Data Analysis

Wikipedia:Data analysis - What is it actually that we call Data Analysis.

1.  HEP Data Analysis

Once there was PAW. After the almost complete failure of the LHC++ project from CERN, we are now left with its competitor, Root. On the other side, some people think that Root is BAD, or that it is the ROOT of all Evil.

  • Root
    • Please note the nice GO4 GUI for Root !
  • PAW and PAW++ are installed on all the PSI Group Linux PCs. They are also used as a on-line display for the Collect? DAQ system and its derivatives.
    It is strongly recommended to read the tutorial and to have a look at the FAQ page
  • OpenScientist is an interesting Data Analysis package, that includes an interesting/curious attempt to revive PAW on a C++ foundation. It's nice to have a KUIP command line again (much easier and faster than Root's CINT command line, IMHO - Sergio) and many underlying concepts, like AIDA - Abstract Interfaces for Data Analysis are good, but it is still far from being complete. It can read HBook, AIDA and some Root data files.
  • HippoDraw is easy, nice, fast, and terribly limited for data analysis - it qualifies more as a graphics package.
  • http://aida.freehep.org/tools.thtml A list of tools that implement AIDA.
    Please note that Geant4 natively supports AIDA.
  • IGUANA interactive visualisation toolkit (CMS, D0)
  • Anaphe once was LHC++. After wasting much time on defining specs, discussing architectures, using Rational Rose to make UML diagrams, trying to squeeze decent performance from ObjectivityDB, finally the CERN IT engineers behind it are trying to converge on AIDA. But it's not clear whether its Python-based interactive interface Lizard is ready for use.
  • http://www.cedar.ac.uk/

RPM packages for High Energy Physics things, from http://grid.tsl.uu.se/RTEs .

RPM packages of Geant4, Root and cernlib are available locally - see Scientific Linux.

2.  Generic Visualization & Data Analysis

  • Octave is a software that is almost a perfect replacement for the well know and expensive MathCAD. It is available in ScientificLinux.
  • Grace does not really qualify as a Data Analysis tool, but is nice and quick to display limited amounts of data.
  • QtiPlot wants to be a cross-platform clone of Origin. Must try it...
  • ParaView Interesting for the advanced visualization and parallel computing
  • OpenDX Advanced visualization
  • MicW Plot nice plotting program for OsX
  • bc is a calculator available from the command line in linux.
    Easy and quick, it has the important advantage over "graphical" one of keeping a history of your commands
    Use scale=n to get N digits of precision.
    See an introduction, or man bc for the details.
    See X-bc for a GUI to bc - with advanced math functions and constants that you can also use with plain bc.
    There is also dc, if you happen to like Reverse Polish Notation.

3.  Nuclear data

4.  Images

Images from scientific-use instruments can have 16-bit integer gray scale (like the fancy microscope at ESRF ID-19), or floating point, or other kind of fancy ways to preserve information in a pixel. These images cannot be manipulated with Gimp, which only supports 8-bit RGB or gray.

  • ImageJ Image analysis and processing in Java, supports 16-bit depth images.
    It seems a bit "heavy", eating up quite some RAM, but it feels quite fast, and it has a lot of useful tools (including all the colour mappings that you can imagine!).
  • Photoshop, of course.

5.  Links