Modules > Interaction and Distortion Techniques
Description | Usage
Hints | Learning Task | Discussion
| References | Acknowledgments
Most information visualizations are highly
interactive and support the location and interpretation of items as well
as the exploration of their relationships. Commonly used interaction techniques
All these techniques can be used to explore potentially huge information
spaces. However, the amount of information people need to process is increasing
much faster (doubling every 18 montgs) than the resolution of our computer
displays. Therefore, a number of distortion techniques have been developed
that magnify the area of interest while minimizing all other data, thus
preserving global context.
Focus & Context Techniques
Often, focus and context information is needed to examine a visualization
in order to understand and manage the underlying data set. A number of
so-called focus and context techniques exist. Examples are
- The perspective wall consists of three different
magnification areas. The front wall is used to display the magnified
part of the data in a constant magnification rectangle. The two side
walls show contextual information with decreasing magnification level from
the front wall (Mackinlay et al., 1991).
- The fisheye lens produces a distorted view
similar to a very wide-angle camera lens (Furnas,
- Hyperbolic space applies hyperbolic geometries
to map a potentially infinite space to a (finite) visualization space
- Traditional (linear) magnification techniques
and nonlinear magnification magnifies the focus region without occluding
the context (Leung and Apperley, 1994). Occlusion-free
magnification can be achieved by combining linear magnifications and
constrained transformation domains (Keahey and Robertson,
Each distortion technique comes with a Transformation function
that defines the way in which a point in the original object image is
transformed to the distorted target image and a Magnification
function which describes the degree of distortion which
has been applied to a particular point of interest. Mathematically, the
magnification function is the first-order derivative of the transformation
function. Please consult Leung and Apperley (1994)
for a great review.
Subsequently, we will examine the fisheye
lens and the hyperbolic
space. Please read the respective web pages before you continue.
Read How to
Install, Compile, and Run the XML Toolkit and start the toolkit. Information
on how to run the hyperbolic tree is available here.
Information on how to run the fisheye table is online here.
Hyperbolic Visualization of Tree Data
We will use the hyperbolic tree viewer to visualize, navigate
and make sense of a tree data set that is too large to fit on the screen.
In particular, we are going to visualize the directory
structure of the InfoVis Toolkit. Run
the toolkit and then load this data set via 'File > Open' and visualize
it via 'Visualization > Hyperbolic Tree'. Navigate the tree to become
familiar with the general software architecture. The images below show
a sequence of snapshots taken while navigating into the '/edu/iu/iv/layout'
click smaller images to enlarge them.
Fisheye Distortion of Tabular Data
Next, we will apply the fisheye table to visualize a long scrolling table
on a screen. In particular, you are asked to visualize a table
that provides information on those movies that have grossed over $100,000,000
at the box office during their theatrical runs. All amounts are in USA
dollars and only include theatrical box office receipts (movie ticket
sales) and do not include video rentals, television rights and other revenues.
Totals may include theatrical re-release receipts. Figures are not adjusted
for inflation. Some movies may still be in general release; all figures
are estimated and subject to change. This information and the data set
itself were downloaded from the Internet
To display the data file you must convert it into the tabular format
and load the data file via 'File > Open'. Next click 'Visualization
> Fisheye Table' and a new window will open showing the table in a
distorted fashion. Mouse over the areas that you would to see in detail.
Double click into any table field to edit its value.
It is now your task to visualize a data set of your choice using either
one of the two algorithms. Please keep in mind that the hyperbolic tree
viewer requires a tree data set while the fisheye table takes a table
data set as input.
The discussed techniques provide diverse ways to visualize and examine
very large scale data sets. Promising directions of future research in
this area comprise the investigation of perceptual cues (3-D grids, color
and shading, stereo display) to reveal nature of distortions when applied
to more general data sets, the extension of distortion techniques to three
dimensions, separating magnification and displacement functions, visual
access distortion, and data driven magnification.
- George W. Furnas (1986) Generalized
Fisheye Views. In Proceedings of the Conference on Human Factors in
Computing Systems (CHI'86), pages 16-23. ACM Press.
- T. Alan Keahey and Edward L. Robertson
(1996). Techniques for non-linear magnification transformations. In Proceedings
of the IEEE Symposium on Information Visualization (InfoVis 1996).
IEEE Computer Society Press.
- Y. K. Leung and M. D. Apperley (1994).
A review and taxonomy of distortion-oriented presentation techniques.
ACM Transactions on Computer-Human Interaction (TOCHI), 1(2):126-160.
- Jock D. Mackinlay, George
G. Robertson, and Stuart K. Card (1991) The Perspective Wall: Detail
and Context Smoothly Integrated. In Proceedings of the Conference on
Human Factors in Computing Systems (CHI'95), Information Visualization,
pages 173-179. ACM Press.
- Tamara Munzner (1997) H3: Laying
Out Large Directed Graphs in 3D Hyperbolic Space. In L. Lavagno and
W. Reisig, editors, Proc. IEEE Symp. Information Visualization, pp
- For additional information and examples consult Focus-plus-Context
Techniques by Niki Sahling.
This documentation was compiled by Katy Börner and Bruce William
Herr. We would like to thank Renee LeBeau and Nathan James Deckard for
their help in preparing the sample data sets and parsers.
Information Visualization Cyberinfrastructure
@ SLIS, Indiana University
Last Modified May 10, 2004