Approaches to Visual Mappings

Transcription

1 Approaches to Visual Mappings CMPT 467/767 Visualization Torsten Möller Weiskopf/Machiraju/Möller

2 Overview Effectiveness of mappings Mapping to positional quantities Mapping to shape Mapping to color Mapping to texture Other mappings Glyphs Weiskopf/Machiraju/Möller 2

3 Readings Colin Ware: Chapter 4 (Color) Chapter 5 (Visual Attention and Information that Pops Out) The Visualization Handbook: Chapter 1 (Overview of Visualization) Additional (background) reading J. Mackinlay: Automating the design of graphical presentations of relational information. ACM Weiskopf/Machiraju/Möller ToG, 5(2), ,

4 Visual Language is a Sign System Image perceived as a set of signs Sender encodes information in signs Receiver decodes information from signs Jacques Bertin French cartographer [ ] Semiology of Graphics [1967] Theoretical principles for visual encodings Weiskopf/Machiraju/Möller 4 Semiology of Graphics [J. Bertin, 83]

5 Effectiveness of Mappings Mapping from (filtered) data to renderable representation Most important part of visualization Possible visual representations: Position Size Orientation Shape Brightness Color (hue, saturation). Weiskopf/Machiraju/Möller 5

6 Effectiveness of Mappings Efficiency and effectiveness depends on input data: Nominal Ordinal Quantitative Good visual design Based on psychology and psychophysics Psychological investigations to evaluate the appropriateness of mapping approaches Weiskopf/Machiraju/Möller 6

7 Basic Types of Visual Encodings Marks Points Lines Areas Channels Position Size (Grey)Value Texture Color Orientation Shape Weiskopf/Machiraju/Möller 7 Semiology of Graphics [J. Bertin, 67]

8 Mapping to Data Types Position Size (Grey)Value Texture Color Orientation Shape Nominal Ordinal Quantitative Weiskopf/Machiraju/Möller 8 = Good ~ = OK = Bad

9 Mapping to Data Types Nominal Ordinal Quantitative Position Size (Grey)Value ~ Texture ~ Color Orientation Shape Weiskopf/Machiraju/Möller 9 = Good ~ = OK = Bad

10 Mackinlay s Retinal Variables Weiskopf/Machiraju/Möller [Mackinlay, Automating the Design of Graphical 10 Presentations of Relational Information, ACM TOG 5:2, 1986]

11 Effectiveness of Mappings Effectiveness of visual variables According to Mackinlay [J. Mackinlay: Automating the design of graphical presentations of relational information. ACM Transactions on Graphics, 5(2), , 1986] Weiskopf/Machiraju/Möller 11

12 Stolte / Hanrahan Weiskopf/Machiraju/Möller Polaris: A System for Query, Analysis and Visualization of Multi-dimensional Relational Databases Chris Stolte and Pat Hanrahan 12

13 Effectiveness of Mappings Effectiveness according to neurophysiology Cells in Visual Areas 1 and 2 differentially tuned to each of the following properties: Orientation and size (with luminance) Color (two types of signal) Stereoscopic depth Motion Grapheme Describes a graphical element that is primitive in visual terms Makes use of early stages of human vision Weiskopf/Machiraju/Möller 13

14 Mapping to Positional Quantities Mapping to positional quantities: Position Size Orientation Geometric mapping Typically, very effective visual parameters Generic diagram methods Weiskopf/Machiraju/Möller 14

15 Mapping to Positional Point diagrams Quantities E.g. scatter plots: visual recognition of correlations Weiskopf/Machiraju/Möller 15

16 Mapping to Positional Quantities Line and curve diagrams: Effective perception of differences in Position and Length Weiskopf/Machiraju/Möller

17 Bar graph: Mapping to Positional Quantities Discrete independent variable (domain): nominal/ordinal/quantitative Quantitative dependent variable (data) Weiskopf/Machiraju/Möller

18 Pie charts: Mapping to Positional Quantities Quantitative data that adds up to a (fixed?) number Weiskopf/Machiraju/Möller 18

19 Mapping to Shape Especially useful for data with direct relationship to locations Convey spatial structures Examples Isolines (contour lines) Height fields Function graphs (function plots) Direct encoding of qualitative data Typically in the form of glyphs Weiskopf/Machiraju/Möller 19

20 Mapping to Color Issues: What kind of data can be color-coded? What kind of information can be efficiently visualized? Areas of application Provide information coding Designate or emphasize a specific target in a crowded display Provide a sense of realism or virtual realism Provide warning signals or signify low probability events Group, categorize, and chunk information Convey emotional content Provide an aesthetically pleasing display Weiskopf/Machiraju/Möller 20

21 Mapping to Color Possible problems: Distract the user when inadequately used Dependent on viewing and stimulus conditions Ineffective for color deficient individuals (use redundancy) Results in information overload Unintentionally conflict with cultural conventions Cause unintended visual effects and discomfort Weiskopf/Machiraju/Möller 21

22 Mapping to Color Nominal data Colors need to be distinguished Localization of data Around 8 different basis colors co-citation analysis Weiskopf/Machiraju/Möller 22

23 Mapping to Color Nominal data Weiskopf/Machiraju/Möller 23

24 Mapping to Color Ordinal and quantitative data Ordering of data should be represented by ordering of colors Psychological aspects x 1 < x 2 <... < x n E(c 1 ) < E(c 2 ) <... < E(c n ) Color coding for scalar data Assign to each scalar value a different color value Assignment via transfer function T T : scalarvalue colorvalue Code color values into a color lookup table Scalar (0,1) 0 R i G i B i Weiskopf/Machiraju/Möller (0,1) (0,N-1)

25 Mapping to Color Pre-shading vs. post-shading Pre-shading Assign color values to original function values (e.g. at vertices of a cell) Interpolate between color values (within a cell) Post-shading Interpolate between scalar values (within a cell) Assign color values to interpolated scalar values Linear transfer function for color coding Specify color for fmin and for fmax (Rmin,Gmin, Bmin) and (Rmax, Gmax, Bmax) Linearly interpolate between them Weiskopf/Machiraju/Möller 25

26 Mapping to Color Different color spaces lead to different interpolation functions In order to visualize (enhance/suppress) specific details, non-linear color lookup tables are needed Gray scale color table Intuitive ordering Rainbow color table Less intuitive HSV color model Perceptual issues Temperature color table Weiskopf/Machiraju/Möller 26

27 Mapping to Color Bivariate and trivariate color tables are problematic: No intuitive ordering Colors hard to distinguish Many more color tables for specific applications Design of good color tables depends on psychological / perceptional issues Often interactive specification of transfer functions to extract important features Weiskopf/Machiraju/Möller 27

28 Mapping to Texture Main parameters for texture Orientation Size Contrast Alternatively [Tamura 78]: Coarseness Roughness Contrast Directionality Line-likeness Regularity Weiskopf/Machiraju/Möller 28 [C. Ware, Information Visualization]

29 Mapping to Texture Goal: Avoid visual "crosstalk Orthogonal perceptual channels Restricts range of parameters E.g. approximately 30 degrees difference in orientation needed to distinguish textures Main application for textures: nominal data Some applications for direct visualization of orientations Weiskopf/Machiraju/Möller 29

30 Mapping to Texture Generate texture Gabor func. as primitives Parameters: Orientation Size Contrast Randomly splatter down Gabor functions Blending yields continuous coverage Stochastic texture model Visualization of a magnetic field [C. Ware, Information Visualization] Weiskopf/Machiraju/Möller 30

31 Mapping to Texture Other stochastic texture models: LIC (Line integral convolution) for vector field visualization Structural models Procedural description of texture generation E.g. Lindenmayer systems (L-systems) Weiskopf/Machiraju/Möller 31

32 Other Mappings More advanced mappings possible Examples for other visual variables Motion Blink coding Explicit use of 3D Multiple attributes Typical combination of attributes: Geometric position, e.g., height field Color: saturation, intensity, tone Texture Issue: Interference? Weiskopf/Machiraju/Möller 32

33 Glyphs Glyphs and icons Consist of several components Features should be easy to distinguish and combine Icons should be separated from each other Mainly used for multivariate discrete data Weiskopf/Machiraju/Möller 33

34 Glyphs More integral coding pairs Multi-dimensional coding Perceptual independence? Integral display dimensions Two or more attributes perceived holistically Separable dimensions Separate judgments about each graphical dimension Simplistic classification, with a large number of exceptions and asymmetries [C. Ware, Information Visualization] Weiskopf/Machiraju/Möller More separable 34 coding pairs

35 Glyphs Interesting graphical attributes for basic glyph design [according to C. Ware, Information Visualization] Visual variable Spatial position of glyph Color of glyph Shape Orientation Surface texture Motion coding Blink coding: The glyph blinks on and off at some rate Dimensionality 3 dimensions: X, Y, Z 3 dimensions: defined by color opponent theory 2 3? dimensions unknown 3 dimensions: corresponding to orientation about each of the primary axes 3 dimensions: orientation, size, and contrast 2 3? Dimensions largely unknown, but phase may be useful 1 dimension Weiskopf/Machiraju/Möller 35

36 Glyphs Color icons [Levkowitz 91] Subdivision of a basic figure (triangle, square, ) into edges and faces Mapping of data to faces via color tables Grouping by emphasizing edges or faces Weiskopf/Machiraju/Möller 36

37 Glyphs Stick-figure icon [Picket & Grinstein 88] 2D figure with 4 limbs Coding of data via Length Thickness Angle with vertical axis 12 attributes Exploits the human capability to recognize patterns/textures Weiskopf/Machiraju/Möller 37

38 Glyphs Stick-figure icon Weiskopf/Machiraju/Möller 38

39 Glyphs Circular icon plots: Star plots Sun ray plots etc Follow a "spoked wheel" format Values of variables are represented by distances between the center ("hub") of the icon and its edges Weiskopf/Machiraju/Möller 39

40 Glyphs Star glyphs [S. E. Fienberg: Graphical methods in statistics. The American Statistician, 33: , 1979] A star is composed of equally spaced radii, stemming from the center The length of the spike is proportional to the value of the respective attribute The first spike/attribute is to the right Subsequent spikes are counter-clockwise The ends of the rays are connected by a line Weiskopf/Machiraju/Möller 40

41 Glyphs Sun ray plots Similar to star glyphs/plots Underlying star-shaped structure Weiskopf/Machiraju/Möller 41

42 Glyphs Chernoff faces/icons [H. Chernoff (1973). The use of faces to represent points in k-dimensional space graphically. Journal of the American Statistical Association, 68 : ] Each facial feature represents one variable Human ability to distinguish small features in faces Weiskopf/Machiraju/Möller 42

43 Glyphs Chernoff faces/icons Possible assignment in the decreasing order of importance: Area of the face Shape of the face Length of the nose Location of the mouth Curve of the smile Width of the mouth Location, separation, angle, shape, and width of the eyes Location of the pupil Location, angle, and width of the eyebrows Coding of 15 attributes Additional variables could be encoded by making faces asymmetric Weiskopf/Machiraju/Möller 43

44 Glyphs Chernoff faces/icons Weiskopf/Machiraju/Möller 44

45 Glyphs Face morphing [Alexa 98] Weiskopf/Machiraju/Möller 45