Transcript ENVR 384: Introduction to Geographical Information Systems

Data Visualization Seminar
NCDC, April 27 2011
Todd Pierce
Module 1 Data Visualization
Introduction
This seminar will look at visualization from the
viewpoint of human perception and cognition
How do humans perceive and use visuals?
What are some principles that can be applied
to visualizations to make them more effective?
The seminar is a summary of the first half of the
UNC Asheville class “Tools for Climate Data
and Decision-Making”
Outline
1 Data Visualization – history, uses, good and bad
visuals
2 Human Perception – visual attendance, patterns, and
working memory
3 The Eightfold Way – principles for effective
visualizations
THEORY
Lunch break
4 Best Practices – color, parts of a graph, picking the
correct graph
5 Types of Graphs – types of analysis supported, do’s
and don’t’s
6 Maps – (if time allows)
PRACTICE
Sources
Sources
Sources
Let’s Get Started
Facebook Friends Graph
http://www.facebook.com/notes/facebook-engineering/visualizingfriendships/469716398919
Need for Climate Change Communication
Why are the skills in this course important?
- Climate Data needs to be a part of decision
making as humans must start enacting climate
mitigation and climate adaptation programs
- Climate Data is overwhelming in its quantity
and needs to be better presented in
visualizations – maps, charts, graphs – that
can be used in decision making
Need for Climate Change Communication
According to Global Climate Change Impacts in
the United States
-Global warming is unequivocal and primarily
human-induced
-Climate changes are underway in the US and
are projected to grow
-Widespread climate-related impacts are
occurring now and are expected to increase
-Future climate change and its impacts depend
on choices made today
Need for Climate Change Communication
Despite the need for choices to be made now,
climate change skepticism abounds
http://environment.yale.edu/uploads/SixAmericasJan2010.pdf
Need for Climate Change Communication
There is a need to counteract the skeptics, but
how? Climate Change is not a sound bite – it
has complex concepts and counterintuitive
findings as well as mountains of data.
Some examples…
Need for Climate Change Communication
Skeptics vs Scientific Consensus
http://www.informationisbeautiful.net/visualizations/climate-changedeniers-vs-the-consensus/
Increasing Sea Levels
http://www.informationisbeautiful.net/visualizations/when-sea-levels-attack/
Need for Climate Change Communication
Moscow Summer Heat Wave 2010
http://www.climatecentral.org/gallery/graphics/how_unusual_was_the_russi
an_heat_wave_of_2010/
Need for Climate Change Communication
Increased US Snow
http://www.climatecentral.org/gallery/graphics/arctic-paradox-warmerarctic-may-mean-colder-winters-for-some/
Data Visualization
So…data visualization can help explain climate
change data (as well as many other things)
Let’s look at data visualization
why use it?
when did it get started?
what makes a good or bad visualization?
Why Use Visualizations?
To explain and to persuade
“picture is worth a thousand words”
Visuals help meet several objectives
Why Use Visualizations?
Objectives for Visuals
-Clarity: make technical or numerical data easier
to understand
-Simplification: break down narrative
description into smaller parts (flow chart)
-Emphasis: draw attention to certain facts
-Summarization: show conclusions or main
points
Why Use Visualizations?
Objectives for Visuals
-Reinforcement: complement text and use
repetition to help remember idea
-Interest: break up blocks of text
-Impact: grab reader’s attention and keep it
-Credibility: impress reader with data validity
(“pictures don’t lie” ?)
-Coherence: help show how related parts of a
document work together
Definition
Data visualization: the visual
representations that support
the exploration, examination,
and communication of data.
• Information visualization:
abstract data
• Scientific visualization:
physical data, such as through
X rays or MRI scans
History
• Tables date to 2nd century CE, first ones in Egypt for
astronomical data for navigation
• Descartes created the Cartesian graph in the 17th
century, but for mathematical analysis, not for
information visualization
source: Stephen Few
History
• In late 18th/early 19th century, William Playfair
created or improved graphs for use in information
visualization – invented the bar graph, used line
graphs to show time trends, and invented the pie
chart.
source: Stephen Few
History
• First college course in graphs in 1913 at Iowa State –
today few courses offered outside of statistics classes
• John Tukey in 1977 started exploratory data analysis
as a tool for statistics – invented tools such as the
box plot to help show trends in data and prove
power of visualization for data exploration
source: Stephen Few
History
• Edward Tufte in 1983 published The Visual Display of
Quantitative Information, the first book to really
show effective and beautiful ways existed to show
data, and that most visuals did not use them
source: Stephen Few
History
• In 1984 the Apple Macintosh debuted – the first
affordable PC with a graphical interface
• William Cleveland in 1985 published The Elements of
Graphic Data – expanded on Tukey and improved use
of visualization in statistics
source: Stephen Few
History
• The National Science Foundation started efforts in
scientific visualization in 1986
• By 1999, information visualization was recognized as
distinct discipline within visualization in general
• Two conditions needed for modern information
visualization:
– graphical computers
– lots of readily accessible data.
– Before, data was limited to the printed page, which can
only be physically manipulated – the data is locked on the
page and can’t be changed. With computers, users can
interact with the data and explore ways to show it.
What Makes a Good Visual?
Easy to understand
Combines multiple data sources
Tells a story
Encourages aha! Moments
Leads to new insights and predictions
Often used in unrelated areas
“forces us to notice what we never expected to see”
– J W Tukey
What Makes a Good Visual?
Easy to understand
What Makes a Good Visual?
Easy to understand
What Makes a Good Visual?
Combines multiple data sources
What Makes a Good Visual?
Combines multiple data sources
What Makes a Good Visual?
Tells a story
What Makes a Good Visual?
Encourages aha! moments
What Makes a Good Visual?
What Makes a Good Visual?
Leads to new insights and predictions
What Makes a Good Visual?
Leads to new insights and predictions
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Good Visual?
Often used in unrelated areas
What Makes a Bad Visual?
Misleading or wrong
Ignores context
Ugly
Confusing
Obscures message
With computers, it is very easy to make a bad
chart, graph, or map
What Makes a Bad Visual?
Misleading or wrong (perspective issues)
What Makes a Bad Visual?
Misleading or wrong (area used for linear value)
What Makes a Bad Visual?
Misleading or wrong
What Makes a Bad Visual?
Misleading or wrong
What Makes a Bad Visual?
Misleading or wrong
What Makes a Bad Visual?
Misleading or wrong – track removed
What Makes a Bad Visual?
Misleading or wrong
What Makes a Bad Visual?
Ignores context
What Makes a Bad Visual?
Ignores context
What Makes a Bad Visual?
Ugly (“chart junk”)
What Makes a Bad Visual?
Confusing
What Makes a Bad Visual?
Obscures message
What Makes a Bad Visual?
Obscures message – better version
Next Module: Human Perception
How can we make visuals better, so they show
more of the ‘good’ qualities and less of the
‘bad’ qualities?
We can consider principles of human
perception.