i247: Information Visualization and Presentation Marti Hearst

1
i247 Information Visualization and
PresentationMarti Hearst
Data Types and Graph Types    
2
Outline

• The Roles and Stages of Visualization (briefly)
• Data Models and Types of Data
• Which Kinds of Graphs for Which Types of Data?
• Class Exercise

3
The Roles and Stages of Visualization
4
What Visualization Can Do (Ware)

• Allows comprehension of huge amounts of data.
• Allows perception of emergent properties
• Enables problems with the data to stand out
• Facilitates understanding at both large and small
  scales patterns linking local features
• Facilitates hypothesis formation.

5
What Visualization Can Do (Tufte 83)

• Show the data
• Induce to viewer to think about the data
• Avoid distorting what the data have to say
• Present many numbers in a small space
• Make large data sets coherent
• Encourage the eye to compare different pieces of
  data
• Reveal the data at several levels of detail, from
  overview to fine structure
• Serve a clear purpose
• Description, exploration, tabulation, or
  decoration
• Be closely integrated with the statistical and
  verbal descriptions of a data set.

6
Stages of Visualization (Ware)

• Collection and storage of data
• Preprocessing to transform data into something
  understandable
• Hardware and graphics algorithms for producing an
  image on the screen
• Human perceptual and cognitive system.
• (I think hes missing a stage Design of the
  visualization.)

7
Put it Into Questions

• What are our goals?
• What questions do we want to answer?
• What kind of data might we collect?
• How might we convey the information associated
  with this data?

8
Visualization Components
From Melanie Tory
9
Data Models and Types of Data
10
Basic Elements of a Data Model

• A data model represents some aspect of the world
• Data models consist of these basic elements
• objects
• values (also called attributes)
• relations

11
Basic Elements Objects

• Objects are items of interest
• people, plants, cars, films, etc
• Objects allow you to define and reason about a
  domain
• ecosystem ponds, streams, woodlands, mountains,
  plants, animals, etc.

12
Basic Elements Values

• Values (or attributes) are properties of objects
• Two major types
• quantitative
• categorical
• Appropriate visualizations often depend upon the
  type of the data values

13
Basic Elements Relations

• Relations relate two or more objects
• leaves are part of a plant
• a department consists of employees
• Ecosystem
• connections between streams and lakes
• predator/prey network of what eats what

14
Types of Data (Ware)

• Entities
• Relationships
• Attributes of Entities or Relationships
• Nominal / Ordinal / Interval / Ratio (Stevens
  46)
• Categorical / Integer / Real
• Operations Considered as Data
• Mathematical
• Merging lists
• Transforming data, etc.
• Metadata (derived data)

15
Types of Data (Few)

• Quantitative (allows arithmetic operations)
• Categorical (group, identify organize no
  arithmetic)
• Nominal
• Ordinal
• Interval
• Hierarchical

16
Types of Data

• Quantitative (allows arithmetic operations)
• 123, 29.56,
• Categorical (group, identify organize no
  arithmetic)
• Nominal (name only, no ordering)
• Direction North, East, South, West
• Ordinal (ordered, not measurable)
• First, second, third
• Hot, warm, cold
• Interval (starts out as quantitative, but is made
  categorical by subdividing into ordered ranges)
• Time Jan, Feb, Mar
• 0-999, 1000-4999, 5000-9999, 10000-19999,
• Hierarchical (successive inclusion)
• Region Continent gt Country gt State gt City
• Animal gt Mammal gt Horse

17
Which Types of Graphs for Which Kinds of Data?
18
Quantitative Against Categorical
19
Quantitative against Quantitative
20
Questions to ask when creating a graph

• Is a graph needed?
• Yes, if illustrating relationships among
  measurements
• What information is being conveyed?
• What is most important?
• Start by writing a title

21
Questions to ask when creating a graph

• What data is needed to answer specific questions?
• Overview? Relationships?
• Grices maxims
• combine relevant information together
• dont show extraneous information
• Who is your audience?

22
What Format to Use?

• Bertin has a notion of efficiency
• Tufte says show the data
• Lets start with familiar graph types
• line graphs
• bar charts
• scatter plots
• layer graphs
• When to use each?

23
Anatomy of a Graph (Kosslyn 89)

• Framework
• sets the stage
• kinds of measurements, scale, ...
• Content
• marks
• point symbols, lines, areas, bars,
• Labels
• title, axes, tic marks, ...

24
When to use which type?

• Line graph
• x-axis requires quantitative variable
• differences among contiguous values
• familiar/conventional ordering among ordinals
• Bar graph
• comparison of relative point values
• Scatter plot
• convey overall impression of relationship between
  two variables

25
What to put on the x axis?

• Independent vs. Dependent variables
• we often measure one quantitative variable
  against another
• the value of one changes in relation to the other
• the dependent variable changes relative to the
  independent one
• the independent variable acts as a measuring
  stick
• Independent usually goes on the x (horizontal)
  axis

26
Independent vs. Dependent

• Independent vs. Dependent variables
• heat in degrees against time
• sales against season
• tax revenue against city
• What happens when there is more than one
  independent variable?
• Choose one for the x axis, and another as a
  variation in the mark (color, shape)

27
Few on How to Show Information

• The best way to show a single value?
• Use a textual representation.
• Why?
• How to draw attention to a number?

28
Few on How to Show Information

• What are tables good for?
• Data lookup
• Hierarchical relationships

29
Class Exercise
30
How to Combine Data Types?

• Class Exercise
• Using data about autos from the 70s
• Each person get a column of data
• First, identify the data type
• Then, stand up
• Then, repeat the following several times
• Walk up to someone else. If they have a
  different column than you do, discuss whether and
  how you should plot your two columns.
• If yes, what question are you answering?
• If no, why not?
• Then, repeat this, but with groups of three
  people.