Getting Started - Framework

CCruncher is designed to work in batch mode, without graphical support. The user creates a xml file with the description of the portfolio. CCruncher takes this file and simulates the portfolio N times. The simulated values are stored in a csv file. Finally, a R script takes the simulated values and calculates some statistics on them to generate the graphs and the risk indicators, such as the Expected Loss, Standard deviation, Value at Risk and Expected Shortfall.

Getting Started - Installation

Step 1. If you have selected a packaged distribution, unpack it.

  tar -xvzf ccruncher-X.Y_ZZZ.tgz
  cd ccruncher-X.Y
    

Step 2. CCruncher uses a library named gsl not included in CCruncher distribution packages. Check that you have gsl installed (find file libgsl.so , usually in folder /usr/lib). In case you don't have it, install it from the gsl site (if you use a RPM system, install the package named gsl). CCruncher source distribution requires that you have gsl-devel installed (find file gsl_math.h, usually in folder /usr/include/gsl). In case you don't have it, install it from gsl site (if you use a RPM system, install the package named gsl-devel).

Step 3. CCruncher uses a library named expat not included in CCruncher distribution packages. Check that you have expat installed (find file libexpat.so or libexpat.dll, usually in folder /usr/lib). In case you don't have it, install it from expat site (if you use a RPM system, install the package named expat). CCruncher source distribution requires that you have expat-devel installed (find file expat.h, usually in folder /usr/include). In case you don't have it, install it from expat site (if you use a RPM system, install the package named expat-devel).

Step 4. CCruncher uses a library named zlib not included in CCruncher distribution packages. Check that you have zlib installed (find file libz.so or zlib1.dll, usually in folder /usr/lib). In case you don't have it, install it from zlib site (if you use a RPM system, install the package named zlib). CCruncher source distribution require that you have zlib-devel installed (find file zlib.h, usually in folder /usr/include). In case you don't have it, install it from zlib site (if you use a RPM system, install the package named zlib-devel).

Step 5. CCruncher uses a statistical package named R to compute risk indicators. Install it from the R project site. If you plan to realize the statistical computations with your own tools (eg. Excel) you can skip this step. If you use a RPM system, install package R.

Step 6. If you have a source distribution, run the following commands from within the CCruncher directory:

  ./configure       # prepare compilation
  make              # compile ccruncher
  bin/src2bin.sh    # install and clean
    

Getting Started - Usage

Step 1. Create a XML ccruncher input file. First, read the Technical Document and the Input File Format document. You can use a file from directory samples as a template. Take into consideration that xml input can be gzipped (gz extension) to reduce the file size. Caution, gzip is a compression algorithm different from zip.

Step 2. Run ccruncher

  bin/ccruncher --help
  bin/ccruncher -f --hash=100 --path=data/ samples/test03.xml
    

Step 3. Check the output files, which are the csv files. Each column has the simulated segment values. The first line is a header that contains the segments names.

  ls -l data/*
  more data/portfolio.csv
    

Step 4. Use your prefered data analysis tool to compute the risk indicators.

Ccruncher provides a basic R script located in $CCRUNCHER/bin/ccreport.R that implements the risk indicators described in documentation. To run it, open a R console and type:

  source("bin/ccreport.R")
  ccruncher.summary("data/portfolio.csv")
  ccruncher.graphic("data/portfolio.csv")
  quit(save='no')
    

Linux/UNIX version has an additional script to create a report:

  bin/ccreport.sh data/portfolio.csv
  firefox data/portfolio.html