Interactive Visual Analysis with different levels of complexity

Helwig Hauser

MISC, June 24, 2010

Abstract

Interactive visual data exploration and analysis is a powerful methodology for enabling insight into complex and also large data. The iterative process of visualization and interaction (and back to visualization, aso.) can be seen as a visual dialog between the user and the data. Thereby, powerful data analysis schemes are enabled such as a step-by-step information drill-down, steered by the user’s perception, cognition, and knowledge. In this talk, we look at different levels of this methodology (in the sense of levels of complexity), starting at the first level of ``show & brush'' continuing then via ``relational analysis'' to a third level that we call ``complex analysis.'' The hypothesis is stated that it indeed is useful to have these different levels of complexity for interactive visual data analysis: a large share of all addressed problems can be satisfyingly solved with the ``simple'' level of ``show & brush,'' while the more complex levels of this methodology are only paying off in special cases. Along with a characterization of these levels, we also take a look at a number of illustrative examples.

Published

Invited talk at TU Delft

• Location: Delft, The Netherlands
• Date: June 24 2010

Media

BibTeX

@misc{hauser10levelsOfComplexity,
  author = {Helwig Hauser},
  title  ={Interactive Visual Analysis with different levels of complexity},
  year = {2010},
  month = {June 24},
  howpublished = {Invited talk at TU Delft},
  location = {Delft, The Netherlands},


  abstract = {Interactive visual data exploration and analysis is a powerful 
methodology for enabling insight into complex and also large data. The 
iterative process of visualization and interaction (and back to 
visualization, aso.) can be seen as a visual dialog between the user and 
the data. Thereby, powerful data analysis schemes are enabled such as a 
step-by-step information drill-down, steered by the user’s perception, 
cognition, and knowledge. In this talk, we look at different levels of 
this methodology (in the sense of levels of complexity), starting at the 
first level of ``show \& brush'' continuing then via ``relational analysis'' 
to a third level that we call ``complex analysis.'' The hypothesis is 
stated that it indeed is useful to have these different levels of 
complexity for interactive visual data analysis: a large share of all 
addressed problems can be satisfyingly solved with the ``simple'' level of 
``show & brush,'' while the more complex levels of this methodology are 
only paying off in special cases. Along with a characterization of these 
levels, we also take a look at a number of illustrative examples.}
}