Info
Alle Inhalte des Nutzerportal sind nur auf Englisch verfügbar.

SimVis

SimVis is a software which strongly supports exploratory visualization. This means that the software provides tools and techniques that allow a research on the data and the individual variables themselves. One of these techniques is the so called Linking & Brushing, using which the user can specifically select or de-select portions of the data, and use this selection to map other variables onto these data points. Using techniques likes scatterplots, histograms or timelines, this data-dependent selection process becomes easy and intuitive.

An example of this technique would be to use a scatterplot to select only those data points that have a high temperature AND a low humidity. In the scatterplot, one would draw on one axis the variable temperature and on the other humidity, and select those parts that are high respective low. See also the following figure, which shows a screenshot of SimVis with REMO data:

 

Simvis_Remo

 

Focus & Context

The following image shows a vector visualization of the zonal velocity in m/s of the ocean model MPI-OM. Here only those data points are drawn in color that are above a certain velocity threshold. The data selection was made using a histogram on the variable velocity magnitude. The interesting data points (above threshold) are highlighted and drawn in color, while all other data points provide contextual information, and are drawn in gray. 

 

simvis_mpiom

 

SimVis on Mistral

SimVis is installed on all GPU nodes. We have developed a brief SimVis Tutorial on how to use the software, with a special focus on the visualization requirements in earth and climate research:

SimVis Tutorial
Contents:

  •     Converting data to be used in SimVis
  •     Basic scalar and vector visualization techiques
  •     Data analysis and specific data selection/deselection techniques
  •     Deriving new data variables
  •     Deriving quantitative information
  •     Capturing screen shots and creating animations
  •     (Working with ICON data - if interested)

 

 

 

 

 

Artikelaktionen