Multidimensional Visualization David Yucht ICS 280 Spring 2001

Multidimensional Visualization David Yucht ICS 280 Spring 2001

The Problem at Hand l How do we work with the constraints of our reality (3 Spatial Dimensions + Time/Motion) in order to visualize data sets containing potentially infinite numbers of attributes (n. Space)?

Potential Solutions l l Question is moot for < 3 attributes For > 3 attributes – – Use complex 2 D displays Use multiple 3 D displays Add other distinguishing image properties Use another medium such as sound

Why use 3 D? l l l Utilizes all Euclidian spatial dimensions Humans have an innate ability to process 3 D images quickly without much effort Images can be enriched with other properties – – Time/Motion (4 D) Color Texture Shape

Problems with 3 D l l l Computationally expensive True 3 D Output & Interaction requires special equipment Navigation can be difficult Data with > 3 attributes require multiple 3 D images and/or additional properties Depth & Occlusion

Early Test Systems l l W. Wright: Information Animation Applications in the Capital Markets Feiner & Beshers: Worlds Within Worlds: Metaphors for Exploring n-Dimensional Virtual Worlds

Information Animation l l l 4 D System for visualizing financial markets Displays Equity & Bond markets in real-time Scalable from 1 stock to entire markets Allows for querying of in-depth information upon request Utilizes movement/animation to alert users to events and give a “feel” for the market

Information Animation (2) l Animation takes advantage of human visual processing – 4 systems within visual cortex l l 1 handles motion 1 handles color 2 handle object form The system thus uses motion, colors, shapes, and sizes to represent information

Financial Properties - Equity l l l Price Net changes (days, weeks, quarters, etc. ) Trading Volume at a Price Trades at bid Trades at offer

Financial Properties – Bond l l l Price Net changes Trading Volume Interest Rates Yield curve (instantaneous & historical)

Significance of Info Animation l l l 3 D can hold 10 x or more data than 2 D 3 D allows users to more quickly see market trends 4 D allows users to examine spreads as they occur Eases interaction & dimensional transformation problems by allowing queries Avoids occlusion problem by relying on display design and image arrangement

n-Vision l l Testbed system for representing n-Dimensions Uses true 3 D imaging for display & interaction – – l Stereoscopic shutter glasses VR Data. Glove Handles n-Space to Euclidian spatial transformation by nesting 3 D images within one another

n-Vision (2) l l Does utilize motion; can be considered a 4 D system Does not utilize other image properties such as texture and color to represent attributes

Nesting Worlds l l l n-Vision nests multiple 3 D images in order to perform dimensionality reduction; it slices away dimensions Each 3 D image is displayed on top of another, with reference marks to illustrate the values where that particular slice lies Uses recursive parent-child relationships to display & control multiple images

The Significance of n-Vision l l Provides a metaphor for > 4 D visualizations Sidesteps the occlusion problem with true 3 D interaction – – – Users can grab and rearrange images Stereoscopic images give accurate representation of depth Provides other tools such as the dipstick and waterline to further reduce occlusion and user confusion