16. [Proposed] Time Plan
29 october 2012
- Data + Perception + Data Mapping
- Bad/Good Infographic Guidelines
5 november 2012
- Infovis, Storytelling,Research Results, ...
- Compelling Dataviz Examples
18. Six Degrees of Mohamed Atta
http://business2.com/articles/mag/0,1640,35253,FF.html
19. US Terrorism Response Org Chart
http://www.cns.miis.edu/research/cbw/domestic.htm#wmdchart
20.
21.
22. Space Shuttle Launch
. O-ring damage data
. launch or not launch?
. risk of human lives versus loosing reputation
. ambient temperature at launch: 25-30 degrees F
36. Choice of “Visual” / “Data”
. anything can be ‘translated’ in anything
. can be ineffective (wrong answers)
. can be inefficient (takes too much effort)
. can be disengaging (no users, giving up, ...)
37. How to Design Visualization?
. 1. understanding properties of the image
. 2. understanding properties of the data
. 3. understanding how to map data to an image
42. Pre-Attentive Features
. time taken to make a decision is constant
. and is less than 200-250ms (< eye movement)
. independent of number of added detractors
. primitive features, low-level visual processing
. salience depends on: ‘strength’, and ‘context’
43. ‘Pop-out’ Features
. form: line orientation, length, width, visual marks,...
. color: hue, intensity, ...
. motion: flicker, direction of motion, ...
. spatial position: depth, convex/concave shape,...
44. orienta(on size
length,
width closure
curvature density,
contrast number,
es(ma(on colour
(hue)
Some already known pre-attentive visual features
45. Viewer can rapidly & accurately determine
whether the target (red circle) is present or absent:
difference detected in color
Pre-Attentive Processing - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
46. Viewer can rapidly & accurately determine
whether the target (red circle) is present or absent
difference detected in form (curvature)
Pre-Attentive Processing - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
47. Viewer cannot rapidly & accurately determine whether target is present or
absent when target has combined two or more features, also present in the
distracters. Viewer must search sequentially
Pre-Attentive Processing - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
48. Hue-on-form feature hierarchy: (a) a horizontal hue boundary is pre-
attentive identified when form is held constant; (b) a vertical hue boundary
is pre-attentively identified when form varies randomly in the background
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
49. Hue-on-form feature hierarchy: (c) a vertical form boundary is preattentively
identified when hue is held constant; (d) horizontal form boundary cannot be pre-
attentively identified when hue varies randomly
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
50. Viewer cannot rapidly & accurately determine border by a conjunction of
features (red circles & blue squares on the left, blue circles and red squares on
the right)
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
51. Target has a unique feature with respect to distractors (i.e. open sides).
The group can be detected pre-attentively.
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
52. Target does not have a unique feature with respect to distractors.
The group cannot be detected pre-attentively.
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
53. A sloped line among vertical lines is pre-attentive.
A sloped line among other sloped ones is not.
Feature Hierarchy - http://www.csc.ncsu.edu/faculty/healey/PP/index.html
68. Law of Symmetry
We assume the horizontal / vertical part is identical. Good for before/after comparisons...
69. •pattern
recognition
•we
are
constantly
grouping
objects
based
on
color,
shape,
direction,
proximity,
closure/enclosure,
...
•so:
where
should
a
chart
legend
be
located?
•
keep
it
‘simple’
•we
like
simple,
close,
smooth,
symmetrical,
easy-‐to-‐
process
shapes...
•design
accordingly
•if
we
are
aware
of
these
laws
we
can
take
advantage
of
them
to
design
better
charts
or
dashboards.
•be
careful
•be
aware
of
their
negative
impact:
we
shouldn’t
force
the
reader
to
see
groups
that
aren’t
really
there
http://www.excelcharts.com/blog/data-visualization-excel-users/gestalt-laws/
80. Information Visualization (Scientific) Visualization
Examples Examples
stock market, friend clouds, microscopic events,
network, DNA functions, ... human organs, ...
Data Characteristics Data Characteristics
Abstract Concrete
Multi-dimensional 2 or 3 dimensional
Mostly time-dependent
Requirements. Requirements
Visual metaphor 3D and fast rendering
User interaction User interaction
Focus: Exploration Focus: Analysis
then Analysis then Exploration
then Presentation then Presentation
81. Kinds of data...
S Stevens “On the theory of scales and measurements” (1946)
84. Categorical / Hierarchical Data
. can be categorized
. e.g. alphabetically, thematically, functionality
. e.g. desktop folder hierarchy, work hierarchy, ...
85
85. Network / Relational Data
. one-to-many, many-to-one relationships
. often these are ‘weighted’
. e.g. social networks, people working on projects, ...
86
86. Nominal / Unstructured Data
. no order, no units, only “equal or different”
. e.g. Australia, Belgium, Mexico
87
87. Temporal / Dynamic Data
. time dependent
. related to progress of time, history, ...
. e.g. stock market, news stories, sensor readings,...
88
89. § dimensionality
§ # of attributes
§ scale / size
§ # of items
§ value range
§ bits/value, min/max, etc.
§ time
dependency?
90. “The current complexity of data is
staggering, and our ability to collect
data is increasing at a faster rate that
our ability to analyse it.”
Data Complexity
91. How is data ‘complex’?
. size: number of records
. dimensionality: number of attributes
. time-dependency: data changes over time
92. Data is more complex now?
. complexity of human society always increasing
. quantity always increasing, never decreasing!
. speed of data creation always increasing
99. because the data is abstract...
“the challenge is to invent
new metaphors for
presenting information &
developing ways to
manipulate these
metaphors to make sense
out of the information...”
Information Visualization ‘Design’ Challenge
100. Data + data mapping = visual representation that can be (relatively fairly) interpreted
101. data insight
10010110 knowledge
transfer
data mapping
mapping
inversion
visualisation comprehension
!
visual transfer
Data Mapping Methodology
102. data §data mapping requirements?
§ computable
10010110 § no user interaction required
§ algorithm: data -> value
data § comprehensible
mapping § user understands
§ intuitively, within short time
visualisation § invertible
§ mapping backwards from
§ form to exact data value
103
Data Mapping Methodology
106. 1 items
data
ê
???
2 attributes
data
ê
???
Data Characteristics
From abstract data to visual form
107. 1 data items
ê
visual objects
2 data attributes
ê
visual object properties
Data Mapping
From abstract data to visual form
108. objects
point, line, area, volume, ...
properties
position
size, length, area, volume
orientation, angle, slope
color, texture, transparency
shape
animation, time, blink
Visualization “Language”
From abstract data to visual form
109. Example scatterplot of movie dataset
Year → X
Length → Y
Popularity → size
Subject → color
Award? → shape
110. Data Mapping Limitations
. data scale: one object for each tuple?
. data dimensionality: visual cue for each attribute?
. value range: metaphorical range?
. time dependency: can metaphor change?
120. “‘Graphical Excellence’ is that which
gives to the viewer a great number of
ideas in the shortest time, with the
least ink, in the smallest space”
. “show data variation”, not “design variation”
. communication: clarity, precision, efficiency
. simplicity of design - complexity of data
121. French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
124. 1. Show comparisons, contrasts, differences
2. Show causality, mechanism, explanation,
systematic structure
3. Show multivariate data; that is, show more
than 1 or 2 variables
4. Completely integrate words, numbers, images,
diagrams
5.Thoroughly describe the evidence: title, authors
and sponsors, data sources, add measurement
scales, highlight relevant issues
6.Analytical presentations ultimately stand or fall
depending on the quality, relevance and
integrity of their content
Principles for the Analysis and Presentation of Data - Tufte
126. 2. Show causality, mechanism,
explanation, systematic structure
French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
127. 3. Show multivariate data; that is,
show more than 1 or 2 variables
French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
128. 4. Completely integrate words,
numbers, images, diagrams
French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
129. 5.Thoroughly describe the
evidence: title, authors and
sponsors, data sources, add
measurement scales, highlight
relevant issues
French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
130. 6.Analytical presentations
ultimately stand or fall depending
on the quality, relevance and
integrity of their content
French Invasion of Russia (Minard, +-1864)
Napoleon Retreat (Minard, +-1864)
139. 2. Data-Ink Ratio
2. Data-Ink Ratio
Unnecessary & distracting patterns - graphics emphasizing style over information
140. 2. Data-Ink Ratio
2. Data-Ink Ratio
Unnecessary & distracting patterns - graphics emphasizing style over information
141. Five Laws of “Data-Ink”
• Above all else show (only) the data
• Maximize the data-ink ratio
• Erase non-data ink
• Erase redundant data-ink
• Revise and edit
2. Data-Ink Ratio
Unnecessary & distracting patterns - graphics emphasizing style over information
172. 5. Expressiveness
Avoid size as quantitative
junkcharts.typepad.com value. If so, map values as surface area (and never as radius!)
173. 5. Expressiveness
Avoid size as quantitative
junkcharts.typepad.com value. If so, map values as surface area (and never as radius!)
174. 5. Expressiveness
Maps only useful for spatial distribution. They do not take into account population density
junkcharts.typepad.com
(which explains trends), physical size of districts (which are visually more prominent),...