CSE 578 Midterm 1 with complete updated
questions & answers
Give 5 reasons for visualizing data? - answer - Analyze data to support
reasoning
- Answer questions
- Communicate complex ideas to others
- Confirm a hypothesis
- Present findings
- Explore a dataset
- Reveal patterns
- Understand trends
- Find outliers
Visualization (for data) - answer Computer-based visualization systems
provide visual representations of datasets designed to help people carry
out tasks more effectively.
When is visualization the most suitable? - answer When we need to
augment human capabilities rather than replace them with purely
computational methods
,Give five examples of when data visualization is the most suitable? -
answer - No trusted automatic (pure computation) solution exists
- Don't know the question to ask (unfamiliar problem space)
- Exploratory data analysis (humans in-loop)
- Presenting data to others
- Stepping stone to automation: 1. Before model creation, 2. During
development, 3. During deployment
Why use external representation? - answer External data representation
helps replace cognition (thinking) with perception (first seeing and using
that information for cognition).
What are the advantages of vision in data representation? - answer -
Human visual system is a high-bandwidth channel to the brain while
other senses are lower bandwidth with different semantics (sound =
sequential stream only)
- See everything simultaneously and processing occurs in parallel and
pre-attentively
Why is data represented visually? - answer - Summaries lose
information, details matter (signal vs. noise)
- Helps confirm what is expected and what is unexpected
- Assesses the validity of statistical models
,Anscombe's Quartet - answer Illustrates the importance of looking at a
set of data graphically and not only relying on basic statistic properties
Explain why Effectiveness requires matching the visual representation
to the data/task? - answer - Set of all possible representations is huge
- Many representations are ineffective for the specific data/task
- There is a better chance of finding an effective visualization if you
understand the full space of possibilities
What representations are effective? - answer - Novel: enables new
kinds of analysis
- Efficiency: Speeds up existing workflow
- Correct: Represents the actual interpretable data
How do you validate effectiveness? - answer Many methods; must pick
the appropriate one for your context!
What were some of the earliest data visualization examples? - answer
Maps/cartography
William Playfair (1759-1823) - answer Scottish architect, economist, and
pioneer of statistical graphics in the late 1700s; credited with
"inventing" line, bar, and pie charts.
, Jon Snow's cholera map (1854) - answer
Limits of Visualization? - answer - Computational: Processing Time,
System memory
- Human: Attention, memory, cognition, perception
- Display: Pixels, Information density
Information density - answer Ratio of space used to encode information
vs. unused whitespace.
Why does visualization work? - answer - Because humans have limited
memory and cognition (Change blindness, working memory)
- Power of perception to reveal (Vs example, Find red circle)
Why analyze visualizations? - answer To impose structure on a huge
design space:
- Think systematically about encoding/interaction choices
- Analyze existing idioms as stepping stones to designing new ones
- Most possibilities are ineffective for particular data/task combos
questions & answers
Give 5 reasons for visualizing data? - answer - Analyze data to support
reasoning
- Answer questions
- Communicate complex ideas to others
- Confirm a hypothesis
- Present findings
- Explore a dataset
- Reveal patterns
- Understand trends
- Find outliers
Visualization (for data) - answer Computer-based visualization systems
provide visual representations of datasets designed to help people carry
out tasks more effectively.
When is visualization the most suitable? - answer When we need to
augment human capabilities rather than replace them with purely
computational methods
,Give five examples of when data visualization is the most suitable? -
answer - No trusted automatic (pure computation) solution exists
- Don't know the question to ask (unfamiliar problem space)
- Exploratory data analysis (humans in-loop)
- Presenting data to others
- Stepping stone to automation: 1. Before model creation, 2. During
development, 3. During deployment
Why use external representation? - answer External data representation
helps replace cognition (thinking) with perception (first seeing and using
that information for cognition).
What are the advantages of vision in data representation? - answer -
Human visual system is a high-bandwidth channel to the brain while
other senses are lower bandwidth with different semantics (sound =
sequential stream only)
- See everything simultaneously and processing occurs in parallel and
pre-attentively
Why is data represented visually? - answer - Summaries lose
information, details matter (signal vs. noise)
- Helps confirm what is expected and what is unexpected
- Assesses the validity of statistical models
,Anscombe's Quartet - answer Illustrates the importance of looking at a
set of data graphically and not only relying on basic statistic properties
Explain why Effectiveness requires matching the visual representation
to the data/task? - answer - Set of all possible representations is huge
- Many representations are ineffective for the specific data/task
- There is a better chance of finding an effective visualization if you
understand the full space of possibilities
What representations are effective? - answer - Novel: enables new
kinds of analysis
- Efficiency: Speeds up existing workflow
- Correct: Represents the actual interpretable data
How do you validate effectiveness? - answer Many methods; must pick
the appropriate one for your context!
What were some of the earliest data visualization examples? - answer
Maps/cartography
William Playfair (1759-1823) - answer Scottish architect, economist, and
pioneer of statistical graphics in the late 1700s; credited with
"inventing" line, bar, and pie charts.
, Jon Snow's cholera map (1854) - answer
Limits of Visualization? - answer - Computational: Processing Time,
System memory
- Human: Attention, memory, cognition, perception
- Display: Pixels, Information density
Information density - answer Ratio of space used to encode information
vs. unused whitespace.
Why does visualization work? - answer - Because humans have limited
memory and cognition (Change blindness, working memory)
- Power of perception to reveal (Vs example, Find red circle)
Why analyze visualizations? - answer To impose structure on a huge
design space:
- Think systematically about encoding/interaction choices
- Analyze existing idioms as stepping stones to designing new ones
- Most possibilities are ineffective for particular data/task combos
