iliinsky uie vs 12 copy

Transcription

iliinsky uie vs 12 copy
Telling the Right Story
with Data Visualization
UIE Virtual Seminar
February 2nd, 2012
Noah Iliinsky
@noahi
Why Visualization?
Visualizations
leverage awesome
capabilities of our
eyes & brains
Visualizations give
access to huge &
ubiquitous data
“Stories” give
context to data
http://www.youtube.com/watch?v=hVimVzgtD6w
Success Depends on Inputs...
Know your
motivations & goals
Reader
Consider their tasks,
needs, everything
Your
visualization
Know the “shape” of
your data
Data
... and Outputs
Designer
Two Phases
1. What to visualize
2. How to visualize
Consider your inputs
Select axes for your
most important data &
relationship
your goals
their needs
shape of the data
Write a spec for your
visualization
Consider & apply
encodings for all
other data dimensions
Experiment, iterate,
etc.
select data to
include
Phase 1: Know your data
handwave the UX stuff
(you know this already)
Acquire data
buzzdata.com
infochimps.com
data.gov
Examine the data
http://buzzdata.com
Data Sets
Health care
expenditure
1995-2009
life expectancy
1965-2009
By Geography:
Countries
Regions
Classifications
http://buzzdata.com
Calculated Data Set
Health care
expenditure 1995
life expectancy 1995
life expectancy 2009
Health care
expenditure 2009
Change in LE (years)
Change in spending ($)
Change in LE (%)
Change in spending (%)
Geography
geography is categorical, all else numeric.
Write the Spec!
The spec gives you guidance on direction and a
standard for successful completion.
What’s an interesting question to answer?
“How have changes in health care spending
affected life expectancy in different
countries between 1995 and 2009.”
Write the Spec!
“How have changes in health care spending
affected life expectancy in different
countries between 1995 and 2009.”
Implied data requirement:
change in spending
change in LE
per Country
(Hint: it’s not quite that easy.)
End Phase 1: What
~ intermission ~
Next up Phase 2: how
Two Phases: on to Phase 2
1. What to visualize
2. How to visualize
Consider your inputs
Select axes for your
most important data &
relationship
your goals
their needs
shape of the data
Write a spec for your
visualization
select data to
include
Consider & apply
encodings for all
other data dimensions
Experiment, iterate,
etc.
Successful Encodings
Axes Define Your World
Select axes for your most important data &
relationships
“How have changes in health care spending
affected life expectancy in different
countries between 1995 and 2009.”
Key relationship is between change in
expenditure and life expectancy
Therefore: Axes should be change in
expenditures and life expectancy
Piece of cake!
First Iteration: all data
Second Iteration: no outliers
Third Iteration: regions only
Fourth Iteration: new axes
Fifth Iteration: newer axes
Selecting Additional Encodings
Expenditure! (quantitative)
size is good
Selecting Additional Encodings
Financial groupings (categorical)
color is good, how about green for $
Selecting Additional Encodings
Regional groupings (categorical)
shape is good, how about round for globe
Selecting Additional Encodings
Expenditure! (quantitative)
size is good
Financial groupings (categorical)
color is good, how about green for $
Regional groupings (categorical)
shape is good, how about round for globe
Sixth Iteration: With Encodings
Two Phases: Phase 2 Complete
1. What to visualize
2. How to visualize
Consider your inputs
Select axes for your
most important data &
relationship
your goals
their needs
shape of the data
Write a spec for your
visualization
Consider & apply
encodings for all
other data dimensions
Experiment, iterate,
etc.
select data to
include
Thank you!
Noah Iliinsky
@noahi
Intentional Communication from Data to Display
/Theory/In/Practice
Designing
Data
Visualizations
Noah Iliinsky & Julie Steele
How to do it
Beautiful
Visualization
Looking at Data Through the Eyes of Experts
Edited by
Julie Steele
& Noah Iliinsky
How they did it