Lecture 04 VisualizationDesign
Lecture 04 VisualizationDesign
Design
DataVis 2020
http://datavis2020.github.io
understandable
?
Too
simple:
Too
complex:
Good
:
What makes a good
visualization?
Visualization
Guidelines
How to decode
bubbles?
a
)
Wrong
sizes
Guideline
#1
Exception: none
Common flaws (from 8 graphics
only..)
1. Missing scales 1. Wrong legends
2. 3D distortion 2. Bad visual mappings
3. 3D occlusion 3. Uncommon / unknown
4. Missing color contrast design choices
5. Colors are too similar 4. Too much clutter
6. Not B/W safe 5. Misplaced data points
7. overuse of texture 6. Long lines, which are hard to
8. Volume as visual variable follow
9. Un-proportional 7. Missing legends
mappings 8. Too much information in
10. Truncated axes one graphic
11. Missing titles 9. Missing descriptions
12. Missing axis labels 10. Overlapping graphical
13. Unnecessary shapes/ink elements
14. Incomplete data 11. Too many charts in one
Data-ink
ratio
Guideline
#5
Exception: ...
Embellishment:
Metaphors
Embellishme
nt
Metaphor Distractors
Embellishme
nt
Exceptions: none
Highlig
ht
Highlig
ht
Highlig
ht
Messag
e
Messag
e
Messag
e
Guideline
#8
"Annotate w h a t is important
"
Exceptions: ?
Uncertain
ty
Clarity vs.
Memorability Novelty
Uncertainty
Reader-driven vs.
Author-driven
...
Closing
Quotes
Nigel Holmes states: “As long as the artist
understands t hat the primary function is to convey
statistics a n d respects that duty, then you c a n have
fun (or be serious) with the image; that is, the form in
which these statistics appear.”