Textual Data; Visual Variables

1
Textual Data Visual Variables
2
Administration

• Sprechstunden
• T 14-15, Th 16-17
• (on the web site)
• Online collaboration tools?
• e.g. wikis

3
Languages

• What languages do people know, and to what
  extent? (Basic, Intermediate, Advanced)
• CuC
• Advanced English
• Intermediate French, German, Italian, Spanish
• Basic Bambara, Fula, Dogon

4
Visualizing the metadata

• How might we visualize this metadata?
• Better how do we think about visualizing it?
• Important aspects
• What is the data? (The data)
• Who will be using it? (The user)
• What will they be trying to do? (The task)
• cf. Data repository project

5
The project

• Goal develop a scientific visualization of some
  kind of linguistic data
• Start thinking about what kind of data you want
  to visualize, and where you'll get it
• Who Small groups
• If you are inexperienced in programming, work
  with someone who is more experienced
• Progress? Questions?

6
Current visualizations

• What LInfoVis visualizations have you used?
• For what purposes?
• Did/do you wish they could be better? How?

7
About the references in the tutorial

• One more
• Ware, Colin. 2004. Information Visualization,
  Second Edition Perception for Design.
• Bertin seminal work on visual variables
• Card, McKinlay, Shneiderman, Ware Important
  research contributions to InfoVis
• Hearst Interesting combination of InfoVis, UI
  design, information retrieval
• Tufte Very influential (sometimes controversial)
  about presentation of information

8
Data types(cf. Tutorial, Hearst, 2009)

• Quantitative data numbers, etc. that can be
  processed arithmetically
• Categorical data everything else
• Interval ordered data with measurable distances
  (e.g. months)
• Ordinal ordered data without measurable
  distances (e.g. hot-warm-cold)
• Nominal data without (relevant) organization
  (e.g. weather types, a collection of names)
• Hierarchical data without order, arranged into
  subsuming groups (e.g. mammals, bear ,
  cat lynx,,, , etc.) cf. GermaNet
• ? Quantitative, interval, and ordered data are
  easier to convey visually than nominal data.

9
More on hierarchical data

• Some hierarchical data in linguistics is a bit
  more complicated.
• Consider a linguistic analysis tree.
• data without order
• arranged into subsuming groups
• Now consider a treebank (collection of trees)
• data without order (trees dont have, but what
  about metadata?)
• Arranged into subsuming groups (possible, but not
  natural)

10
Textual data(cf. tutorial)

• We are interested in the information/properties
  of textual elements e.g. word frequency,
  syntactic structure, emotion content, etc.
• However, in many cases, the actual textual items
  are important for understanding the information,
  so they must be indicated in the visualization.
• The categorical nature of text, and its very
  high dimensionality, make it very challenging to
  display graphically. (Hearst, 2009)

11
Non-mappability of text (cf. tutorial)

• The problem is less about the information about
  textual elements, but the textual elements
  themselves they take up space.
• Textual items are not mappable
• We (usually) cannot effectively represent them by
  something else meaningful (e.g. shape, color,
  position, etc.)
• Textual items are too variable (Hearsts high
  dimensionality) and too complex to be reduced to
  a more compact representation, even a label
• The details of the textual items are often
  crucial to understanding the data (e.g. context
  in a concordance)

12
What to do? (cf. tutorial)

• This is a huge challenge for LInfoVis!
• Not a solved problem
• Interactive visualizations will be the key to the
  solution(s)
• Show only some of the data, and interact to get
  more
• Data filtering/selection will also be key
• Cf. Data filtering/selection in InfoVis more
  generally
• Need domain and task specific information

13
Visual variables

• A visual variable is a visual property that can
  be varied in order to convey (encode) information.

14
Visual data transcription visual variables
(from tutorial)
Value Brightness
Taken from M. Carpendale, "Considering visual
variables as a basis for information
visualisation, Dept. of Computer Science,
University of Calgary, Canada, Tech. Rep.
2001-693-16, 2003, Table 1.
14
15
Visual variables characteristics (1) (from
tutorial)

• 5 key characteristics
• Selectivity Different values are easily seen as
  different
• Is A different from B?
• Worst case visual properties of all objects need
  to be looked at one by one

15
16
Visual variables characteristics (2) (from
tutorial)

• Associativity Similar values can easily be
  grouped together
• Is A similar to B?
• Positioning gt size, brightness gt color,
  orientation (for points) gt texture gt shape

Full selectivity / associativity
No selectivity / associativity
16
17
Visual variables characteristics (3) (from
tutorial)

• Order Different values are perceived as ordered
• Is A more/greater/bigger than B?
• Size and brightness are ordered
• Orientation, shape, texture are not ordered
• Hue is somewhat ordered

17
18
Visual variables characteristics (4) (from
tutorial)

• Quantity A number can be deduced from
  differences
• How much is the difference between A and B?
• Position is quantitative, size is somewhat
  quantitative
• The other variables are not quantitative

18
19
Visual variables characteristics (5) (from
tutorial)

• Length The number of distinctions possible using
  the variable
• How many different things can we represent with
  this variable?
• Shape, Texture infinite, but
• Brightness, hue 7 (Association) 10
  (Distinction)
• Size 5 (Association) -20 (Distinction)
• Orientation 4

19
20
An experiment
21

Made with http//www.wordle.net/
22
Made with http//www.tocloud.com/
23
Combining visual variables

• When we use 2 visual variables for an element,
  how independently do we perceive the two
  variables?
• If we perceive them separately, they are
  separable.
• If we perceive them together, they are integral.

24
Test 1

C
B
A
25
Test 2

C
B
A
26
Integral/Seperable

• Red/green and yellow/blue
• Width and height
• Size and orientation
• Color and shape
• Color and motion
• 
  
  Ware 2004181
• Color and location

27
Integral/Seperable

• Red/green and yellow/blue
  More integral
• Red/green and black/white
• Width and height
• Shape and size
• Color and size
• Shape and direction of motion
• Color and shape
• Color and direction of motion
• Position and size OR shape OR color More
  separable
• 
  
  cf. Ware 2004180

28
Gestalt psychology and perception

• Early 20th century, with a lot of work on aspects
  of visual perception and how people organize what
  they see.
• Relevant for visualization
• Reification (somewhat)
• we perceive more information than
• is present
• e.g. illusory contours
• Source http//en.wikipedia.org/wiki/FileKanizsa_
  triangle.svg

29
Gestalt Principles of Grouping

• Some basic sources
• http//psychology.about.com/od/sensationandpercept
  ion/ss/gestaltlaws.htm
• http//en.wikipedia.org/wiki/Principles_of_groupin
  g
• http//en.wikipedia.org/wiki/Gestalt_psychology
• More detail
• http//www.scholarpedia.org/article/Gestalt_princi
  ples

30
Gestalt Principles of Grouping

• Similarity
• Objects with common visual attributes are
  perceived as being part of the same group
• Source
• http//psychology.about.com/od/sensationandpercept
  ion/ss/gestaltlaws_2.htm

31
Gestalt Principles of Grouping

• Proximity
• Objects that are near each other in space (or
  time) are perceived as forming a group.

O O O O O
O O O O O
O O
32
Gestalt Principles of Grouping

• Continuity
• A pattern of objects (in space, time) tends to be
  continued.
• Source http//www.scholarpedia.org/article/Gestal
  t_principles

33
Gestalt Principles of Grouping

• Law of Common Fate
• Objects moving in the same direction are
  perceived as a group.
• Note Difficult to use in (L)Infovis, since we
  need to be able to focus easily on the
  information.

34
Using visual variables (1) (cf. tutorial)

• Sameness of a visual element implies sameness of
  what the visual element represents
• (Tufte, 2006)
• Cf. Principles of grouping

35
Using visual variables (2) (cf. tutorial)

• Characteristics of visual variables determine
  their use
• e.g. Ordered values have to be represented by
  ordered visual variables

36
Using visual variables (3) (cf. tutorial)

• Be consistent concerning relations of similarity,
  proportion and configuration

37
Using visual variables (4) (cf. tutorial)

• Adhere to conventional uses of visual variables
• e.g. in cartography use blue color for water
• Scales should be made up of visually equidistant
  values of a variable

38
Using visual variables (5) (cf. tutorial)

• The full range of a visual variable should be
  used
• e.g. when using shades of gray, use from white to
  black
• The number of visual variables of a visualization
  should correspond to the dimensionality of the
  represented information
• But sometimes dual encoding can be useful

39
Using visual variables (6)

• When combining two visual variables, if people
  should be able to analyze the two attributes
  independently, then separable variables should be
  used.

40
For next time