UNIT 49 - VISUALIZATION OF SPATIAL DATA

• Maps
• Computer-generated displays

• Visualization
• What is the image supposed to show?
• To whom?
• Ideal display

• 1. Location
• 2. Value
• 3. Hue
• 4. Size
• 5. Shape
• 6. Spacing
• 7. Orientation

• Explicit uncertainty codes
• Graphic ambiguity
• Examples

• Basic strategies

• Contours
• Hypsometric mapping
• Simulating oblique views of surface

UNIT 49 - VISUALIZATION OF SPATIAL DATA

Compiled with assistance from Matt McGranaghan, University of Hawaii

A. INTRODUCTION

Maps

• are limited to two-dimensions
  • must show 3-D data projected onto a flat surface
  • give a distorted impression of spatial distributions on the globe

• are static, cannot show change through time or animate

• have difficulty showing interactions or flows between places

• are limited by the tools used to make maps
  • pens of constant width
  • constant color or tone
  • the airbrush adds flexibility but is difficult to use, control

• have difficulty showing uncertainty in data
  • give a false impression of accuracy

• include screens, plots, printer output

• include raster and vector

• can be animated

• can show continuous gradations of color, texture, tone

• can show 3-D using stereoscopic technology and pairs of images

• the computer is a powerful tool for visualizing spatial information
  • this unit looks at some of the issues involved in combining the knowledge of cartography with the power of digital technology
  • all too often these issues are ignored when output maps and displays are created from GIS
  • although GIS display and mapping has much to learn from principles of cartographic design, it also provides entirely new possibilities

• must consider the objective of display

• process for putting (complex) images into minds

• examples:
  • the shape of a mountain - poorly conveyed by contours
  • pattern of growth of an urban area - may need animation to show changes through time effectively
  • air-flows over a patch of terrain - needs 3-D capabilities plus animation to show true pattern of directions, speeds of flow
  • movements of people in an area - needs ability to generalize individual movements into meaningful aggregate patterns

• components of visualization system:
  • database containing information
  • hardware device used to generate display
  • human visual system
  • processing of perceived image in the brain
  • correct perception depends on functioning of all of these components

• what impressions does the analyst wish to create in the mind?

• what relationship do these have with the contents of the database?
  • database contents are abstract version of geographical reality
  • system should create an impression of reality, not of the contents of the database
  • aspects of relationship between database and reality, e.g. accuracy, should be important part of display

• geography is complex
  • display is a filter removing unwanted complexity to show trends, patterns
  • display must show level of detail required by user, from general overview to detailed insights

• effective visualization may require familiarity with symbols on the part of the user
  • some people may never master skills of map-reading, i.e. using maps to visualize geography

• how much familiarity should be assumed?
  • it may generally be better to assume low familiarity
  • people can learn to work with complex displays, but may lose interest and look for alternative sources of information

• communicates intended message perfectly to all users
  • is not mis-understood

• offers complete design flexibility
  • put any symbol anywhere, at any size, etc.

• classes of symbols correspond to classes of objects
  • point

  • line
  • area

• visual differences among map symbols convey information

• where the symbol is

• determined primarily by geography

• the primary means of showing spatial relations

• the brain computes relations like "is within", "crosses" on the fly from the eye's perceived image of the map

• compare GISs
  • some compute these relationships on the fly, others store them in the database to avoid the processing required to compute them
  • compared to the brain, current GIS technology is amazingly crude

• lightness or darkness of a symbol

• very important visually - the eye tends to be led by patterns of light and dark

• usually used to represent quantitative differences

• tradition suggests darker symbols should mean "more" - however this may reverse on dark backgrounds which are common on computer displays - on dark backgrounds, lighter may mean "more"

• color

• important aesthetically

• usually represents qualitative differences - continuous grading of color is difficult and expensive to achieve on printed maps

• how large the symbol is

• conveys quantitative difference

• brain has difficulty inferring quantity accurately from the size of a symbol
  • if proportional circles are used to portray city population, doubling the radius of a circle (quadrupling its area) is perceived as indicating more than twice the population, but not four times
    • i.e. the brain infers population from some mixture of the radius and the area of the symbol

• geometric form of the symbol

• used to differentiate between object classes

• used to convey nature of the attribute, e.g. population indicated by images of people, housing by house symbols

• arrangement, density of symbols in a pattern

• used to show quantitative differences, e.g. dot density to show population

• of a pattern, to show qualitative differences

• of a linear symbol, to show quantitative (directional differences)

• symbol differences must be perceptible to be of use
  • JND - just noticeable difference - the smallest difference which can be reliably perceived between symbols, sizes, colors, shapes etc.
  • LPD - least practical difference - the smallest difference which can be produced by the cartographic process

• eye's sensitivity to various graphic codes
  • some codes "get through" better
    • e.g. use of yellow for fire trucks allows them to stand out better in the visual field
  • sensitivity varies across visual field
    • "peripheral vision" is enhanced by movement, varies among individuals

• cognitive aspects
  • indications that perception is dependent on cognition - knowledge understanding of phenomena
  • color categories/nameability - certain colors may have associations with names, concepts

• digital devices provide finite resolution

• spatial - where symbols might be and their shapes
  • display device has a set screen or paper size
  • display pixels have a set size, finite number of spatial locations
  • aliasing - line (or point) mapped onto closest pixel(s)
    • produces stepped (straight) lines

• color - what colors things might be
  • limit on number of colors available (palette) - plotter may have only 8 pen colors - screen may have millions of possible colors
  • limit on range of luminance and contrast

  • how many colors displayable at one time - 2n where n is the number of bit planes
  • what the colors are

• temporal limits
  • data retrieval from mass storage or from core memory?
  • how much data processing needed to compute display?
  • writing to the display device
    • speed limited by communication overhead & bus contention (competition from other activities)
  • these factors may preclude using some types of display image
    • animation requires fast through-put
    • complex images require fast data retrieval
  • acceptable response time
    • people don't like to sense a pause in the system
    • typical goal: maximum of two seconds for complex operations, instantaneous for all others
  • how long should something remain visible to be noticed?

• have to use SOME graphic code
  • don't want its meaning confused with something else
    • e.g. line drawn wide to represent uncertain position confused with wide highway or braided stream

• mark things which are uncertain with a color
  • e.g. red or yellow to suggest caution in using the information

• use graphic ambiguity to create cognitive/visual ambiguity
  • e.g. multiple positions for an uncertainly located item

• dot density or color could be used to show varying probability, e.g. a cloud with highest density in the center

• absence of "hard" lines or edges where they are uncertain

• show uncertain area with a red tint overlay

• show uncertain lines as multiple lines (like a braided stream)

• fuzzy line
  • vary the value or saturation of the line across its width

• blending between adjacent areas to show zones of transition
  • blend the colors

  • choose such that the blend works psychologically
    • red &LT-> purple &LT-> blue
    • blue &LT-> aqua &LT-> green
    • NOT red &LT-> yellow &LT-> orange
  • large set of possible colors are needed to show the appearance of a smooth transition
  • transition can be simulated with a small set of colors by spatially blending pixel colors ("dithering")

Basic strategies

• static maps
  • show a single slice of time
  • show several states at once by careful choice of symbols
  • indicate amount or rate of change

• dynamic maps
  • real time is compressed or scaled into changing display
  • non-moving occurrences - events added and deleted at places through time
  • moving objects - movement is animated on the screen - symbol is deleted at one location, regenerated at adjacent location

Contours

• calculated contours (calculated by contouring algorithms)
  • starting with a grid of elevations, thread contours and display the lines

• visual contours with elevation grid cells (contours are perceived but not computed explicitly)
  • given a sufficiently dense raster of elevations
  • shade pixels according to the elevation value of the central point using specified elevation ranges

• result is apparently (not analytically) a contour map

• set each pixel to a color dependent on its height

• this is easily implemented as table look-up

• range of colors is conventional - dark green for low elevations, through green, yellow, brown, then white at highest elevations

• each pixel's illumination computed from its slope relative to simulated "sun"
  • sun must be placed at top of image for correct visual perception - if sun is at bottom, eye sees surface inverted

• requires assumptions about reflectance of surface
  • lakes, ice, some building materials produce highlights

• single light source makes the surface too "stark"

• assume light source infinitely far away from surface
  • may assume viewer is also infinitely far away to avoid complex perspective calculations

• with TINs or coarse grids, edges of plane patches may be visible because of sharp change of slope
  • discontinuities can be eliminated by varying intensity of illumination continuously over facets
  • many 3-D display systems supply this capability - called Gouraud rendering

Standard texts on map design:

Cuff, D.J., and Mattson, M.T., 198. Thematic Maps: Their design and Production, Methuen, New York.

Dent, B.D., 1985. Principles of Thematic Map Design. Addison- Wesley, New York.

Tufte, E.R., 1983. The Visual Display of Quantitative Information. Graphics Press, Cheshire, CT. A fascinating discussion including many cartographic examples.

Texts on computer graphics:

Durrett, H.J. ed., 1987. Color and the Computer. Academic Press, New York.

Foley, J.D., and Van Dam, A., 1982. Fundamentals of Interactive Computer Graphics. Addison-Wesley, New York.

Myers, R.E., 1982. Microcomputer Graphics. Addison-Wesley, Reading, MA.

Design for digital maps:

Monmonier, M., 1982. Computer-Assisted Cartography: Principles and Prospects. Prentice-Hall, Englewood Cliffs NJ.

Techniques for displaying topography:

Kennie, T.J.M., and McLaren, R.A., 1988. "Modelling for digital terrain and landscape visualisation," Photogrammetric Record 12(72):711-45.

EXAM AND DISCUSSION QUESTIONS

1. Summarize the ways in which digital displays offer greater flexibility for visualizing spatial data.

2. The visual system is not the only way in which spatial information might be conveyed to a user. Discuss the prospects for using other methods of communication, either alone or in combination with visual methods. What kind of user interface would be appropriate for a GIS for visually impaired users, and what applications might such a system have?

3. Review the methods of visualization available in any GIS to which you have access. How limited are they, and how could they be improved?

4. How would you adapt the concept of an atlas to a digital system with capabilities for animation?

Back to Geography 470 Home Page

Back to GIS & Cartography Course Information Home Page