Lecture 2 Flashcards
What(Data)-Why(task)-How(idiom)
What Data does the user see; (Data)
Why does the user intend to use the visualization tool; (Task)
How are the visual encoding constructed in terms of design choices (Idiom)
We have four looped levels of visualization design
The first level is domain
The second level is abstraction
The third level is idiom
The fourth level is algorithm
Domain in visualization refers to to a particular field of interest like e-commerce , education,
True
Group target people can be scientific research people , public people , specific group of people
True
User centered design or human centered design
We are doing four nested levels of design
Working with a specific target audience to iteratively refine a design
The outcome of the design process is to ensure that the designer reaches the needs of the user
true
Data/ task abstraction
Abstracting tasks and data from the domain
tasks: I want to compare, visualize, summarize
Data: data needed (raw or processed)
Many visualization idioms (اساليب) are specific to a particular data type
true
like tables are specific for quantities and number
Human abilities such as perception and memory need to be taken into account when we are doing the visual encoding/interaction idiom (2sloob) in the four looped visual design
true
Visual encoding/interaction idiom is choosing a specific way to create and manipulate the visual representation of the abstracted data and tasks
True
Visual encoding/ interaction idiom has two main concerns
visual encoding (what users see) interaction (how users change and what they see)
Algorithms : The goal at this level is to efficiently handle the visual encoding and interaction idioms
true
Features that promote and algorithm over other are :
Computational speed
How much computer memory is required
True
Approaches we can have
Problem driven approach (top down):
Start by domain situation then move to data /task abstraction then move to visual encoding /interaction idiom then move to algorithm
Technique driven approach (down top):
We start at algorithm level or idiom level in order to create better idioms or algorithm to better support existing abstraction
Slide 25 see them and see how to validate first and after
true
There are three major dataset types
1- Tables
2- Networks
3- Spatial
In tables, the attributes (properties) are columns, the items are rows, and the cell contains the value
True
In networks we have nodes and links
True
In spatial we have fields and gematry
True
Semantics of data is the underlying meaning of data
True
The two main aspects of data are the semantics of the data and the type of data
True
metadata
data about data
Five main data types
Item Attribute Link Grid Position
Item
Discrete individual entity (row in a simple table or a node in a network)
Attribute
also called variable or dimension ( specific property that can be measured, observed ,or logged)
Link
relationship between two items, typically an edge in a network
Grid
strategy for sampling continuous terms of geometric and topological relationships between its cells
Position
spatial data (location in 2D or 3D space)
Review from 40 to 53
True
Dataset Availability
1- Static (The entire dataset is available all at one to visualize)
2- Dynamic (The dataset information change over the course of the visualization process)
There are two attributes types
Categorical ( small Box, large box)
Ordered (Ordinal and Quantitative)
Key attribute
index that could be use to look up value attributes ( like weight is a key , 28 kg is an attribute, ID is a key )
Multidimensional tables have multiple keys
true
Unlike tables, fields contain continuous rather than discrete data
true
Each cell in a field refer to a unique range of continuous domain
true
n contrast with tables, attribute values in
fields are returned for locations throughout the sampled range and not just the exact points where data was recorde
True
In context of fields, independent variables refers to key and dependent variable refer to value.
true
We have 3 types of fields
Scalar fields
Vector fields
Tensor fields
Properties of scalar fields
-Univariate (single value attribute at each point in space.
If no connection between points in space, then we will have multiple separate scalar fields.
Properties of vector fields
Multivariate, with multiple attributes at each point.
Each point has a direction and a magnitude.
Tensor fields
array of attributes at each point
A dataset is said to have time-varying semantics when one of its “key” attributes is time
true
Data abstraction operation
slide 71 -75
