Data Analytics Chapter 2 - VIS
Data Visualization
The process that transforms data into graphical representations for the purpose of exploration, confirmation or communication.
Insight
The relations shown between different data.
- They are made by humans, NOT computers!
High-level action: Visualization for consuming data
- Needs good balance, details and message
- Dataset will not be changed
- End users, non-technical
What is good data visualization?
Combines human and computer strength, makes data accessible, effectivaly enables insight and communicates truthfully.
High-level action: Visualization for producing data
- Extends dataset
- Interactive, more technical users
- Additional meta-data (semantics)
- Additional data (volume)
- Additional dimensions (derived)
Lower level actions: Searching
- Lookup: You know what you look for and where
- Browse: You don't know what, but know where
- Locate: You know what, but not where
- Explore: You don't know what, nor where
Targets:
Things we want from the data.
- Look at trends, outliers and features (task-dependent structures of interest)
- Attributes to look for:
> One: Distribution -> Extremes
>Many: Dependency -> Correlation -> Similarity
Proximity
Objects colse to each other are preceived as a group
Similarity
Objects that are similar (color, shape...) are precived as a group
Continuity
Mind unconsciousely draws a line between points
Ordering Direction
- Sequential: XS < S < M < L < XL
- Diverging: -10...0...32 (can do calculations)
- Cyclic: Repeats every time
Key/independent attribute
Categorical or ordinal
- Usually not quantitative
Value/dependent Attribute
An index used to look up value attributes
Arrangments
How data values determine position and alignments of visual representations
Arrangments (Express)
Out it on the canvas, mapped to s
Arrangments (Seperate)
Emphasize similarity and distinction (categorical attribute)
Arrangment (Order)
Emphadsize ordered attribute
Arrangment (Align)
Emphasize quantitative comparison (quantitative attribute)
Arrangment (Use)
Using exisiting atructure for arrangment (e.g: map)
Munzner's Reference Model
Uses a stepwise approach to analyze and construct visualizations of data:
- Why? - What are the actions and target of the visualizations
- What? - What is the data and how is it structured
- How? - What is the mapping between data items and visual
elements or channels
What are the aspects to consider when vizualizing data?
1. What is the data and how is it structured?
2. Who is the user or recipient?
3. What should the user be able to do? Exploration,
confirmation or communication?
4. Why?
5. How?
Design Strategies
- Use postion to visualize color for categories
- Natural order: position, length, thickness, brightness and
saturation
- Use rebundant encoding
- Limit amount and detail of data in visualization
- Consider defult formats and mapping
Idiom
Indicates chart types
- Bar charts, scatter plots, histogram etc...
What are the idiom types?
1. Data - What is the data behind the chart?
2. Marks - Which marks are used?
3. Channels - How will the data be encoded in visual channels?
4. Tasks - What are the supported tasks?
Effectiveness
How well the visualization helps a person with their tasks:
- Indicate how values relate to one another
- Represents quantities accuratly
- Easy to compare qunatities
- Easy to see ranked order of values
- Make it obvious how people should use the information, what to use to accomplish and encourage them
Data-Ink
Ratio of ink on a graph that represents data
- Erase as much non-data-ink as possible
Chart Junk
Exessive and unnecessary use of graphical efforts in graphs.
- Careful with pies, donuts, 3D
>Insufficient task support (difficult to compare and read)
Ethical Considerations
- Scientific Integrity
- Data protection and privacy
- Misleading visuals (Make sure not to have)
- Persuasion