R/Qt, an overview


Last year, I attended the DSC 2009 conference in Copenhagen (a very nice place!). Sarkar Deepayan had a talk about a newly designed graphical interface based on Qt. To illustrate the power of this glyph-based device, he showed how a scatterplot matrix of 100 random gaussian vectors renders very slowly through lattice compared to QT-based plot.

Although not ready for production, the qtpaint package offers some visualization functionalities that I’d like to experiment. Obviously, the installation doesn’t get right out of the box…

Here is a couple of tricks I used to compile a 64-bits version of qtpaint.

  • Install Qt, not from binary package which is 32-bits only, but from source: wget http://get.qt.nokia.com/qt/source/qt-everywhere-opensource-src-4.6.2.tar.gz (153 Mb), then ./configure -arch x86_64; make (I think a flag like -cocoa might also works) and sudo make install: the compilation takes about 3 hours (on a 2.8 GHz Core 2 Duo), so be patient :)
  • Download the SVN qtinterfaces bundle package from Rforge, e.g. svn checkout svn://svn.r-forge.r-project.org/svnroot/qtinterfaces
  • CMake ≥ 2.8.1, grab the source from here, then ./bootstrap; make and sudo make install upon completion (use gcc and g++ to compile, not the llvm suite, e.g. export CXX=g++; export CC=gcc)
  • Build the R package qtbase, using R CMD build qtbase in the svn directory (go to pkg/ from svn root directory)
  • Install the R package with R CMD install qtbase_0.6-8.tar.gz; this is also rather long and you will see several passes over the package components (smoke, etc.)

Ok, it’s a bit long when one’s just want to give R/qt a try.


Articles with the same tag(s):

Multi-Group comparison in Partial Least Squares Path Models
Yet another gray theme for R base graphics
Writing a book
R, pipes and Co.
R Graphs Cookbook
Emacs Org-mode and literate programming
Reproducible research with R
Python for interactive scientific data visualization
Bar charts of counts or frequencies in Stata