Lecture 7: Single View MethodsInformation Visualization

CPSC 533C, Fall 2011

Tamara Munzner

UBC Computer Science

Wed, 28 September 2011

1 / 34

Required Readings

Chapter 5: Single View Methods

Trellis paper moved to Multiple Views on Monday

2 / 34

Further Reading

Milestones in the History of Thematic Cartography, StatisticalGraphics, and Data Visualization. Friendly and Denis.http://www.math.yorku.ca/SCS/Gallery/milestone/

Bars and Lines: A Study of Graphic Communication. Zacks andTversky. Memory and Cognition 27(6):1073-1079, 1999.

Multi-Scale Banking to 45 Degrees. Heer and Agrawala. IEEETVCG 12(5) (Proc. InfoVis 2006), Sep/Oct 2006, pages 701-708.

Overview Use in Multiple Visual Information Resolution Interfaces.Lam, Munzner, and Kincaid. Proc. InfoVis 2007.

VisDB: Database Exploration using Multidimensional Visualization.Keim and Kriegel. IEEE CG&A, 1994

3 / 34

Principles, Methods, and Techniques...

part 1: principles (3 chapters)

why underlying many design decisionsdata, visual encoding, interaction

part 2: methods (4 chapters)

what are the axes of the (current) design spacetaxonomy of design considerations

how many views? single, multiplehow to reduce what’s shown? data, dimensions

part 3: techniques (3 lectures [∼4 chapters...])

analyze techniques by which methods/principles usedtables, graphs, (text/logs), spatialgrouped by data type to follow nested model

technique level design happens after data type chosen atabstraction level

4 / 34

... and Practice

part 4: practice (2 lectures)

problem identification and task abstractionvalidation at problem, abstraction, technique levels

research process/papers

5 / 34

Experiment

which lecture style works best?

summarize chapters thoroughly

last several lecturesif book doing its job, maybe other choices viable!

summarize lightly

also bring up other ideas/approachesmore time for discussiontrying this today

end of class: get feedback from you

6 / 34

Single View Methods

all information integrated in one view

basic visual encodings

spatial positioncolorother channelspixel-oriented techniques

visual layering

global compositingitem-level stacking

glyphs

7 / 34

Spatial Position

most statistical graphics

bar chart, histogram

8 / 34

Spatial Position

most statistical graphics

bar chart, histogram, dot plot, line chart

9 / 34

Statistical Graphics

heavy focus on spatial position for visual encoding

long history for paper-based views of data

springboard for infovishttp://www.datavis.ca/milestones/

many ways to make interactive (more later)

many ways to refine/improve/combine

10 / 34

Line Charts

invented by William Playfair (1759-1823)

also bar charts, pie charts, ...

http://labspace.open.ac.uk/file.php/1872/Mu120 3 021i.jpghttp://www.math.yorku.ca/SCS/Gallery/images/playfair-wheat1.gif

11 / 34

Banking to 45 Degrees

previous work by Cleveland

perceptual principle: mostaccurate angle judgement at 45degrees

pick line graph aspect ratio(height/width) accordingly

[www.research.att.com/∼rab/trellis/sunspot.html]

12 / 34

Multiscale Banking to 45

frequency domain analysis

find interesting regions at multiple scales

[Multi-Scale Banking to 45 Degrees. Heer and Agrawala, Proc InfoVis 2006vis.berkeley.edu/papers/banking]

13 / 34

Choosing Aspect Ratios

FFT the data, smooth byconvolve with Gaussian

find interestingspikes/ranges in powerspectrum

cull nearby regions if toosimilar, ensure overviewshown

create trend curves foreach aspect ratio

14 / 34

Multiscale Banking to 45

[Multi-Scale Banking to 45 Degrees. Heer and Agrawala, Proc InfoVis 2006vis.berkeley.edu/papers/banking]

15 / 34

Bar vs Line Charts

line implies trend, do not use for categorical data

•

•

Female Male0

10

20

30

40

50

60

He

igh

t (in

ch

es)

Female Male0

10

20

30

40

50

60

He

igh

t (in

ch

es)

10-year-olds 12-year-olds0

10

20

30

40

50

60

He

igh

t (in

ch

es) •

•

10-year-olds 12-year-olds0

10

20

30

40

50

60

He

igh

t (in

ch

es)

[Fig 2. Zacks and Tversky. Bars and Lines: A Study of Graphic Communication.

Memory and Cognition 27(6):1073-1079, 1999.]

16 / 34

Scatterplots

encode two input variables with spatial position

show positive/negative/no correllation between variablesshow clusters: clumpiness/density, shape, overlap

[http://upload.wikimedia.org/wikipedia/commons/0/0f/Oldfaithful3.png]

17 / 34

Scatterplots

or compare correlation/clusters for two position attributesagainst more attributes encoded with color/shape

[Fig 1c. Robertson et al. Effectiveness of Animation in Trend Visualization. IEEETrans. on Visualization and Computer Graphics 14(6):1324-1332 (Proc. InfoVis08),2008.] 18 / 34

Colormap Taxonomy

http://www.colorbrewer.org

[Brewer, www.personal.psu.edu/faculty/c/a/cab38/ColorSch/Schemes.html]

19 / 34

Rainbows: The Good, The Bad, The Ugly

[Fig 1. Rogowitz and Treinish. Data visualization: the end of the rainbow. IEEESpectrum 35(12):52-59 1998.] [Fig 2,1. Bergman and Rogowitz and Treinish. ARule-based Tool for Assisting Colormap Selection. Proc. IEEE Vis 1995, p 118-125.][Kindlmann. http://www.cs.utah.edu/ gk/lumFace]

20 / 34

Accuracy/InfoDensity Tradeoff:Position/Color

!"#$ !"#%

!"!#

$"

$"%

&"

$"%&#

$#

$''

&'

$"'

$'(

&)

*

+

$,

$"(

-

.

$#

$"(

/"

!'(

$#%

(0(( "0((

(0(( "0((

!"# !"#"$1234-3+ 3565375 .8+835*982:-;

%"&'()*+&"$ %"&,+-$&5-372:<==

!#>?>!5:5#>>(0'%

$%

@A0#

>%0'

$&

@,0"

>)0,

@,0"

>,0,

@A0(

>%0'

@'0%

>A0)

@A0#

>'0(

@)0)

>)0B

>(0(

>)0#

@'0)

>#0'

@"0A

>'0C

@#0"

>)0,

@)0B

>'0,

@'0B

>'0C

@#0B

>(0A

@"0"

>(0(

@"0(

>"0'

'%

'&

'(

')

'*

'+

',

'-

'.

'%/

'%%

'%&

'%(

'%)

0123456" 172884798:7;8<01:125="79>6;2586147?"68<5;@0123A1# <"2:5;78=":=5"A @0";5"@ ;>:="<10 8=":41#@

!111!222

mmm111

mmm111888

PPP111

mmm111888PPP222

mmm222

mmm333333

PPP333

mmm111333

mmm333000

PPP444

ccc

ddd

mmm777

mmm111000

aaa

bbb

mmm222

mmm111000

BBB111

!333000

mmm222888

000...000000 111...000000

000...000000 111...000000

!EEYY !eeennneeesssfffooorrrwwwaaarrrddd rrreeevvveeerrrssseee bbbiiidddiiirrreeeccctttiiiooonnnaaalll

MMMeeetttaaabbbooollliiittteeesss MMMeeetttrrriiicccsssPPPeeeaaarrrsssooonnnAAALLLLLL

!222 /// !eeennneee222 000...333888

gg11

---555...222

888...333

gg22

---777...111

444...777

---777...111

777...777

---555...000

888...333

---333...888

555...444

---555...222

333...000

---444...444

444...666

000...000

444...222

---333...444

222...333

---111...555

333...999

---222...111

444...777

---444...666

333...777

---333...666

333...999

---222...666

000...555

---111...111

000...000

---111...000

111...333

ss11

ss22

ss33

ss44

ss55

ss66

ss77

ss88

ss99

ss1100

ss1111

ss1122

ss1133

ss1144

PPAATTHHLLIINNEE AA    TTOOOOLL    FFOORR    CCOOMMPPAARRAATTIIVVEE    FFUUNNCCTTIIOONNAALL    GGEENNOOMMIICCSSPPAATTHHWWAAYY MMEETTRRIICC    OOVVEERRVVIIEEWW SSPPEECCIIEESS CCUURRVVEEMMAAPP OOVVEERRLLAAYYSS

[Fig 4b,4a. Meyer et al. Pathline: A Tool for Comparative Functional Genomics.Proc. EuroVis 10, p 1043-1052.]

21 / 34

Tradeoff: Empirical Study

[Fig 1. Lam, Munzner, and Kincaid. Overview Use in Multiple Visual InformationResolution Interfaces. Proc. InfoVis 2007] 22 / 34

Study: Control Room Scenario

Which location has the highest power surge for the given time period?(find extreme value, y-dimension)

A fault occurred at the beginning of this recording, and resulted in atemporary power surge. Which location is affected the earliest? (findextreme value, x-dimension)

[Lam, Munzner, and Kincaid. Overview Use in Multiple Visual Information ResolutionInterfaces. Proc. InfoVis 2007]

23 / 34

Study: Findings

tasks

Max: simple, local, no comparisonMost: complex, dispersed, no comparisonShape: complex, local, comparisonCompare: simple, local, comparison

resultslow-res / high-density used:

simple/local targets

(other findings about focus+context vs overview/detail)

see also horizon graphs study

24 / 34

Pixel-Oriented Methods: VisDB

how to draw pixels?

sort, color by relevance

local ordering

spiral 2D

[VisDB: Database Exploration using Multidimensional Visualization, Keim and Kriegel,IEEE CG&A, 1994]

25 / 34

VisDB Windows

grouped dimensions

separate dimensions

[VisDB: Database Exploration using Multidimensional Visualization, Keim and Kriegel,IEEE CG&A, 1994]

26 / 34

VisDB Results: Separate Dimensions

spiral 2D

[VisDB: Database Exploration using Multidimensional Visualization, Keim and Kriegel,IEEE CG&A, 1994]

27 / 34

VisDB Results: Grouped Dimensions

[VisDB: Database Exploration using Multidimensional Visualization, Keim and Kriegel,IEEE CG&A, 1994]

28 / 34

Visual Layering

beyond simple use of visual channels

method variants

global compositing: everything superimposeditem-level stacking

major consideration

static layers: disjoint ranges in channels safestdynamic/interactive layers: more freedom

29 / 34

Visual Layering: Constellation

global compositing, dynamic layers

video

[Munzner. Constellation: Linguistic Semantic Networks. Interactive Visualization ofLarge Graphs and Networks (PhD thesis) Chapter 5, Stanford University, 2000, pp87-122. http://graphics.stanford.edu/papers/munzner thesis]

30 / 34

Glyphs

compound marks

macro (small picture) vs micro (texture)

channel questions

separabilityeffectiveness principle: importance matching

[Fig 9. Information Rich Glyphs for Software Management, IEEE CG&A 18:4 1998][Fig 2. Smith and Grinstein and Bergeron. Interactive data exploration with asupercomputer. Proc. IEEE Visualization (Vis) 1991, p. 248-254]

31 / 34

Questions/Discussion

32 / 34

Experiment: Feedback

which lecture style works best?

summarize chapters thoroughly

last several lecturesif book doing its job, maybe other choices viable!

summarize lightly

more time for other/further ideas/approachesmore time for discussiontrying this today

your preferences?

33 / 34

Reading For Next Time

Chapter 6: Multiple View Methods

The Visual Design and Control of Trellis Display. R. A. Becker, W.S. Cleveland, and M. J. Shyu (1996). Journal of Computationaland Statistical Graphics, 5:123-155.

34 / 34