2. #DataVizInSixWeeks
Copyright Anne Stevens
Week One
What is data visualization?
Historical context
Week Four
Design issues & best practices
Week Two
Visualization types
Week Five
Big data, data management
Week Three
Perception and cognition
Week Six
Synthesis
Data Viz In Six Weeks
An Introduction to Visual Analytics course taught at OCAD University, Toronto
By Anne Stevens
6. #DataVizInSixWeeks
Copyright Anne Stevens
The representation of numbers, as physically measured on the
surface of the graphic itself, should be directly proportional to the
quantities represented.
(Edward Tufte)
9. #DataVizInSixWeeks
Copyright Anne Stevens
Data / ink ratio
Data-ink ratio =
total ink used to print the graphic
data-ink
proportion of a graphic’s ink devoted to the
non-redundant display of data-information
=
1.0 – proportion of a graphic that can be
erased without loss of data information
(Edward Tufte)
=
17. #DataVizInSixWeeks
Copyright Anne Stevens
High density vs summary information
Edward Tufte
The Visual Display of Quantitative Information
Summary graphics can emerge
from high-information displays, but
there is nowhere to go if we begin
with a low-information design.
(Edward Tufte)
18. #DataVizInSixWeeks
Copyright Anne Stevens
Data-rich designs give context and credibility to statistical
evidence. Low-information designs are suspect: what is left out,
what is hidden, why are we
shown so little? (Edward Tufte)
49. #DataVizInSixWeeks
Copyright Anne Stevens
Week One
What is data visualization?
Historical context
Week Four
Design issues & best practices
Week Two
Visualization types
Week Five
Big data, data management
Week Three
Perception and cognition
Week Six
Synthesis
Data Viz In Six Weeks
An Introduction to Visual Analytics course taught at OCAD University, Toronto
By Anne Stevens
stevensanne.com
stevensanne.com/blog/
@3_ring_binder
Editor's Notes
If you know 1 dataViz guru …
The Visual Display of Quantitative Information
Statistician, very concerned with precision of charts, so that they don’t dumb down the numbers.
Wk: 1: a framework of rules to distinguish data viz from info viz.
Thinks of graphics sim to Golden Age of SG: crunch the numbers first, then find the best viz to represent it.
Rigidly rejects: icon, metaphor, 3D
Artistic vis, graphic design
Chart junk
Book is full of rules, typically straightforward.
The single biggest threat to a presentation is cherry-picked data – or sources.
Ask yourself: am I seeing all the data, or just some of the data?
Book filled with rhetoric of “lies”, deceit, distortion, suspicion, manipulation.
Lie factor < 0.95 or > 1.05 = substantial distortion.
Logarithm of the LF
Context trumps data/ink ratio law
Suspicion, again.
Drill down.
Provide all the data
If you have high density, multi-dimensional, layered viz’s
You need to be able to drill down, move up, pan, zoom, select, filter
Brings us to the importance of interaction.
William Playfair, 1780s, Scottish imports & exports
Static image, for presentation only.
No layers, filters / interaction / live data / aggregation.
TODAY big data/live data demand computation and interaction to make sense of it.
TODAY computers let us do live data, interact, save parameters and searches, scroll in time.
As soon as we are talking about computerization, we have to think about UX and IxD.
What would you rather have: post-card/map of venice or explore it yourself?
what’s left out? / - not live / - not current
Want to explore for yourself
Exploration: multiple readings, experiences
Build your own mental maps: better understanding of underlying data.
Different visitor / user experience
Even worse – query lists - when you have to leave the viz space entirely.
Query lists
Bill Buxton: have to pay as much attention to designing the transitions to designing the static states.
Compare to instant transitions in Tableau and Many Eyes.
Appliance view: add Range. Movement as preattentive feature
Preattentive perception re. movement
Helps preserve context & create meaning
DM and touchscreen
- Multiple ways to interact
- encourages exploration
- Fewer error messages, less frustration
Play with your data
Touch screen possibilities
Preserves context
Preserves context
Watch video up to Births/Deaths
Movement helps show real time activity, while preserving context
Trail Chart: starts 3:40 mins: trails and tails
Refuses to reduce data to primitive forms.
Look at data impressionistically.
Zoom, filter, detail on demand, context, scale, movement
Full size image, not abstract/primitive dots.
Analyse art as image and data at the same time.
Tufte re. metaphor
Largely descriptive metaphors … don’t contribute much to the substance at hand. Such metaphors are inevitably dequantified, turning data into vague, cute shapes.
Metaphor Theory in comm’n & cognition: George Lakoff
Development of metaphorical concepts during normal childhood development
Complex metaphors are made up of simpler ones.
And so it is with visual language.
Like Gestalt phenomena, we have evolved to use metaphor to communicate.
It is important to understand the rich heritage of metaphor.
But also to understand that metaphors are not universal.
But also to understand that metaphors are not universal.
Metaphor is part of visual communication in data viz as much as any other medium.
What about Tableau re. interactivity, metaphor, 3D, colour, graphic integrity