Visualization Analysis & Design
Tamara Munzner Department of Computer Science
University of British Columbia
Illustrations by Eamonn Maguire
~ ~~r~Fr~~i~Zr~up 'e/ Soca Raton London New York
CRC Press is an imprint of the Taylor & Francis Group, an informa business
AN A K PETERS BOOK
Contents
Preface Why a New Book? Existing Books Audience . . · .. . Who'sWho ... . Structure: What's in This Book What's Not in This Book Acknowledgments . . . .
What's Vis, and Why Do lt? 1.1 Tue Big Picture . . . . . . . . . . . . 1.2 Why Have a Human in the Loop? . 1.3 Why Have a Computer in the Loop? 1.4 Why Use an External Representation? 1.5 Why Depend on Vision? .... 1.6 Why Show the Data in Detail? .... . 1.7 Why Use Interactivity? ......... . 1.8 Why ls the Vis Idiom Design Space Huge? 1. 9 Why Focus on Tasks? . . . . . . . . 1.10 Why Focus on Effectiveness? . . . . 1.11 Why Are Most Designs Ineffective? . 1.12 Why Is Validation Difficult? . . . . . 1.13 Why Are There Resource Limitations? . 1.14 Why Analyze? . . 1.15 Further Reading . . . . . . . . . . . . .
2 What: Data Abstraction 2.1 Tue Big Picture . . . . . . . . . . . . . . . . . 2.2 Why Do Data Semantics and Types Matter? 2.3 Data Types .. 2.4 Dataset Types ....... .
2.4.1 Tables . . . . . . . . 2.4.2 Networks and Trees
2.4.2.1 Trees . . .
V
XV XV
xvi xvii
xviii xviii
XX
XX
1 1 2 4 6 6 7 9
10 11 11 12 14 14 16 18
20 21 21 23 24 25 26 27
vi Contents
2.4.3 Fields .... . ..... 27 2.4.3.1 Spatial Fields . 28 2.4 .3.2 Grtd Types 29
2.4.4 Geometry .. .. . . . 29 2.4.5 Other Combinations . 30 2.4.6 Dataset Availability 31
2.5 Attrtbute Types . . . . . . . . 31 2.5.l Categortcal . .. . . . 32 2.5.2 Ordered: Ordinal and Quantitative 32
2.5.2. l Sequential versus Diverging 33 2.5.2.2 Cyclic .... 33
2.5.3 Hierarchical Attrtbutes ... 33 2.6 Semantics ........ .. .... . 34
2.6.1 Key versus Value Semantics 34 2.6.1.1 Flat Tables .... 34 2.6.1.2 Multidimensional Tables 36 2.6.1.3 Fields . ... 37 2.6.1.4 Scalar Fields 37 2.6.1.5 Vector Fields 37 2.6.1.6 Tensor Fields 38 2.6.1.7 Field Semantics 38
2.6.2 Temporal Semantics . . . . 38 2.6.2.1 Time-Varying Data . 39
2.7 Further Reading .. ... . .. ... . 40
3 Why: Task Abstraction 42 3.1 The Big Picture .... .. .... 43 3.2 Why Analyze Tasks Abstractly? 43 3.3 Who: Designer or User 44 3.4 Actions .... . . . . . . . . 45
3.4.l Analyze . ... .. . 45 3.4.1.1 Discover . 47 3.4.1.2 Present 47 3.4 .1.3 Enjoy . . 48
3.4.2 Produce ••• • • t • 49 3.4.2.l Annotate 49 3.4.2.2 Record . 49 3.4.2.3 Dertve 50
3.4.3 Search. . . . ... 53 3.4.3.1 Lookup 53 3.4.3.2 Locate 53 3.4.3.3 Browse. 53 3.4.3.4 Explore 54
Contents
3.4.4 Query . 3.4.4.1 3.4.4.2 3.4.4.3
Identify . Compare. Summarize
3 .5 Targets ............ . 3.6 How: A Preview . . ... . . . 3. 7 Analyzing and Deriving: Examples .
3. 7 .1 Compartng Two Idioms . . . 3.7.2 Deriving One Attribute ... 3 . 7 .3 Deriving Many New Attributes
3.8 Further Reading . . . . . . . . . . . .
4 Analysis: Four Levels for Validation 4.1 Tue Big Picture . . . . 4.2 Why Validate? . . . . . . .
· 4 .3 Four Levels of Design . . 4.3. l Domain Situation 4.3.2 Task and Data Abstraction . 4.3.3 Visual Encoding and Interaction Idiom . 4.3.4 Algorithm .
4.4 Angles of Attack ... . 4.5 Threats to Validity .. . 4.6 Validation Approaches .
4.6. l Domain Validation . 4.6.2 Abstraction Validation 4.6.3 Idiom Validation ... 4.6.4 Algorithm Validation . 4.6.5 Mismatches . . . . . .
4. 7 Validation Examples . . . . . 4. 7 .1 Genealogical Graphs 4. 7 .2 MatrixExplorer . 4.7.3 Flow Maps 4.7.4 LiveRAC .... . 4.7.5 LinLog ..... . 4.7.6 Sizing the Horizon .
4.8 Further Reading . . . . . .
5 Marks and Channels 5.1 Tue Big Picture . . . . . . . . . 5.2 Why Marks and Channels? . . 5.3 Defining Marks and Channels
5.3.1 Channel Types 5.3.2 Mark Types ..... . .
„ 54 54 55 55 55 57 59 59 60 62 64
66 67 67 67 69 70 71 72 73 74 75 77 78 78 80 81 81 81 83 85 87 89 90 91
94 95 95 95 99 99
viii
5.4 Using Marks and Channels ....... . 5.4.1 Expressiveness and Effectiveness 5.4.2 Channel Rankings .
5.5 Channel Effectiveness . 5.5.1 Accuracy .... 5.5.2 Discriminability 5.5.3 Separability . 5.5.4 Popout . . . . . . 5.5.5 Grouping . . . .
5.6 Relative versus Absolute Judgements . 5. 7 Further Reading . . . . . . . . . . . . .
6 Rules of Thumb 6.1 Tue Big Picture . . . . . . . . . . . . . . . . . 6.2 Why and When to Follow Rules of Thumb? . 6.3 No Unjustified 3D . . . . . . .
6.3. l Tue Power of the Plane . . . . 6.3.2 Tue Disparity of Depth . . . . 6.3.3 Occlusion Hides Information . 6.3.4 Perspective Distortion Dangers 6.3.5 Other Depth Cues . . . . . . . . 6.3.6 Tilted Text Isn't Legibile . . . . . 6.3.7 Benefits of 3D: Shape Perception 6.3.8 Justification and Alternatives . .
Example: Cluster-Calendar Time-Series Vis Example: l.a.yer-Oriented Time-Series Vis
6.3.9 Empirical Evidence 6.4 No Unjustified 2D . . . . .. . 6.5 Eyes Beat Memory ... . .. .
6.5.1 Memory and Attention 6.5.2 Animation versus Side-by-Side Views . 6.5.3 Change Blindness ........... .
6.6 Resolution over Immersion . . . . . . . . . . . 6. 7 Overview First, Zoom and Filter, Details on Demand 6.8 Responsiveness Is Required . . . . . .
6.8.1 Visual Feedback . . . . . . . . . 6.8.2 l.atency and Interaction Design 6.8.3 Interactivity Costs . . ..
6.9 Get lt Right in Black and White 6.10 Function First, Form Next 6.11 Further Reading . . . . . . . . .
Contents
99 100 101 103 103 106 106 109 111 112 114
116 117 117 117 118 118 120 121 123 124 124 125 125 128 129 131 131 132 132 133 134 135 137 138 138 140 140 140 141
Contents
7 Arrange Tables 7 .1 The Big Pieture . . . . . . . . 7.2 Why Arrange? . . ... . .. . 7.3 Arrange by Keys and Values 7 .4 Express: Quantitative Values.
Example: Seatterplots 7 .5 Separate, Order, and Align: Categorieal Regions.
7. 5 .1 List Aligrunent: One Key . . . Example: Bar Charts . . . . . Example: Staeked Bar Charts Example: Streamgraphs . .. Example: Dot and Line Charts
7.5.2 MatrixAligrnnent: Two Keys . Example: Cluster Heatmaps . Example: Seatterplot Matrix .
7.5.3 Volumetrie Grid: Three Keys . 7 .5.4 Reeursive Subdivision: Multiple Keys .
7.6 Spatial Axis Orientation . . 7 .6.1 Reetilinear Layouts . . . . .. . 7.6.2 Parallel Layouts . .. ..... .
Example: Parallel Coordinates . 7 .6.3 Radial Layouts ....... .
Example: Radial Bar Charts Example: Pie Charts
7.7 Spatial Layout Density ...... . 7.7.1 Dense .. .. . . . . . .. . .
Example: Dense Software Overviews 7 .7.2 Spaee-Filling
7 .8 Further Reading . . . . . . . . . . . . . . . .
8 Arrange Spatial Data 8.1 The Big Pieture 8.2 Why Use Given? 8.3 Geometry . ...
8.3. l GeograpWe Data . Example: Choropleth Maps
8.3.2 Other Derived Geometry 8 .4 Sealar Fields: One Value . . . . . .
8 .4.1 Isoeontours ... .. . .. . . Example: Topographie Terrain Maps Example: Flexible lsosurfaees . . . .
8.4.2 Direet Volume Rendertng . . . . . . . Example: Multidimensional Transfer Funetions
144 145 145 145 146 146 149 149 150 151 153 155 157 158 160 161 161 162 162 162 162 166 167 168 171 172 172 174 175
178 179 179 180 180 181 182 182 183 183 185 186 187
X
8.5 Vector Fields: Multiple Values 8.5.1 Flow Glyphs ..... . 8.5.2 Geometrie Flow . . ..
Example: Similarity-Clustered Streamlines 8.5.3 Texture Flow . . . . . 8.5.4 Feature Flow . . . . . . . . . . . . .
8.6 Tensor Fields: Many Values . . . . . . ... Example: Ellipsoid Tensor Glyphs
8. 7 Further Reading . . . . . . . . . . . . . . .
9 Arrange Networks and Trees 9.1 The Big Picture . . . . . . . . . . . . . . . .. 9.2 Connection: Link Marks . . . . . . . . . . . .
Example: Force-Directed Placement Example: sfdp . . . . . . . . . . .
9.3 Matrix Views . . . . . . . . . . . . . . . . . . Example: Adjacency Matrix View . .
9.4 Costs and Benefits: Connection versus Matrix. 9.5 Containment: Hierarchy Marks
Example: Treemaps . . . Example: GrouseFlocks
9.6 Further Reading ........ .
1 O Map Color and Other Channels 10 .1 The Big Picture . . . 10.2 Color Theory . . . .
10.2.1 Color Vision 10.2.2 Color Spaces 10.2.3 Luminance, Saturation, and Hue 10.2.4 Transparency . . . . . .
10.3 Colormaps ........... . 10.3.l Categorical Colormaps 10.3.2 Ordered Colormaps .. 10.3.3 Bivariate Colormaps . . 10.3.4 Colorblind-Safe Colormap Design .
10.4 Other Channels . . . . 10.4.1 Size Channels . . . 10.4.2 Angle Channel . . . 10.4.3 Curvature Channel 10.4.4 Shape Channel . . . 10.4.5 Motion Channels .. 10.4.6 Texture and Stippling .
10.5 Further Reading ....... .
Contents
189 191 191 192 193 193 194 194 197
200 201 201 204 207 208 208 209 213 213 215 216
218 219 219 219 220 223 225 225 226 229 234 235 236 236 237 238 238 238 239 240
Contents
11 Manipulate View 11.1 The Big Picture . . . . . 11.2 Why Change? . . . . . . 11.3 Change View over Time
Example: LineUp Example: Animated Transitions .
11.4 Select Elements . . . . . . . . . . 11.4.1 Selection Design Choices . . . . . 11.4.2 Highlighting . . . . . . . . . . ..
Example: Context-Preserving Visual Links 11.4.3 Selection Outcomes ..
11.5 Navigate: Changing Viewpoint 11.5. 1 Geometrie Zooming . . 11.5.2 Semantic Zooming . . . 11.5.3 Constrained Navigation .
11.6 Navigate: Reducing Attrtbutes 11.6.1 Slice .... .. .. . .
Example: HyperSlice 11.6.2 Cut .. .. 11.6.3 Project . .
11. 7 Further Reading
12 Facet into Multiple Views 12.1 The Big Picture .. ... ..... . 12.2 Why Facet? ..... . . . .... . 12.3 Juxtapose and Coordinate Views
12.3. l Share Encoding: Same/Different Example: Exploratory Data Visualizer (EDV)
12.3.2 Share Data: All, Subset, None ... . . . .. . Example: Bird's-Eye Maps ... ...... . . Example: Multiform Overview-Detail Microarrays Example: Cerebra! . . . . . . .
12.3.3 Share Navigation: Synchronize 12.3.4 Combinatlons .. ..
Example: Improvise . 12.3.5 Juxtapose Views .. .
12.4 Partition into Views .... . 12.4.1 Regions, Glyphs, and Views 12.4.2 List Alignments .. 12.4.3 Matrtx Alignments .. .
Example: Trellis . . . . 12.4.4 Recursive Subdivision.
12.5 Supertmpose Layers . . ... .
xi
242 243 244 244 246 248 249 250 251 253 254 254 255 255 256 258 258 259 260 261 261
264 265 265 267 267 268 269 270 271 274 276 276 277 278 279 279 281 282 282 285 288
xii
12.5.1 Visually Distinguishable Layers . 12.5.2 Static Layers . . . . . . . . . ...
Example: Cartographic Layering Example: Superimposed Line Charts . Example: Hierarchical Edge Bundles .
12.5.3 Dynamic Layers 12.6 Further Reading . . . . . . . . . . . . . . .. .
Contents
289 289 289 290 292 294 295
13 Reduce ltems and Attributes 13.1 The Big Picture
298 299 299 300
13.2 Why Reduce? ... . 13.3 Filter ........ .
13.3.1 Item Filtering . 301 Example: FilmFinder 301
13.3.2 Attribute Filtering 303 Example: DOSFA 304
13.4 Aggregate . . . . . . . . . 305 13.4.1 Item Aggregation . 305
Example: Histograms . 306 Example: Continuous Scatterplots 307 Example: Boxplot Charts . . . . . . 308 Example: SolarPlot . . . . . . . . . 310 Example: Hierarchical Parallel Coordinates 311
13.4.2 Spatial Aggregation . . . . . . . . . . . . . . 313 Example: Geographically Weighted Boxplots 313
13.4.3 Attribute Aggregation: Dimensionality Reduction . 315 13.4.3.1 Why and When to Use DR? . . . . . . . . 316 Example: Dimensionality Reduction for Document Collections . 316 13.4.3.2 How to Show DR Data? 319
13.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
14 Embed: Focus+Context 14.1 The Big Picture 14.2 Why Embed? . 14.3 Elide ..... .
Example: DOITrees Revisited 14.4 Superimpose ............. .
Example: Toolglass and Magie Lenses 14.5 Distort ............... .
Example: 3D Perspective . . . . Example: Fisheye Lens . . . . . Example: Hyperbolic Geometry
322 323 323 324 325 326 326 327 327 328 329
Contents
Example: Stretch and Squish Navigation Example: Nonlinear Magnification Fields
14.6 Costs and Benefits: Distortion 14. 7 Further Reading . . . . . . . . . . . . . . . . . . .
15 Analysis Gase Studies 15.1 Tue Big Picture ........ . 15.2 Why Analyze Case Studies? .. 15.3 Graph-Theoretic Scagnostics . 15.4 VisDB . ... ... . ..... . 15.5 Hierarchical Clustering Explorer . 15.6 PivotGraph . 15.7 InterRing . . .. 15.8 Constellation .. 15.9 Further Reading
Figure Credits
Bibliography
Idiom and System Examples Index
Concept Index
xiii
331 333 334 337
340 341 341 342 347 351 355 358 360 366
369
375
397
399
