A Spatial Metaphor for Data Visualization

M. Sheelagh T. Carpendale, David J. Cowperthwaite, F. David Fracchia
School of Computing Science
Simon Fraser University
Burnaby, B.C. V5A 1S6 Canada
carpenda@cs.sfu.ca

Abstract

While multi-scale tools are becoming a common technique for exploring images that are larger than available screen space, there have been many user comments about confusion and disorientation. We directly address the question of making a multi-scale image comprehensible. Our approach integrates use of an intuitive spatial manipulation metaphor with visual cues.

Keywords:

Distortion viewing, 3D interaction

INTRODUCTION

Multi-scaled views make more effective use of screen space by combining various degrees of magnification and compression in a single image. Furnas' studies[3] support this approach showing user predilection for presenting and storing information in this manner. Further studies indicate increased user performance in path finding when relevant information is kept in one image [1].

In spite of this positive evidence there are also reported users' comments about disorientation [1, 5]. Specific comments are vague but complain about such things as sudden moves, not seeing how the image changed, expressing doubts as to whether the rescaled image is actually of the same information and even more vague comments about the images not being `right'. One possible explanation is that the disorientation occurs when the modified visual image conflicts with the users' prior mental memory representation of the image. Some types of visual information such as maps assume that distance is to scale and that scale is consistent. However, a multi-scale view will result in regions of varying scale. Previous knowledge about the map may protect the user from assimilating false information but causes disorientation. In unfamiliar information spaces there is an actual chance of misforming.

Use of a spatial metaphor with low level visual cues helps make the distortions present in a multi-scale view understandable. This support will allow users to maintain an accurate mental representation, learning that the information is not impaired by the distortion.

THE SPATIAL METAPHOR

In our everyday life we deal effortlessly with multi-scale information. Not only do we understand how a three-dimensional world incorporates more than one scale, the natural action of bringing objects of interest closer in order to see them better can form the basis of an intuitive interface metaphor. This idea is incorporated in the three-dimensional pliable surface (3DPS) [2] to create a 3D multi-scale viewing enviroment for two-dimensional visual information. In 3D the 2D information can be thought of as lying on a planar surface. This surface is manipulated with a simple mathematical function (gaussian curve) allowing for arbitrarily shaped multiple foci and control of the extent and pattern of magnification and compression. This in itself does not make a convincing 3D space. Visual information about the form must be provided that will reveal the nature of the distortion even through areas of sparse data.

Ideally these visual cues should not interfere with interpreting the actual information, or create a significant drain on cognitive processing. Presently the user has the option of appling either shading or a grid. %This indicates the use of a low level visual skills. Considering the extensive literature establishing that humans can discern three-dimensional shape from shading alone [4], and the considerable evidence to support the fact that this is a low level precognitive skill, shading should give an intuitive impression of the 3D shape. Such a low level visual routine will interfere less with conscious processing and may even provide an aspect of the interface that requires no learning [6]. The second choice of a regular grid reveals the shape of the distortion by accessing two depth cues: it provides perspective information without requiring edges and serves as a texture gradient. Furthermore, the shape and size of the grid squares give an approximate quantitative reading about relative magnification and distortion. As this use of a grid parallels use of a reference grid in cartography its interpretation will be familiar to many.

Figure1

Figure 1. Undistorted map of Vancouver area

Figure2

Figure 2. Vancouver area with four focal points

Figure 3. Distortion revealed by the grid

Figure4

Figure 4. Distortion revealed by shading

Figures 1, 2, 3 and 4 show the Vancouver area with four focal points. The contrast between the images with and without distortion clues demonstrates how either the shading or the grid can disambiguate the distortions.

CONCLUSION

There are a whole suite of perceptual cues through which humans establish three dimensions. At this point shading and grids are included in 3DPS. We intend to study to what extent these two cues reveal the 3-dimensional form of the distortions. We are also interested in how successfully users map different aspects of the information, for instance distance, to the distorted surface. Presently only very simple shading is included, it may be necessary to extend to shading model for it to be truly effective.

Acknowledgments

This research was suported by NSERC and the Algorithms Lab, the Graphics and Multimedia Research Lab and the School of Computing Science, S.F.U.

References

  1. L. Bartram, A. Ho, J. Dill and F. Henigman. The continuous zoom: A constrained fisheye technique for viewing and navigating large information spaces. In UIST'95: Proceedings of the ACM Symposium on User Interface Software and Technology, pages 207--216, 1995.

  2. M.S.T. Carpendale, D.J. Cowperthwaite, and F.D. Fracchia. 3-dimensional pliable surfaces: For effective presentation of visual information. In UIST: Proceedings of the ACM Symposium on User Interface Software and Technology, pages 217--226, 1995.

  3. G. W. Furnas. Generalized fisheye views. In Human Factors in Computing Systems: CHI'86 Conference Proceedings, pages 16--23, 1986.

  4. V. S. Ramachandran. Perception of shape from shading. In Nature, vol. 331(14), pages 163--166, 1988.

  5. M. Sarkar, M. H. Brown. Graphical fisheye views. In Communications of the ACM, 37(12) vol. 37(12), pages 73--84, 1994.

  6. C. Ware. The foundations of experimental semiotics: a theory of sensory and conventional representation. In Journal of Visual Languages and Computing, vol. 4, pages 91--100, 1993.


ACM:
Sponsored by the ACM's Special Interest Group on Computer-Human Interaction (ACM/SIGCHI)