31
votes
$\begingroup$

Besides gnuplot and ggobi, what open source tools are people using for visualizing multi-dimensional data?

Gnuplot is more or less a basic plotting package.

Ggobi can do a number of nifty things, such as:

  • animate data along a dimension or among discrete collections
  • animate linear combinations varying the coefficients
  • compute principal components and other transformations
  • visualize and rotate 3 dimensional data clusters
  • use colors to represent a different dimension

What other useful approaches are based in open source and thus freely reusable or customizable?

Please provide a brief description of the package's abilities in the answer.

$\endgroup$
1
  • 4
    $\begingroup$ I wonder if it isn't more sensible to ask for methods of visualisation, rather than packages, especially since most of the answers are providing little detail, and many packages provide the same methods. See, for example, stats.stackexchange.com/questions/41326/… $\endgroup$
    – naught101
    Commented Nov 8, 2012 at 22:50

8 Answers 8

Reset to default
13
votes
$\begingroup$

How about R with ggplot2?

Other tools that I really like:

$\endgroup$
1
  • $\begingroup$ ggplot2 is just a graphic package? What about it makes you recommend it for multi-dimensional data? Facetting? $\endgroup$
    – naught101
    Commented Nov 21, 2012 at 6:26
15
votes
$\begingroup$
  • Mondrian: Exploratory data analysis with focus on large data and databases.
  • iPlots: a package for the R statistical environment which provides high interaction statistical graphics, written in Java.
$\endgroup$
2
  • $\begingroup$ +1 for Mondrian - very useful toy, especially for large data $\endgroup$
    – user22
    Commented Jul 20, 2010 at 14:54
  • $\begingroup$ large data != high dimensionality. Is Mondrian any more useful than other packages for high dimensionality? $\endgroup$
    – naught101
    Commented Nov 9, 2012 at 2:39
11
votes
$\begingroup$

The lattice package in R.

Lattice is a powerful and elegant high-level data visualization system, with an emphasis on multivariate data,that is sufficient for typical graphics needs, and is also flexible enough to handle most nonstandard requirements.

Quick-R has a quick introduction.

$\endgroup$
2
  • $\begingroup$ Heh. I can't edit this answer to add this link because it's too short. With 4 upvotes, there must be at least a few people out there familiar enough with lattice that they could add a couple of lines of description, to actually make this answer half-way useful... $\endgroup$
    – naught101
    Commented Nov 21, 2012 at 6:35
  • 1
    $\begingroup$ good point. I've added a blurb and your quick-r link $\endgroup$
    – Jeromy Anglim
    Commented Nov 21, 2012 at 6:38
4
votes
$\begingroup$

ggobi and the R links to Ggobi are really rather good for this. There are simpler visualisations (iPlots is very nice, also interactive, as mentioned).

But it depends whether you are doing something more specialised. For example TreeView lets you visualise the kind of cluster dendrograms you get out of microarrays.

$\endgroup$
3
votes
$\begingroup$

Python's matplotlib

$\endgroup$
3
votes
$\begingroup$

Viewpoints is useful for multi-variate data sets.

$\endgroup$
1
  • $\begingroup$ I can only second that... from what I've seen can you select data with the mouse in one projection while looking how the selected subset looks like in another projection. $\endgroup$
    – Andre Holzner
    Commented Aug 16, 2010 at 6:22
1
vote
$\begingroup$

t-SNE has many open source implementations. One of the easiest to use is probably sklearn.manifold.TSNE. sklearn.manifold contains other manifold learning methods to plot your data to 2D:

enter image description here

$\endgroup$
0
votes
$\begingroup$

Look also SCaVis data plotting library. It works on any platform since Java. It supports many data containers and plot styles (2D, 3D etc.)

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.