language.
  • ocean data view is a graphical program that allows to visualize various ways. It is not a programming environment like the previous ones and it thus easier to use for new users. However, it can also be used in scripts. It is free to use for research and educational purposes.
Personal tools