Quiz 2 Flashcards
Characteristics of Time
scale
ordinal (A before B before C)
discrete (points)
continuous (line)
Characteristics of Time
scope
point based (most common)
interval based
Characteristics of Time
arrangement
linear
cyclic
Characteristics of Time
viewpoint
ordered
branching
multiple perspectives
Characteristics of Time-Oriented Data
scale
quantitative (numbers)
qualitative (words)
Characteristics of Time-Oriented Data
frame of reference
abstract
spatial
Characteristics of Time-Oriented Data
kind of data
events
states
Characteristics of Time-Oriented Data
number of variables
univariate
multivariate
Mapping of time
- mapping time to space: static visualization, time and data in single coherent representation
- mapping time to time: dynamic representation, utilise physical dimension to convey time dependency of data
Categorisation on TimeViz Browser
data
- frame of reference: abstract vs. spatial
- variables: univariate vs. multivariate
Categorisation on TimeViz Browser
time
- arrangement: linear vs. cyclic
- time primitives: instant vs. interval
Categorisation on TimeViz Browser
vis
- mapping: static vs. dynamic
- dimensionality: 2D vs. 3D
Geospatial data
- describes objects with specific location in real world
- map spatial attributes to the two physical screen dimensions resulting in map visualizations
- map: world reduced to points, lines, and areas
Spatial phenomena
- point phenomena
- line phenomena: have length, but no width
- area phenomena: have both length and width
- surface phenomena: have length, width and height
Maps subdivided into:
Map types based on:
Properties of data:
- qualitative vs. quantitative
- discrete vs. continuous
Properties of graphical variables:
- points
- lines
- surface
- volumes
Map projections
mapping the positions on the globe (sphere) to positions on the flat surface
longitude: negative = western degrees
latitude: negative = southern
Cylinder projections
- preserve local angles
- conformal projections
- degrees of longitude and latitude usually orthogonal to each other
Plane projections
- azimuthal projections
- map to a plane that is tangent to the sphere with tangent point corresponding to the center point of projection
- some are true perspective projections
Cone projections
- map to cone that is tangent to the sphere
- degrees of latitude represented as circles around the projection center
degrees of longitude as straight lines emanating from center - designed to preserve distance from center to the cone
Point data
- discrete in nature, buy may describe continuous phenomenon
- discrete: occur at distinct locations
- continuous: defined at all locations
- smooth: data that change in gradual fashion
- abrupt: change suddenly
Dot maps
- place symbol at location
- quantitative parameter mapped to the size or color of the symbol
- problem of overlap or over plotting in highly populated areas
approaches for dense spatial data
- 2.5D visualization showing data points aggregated up to map regions
- individual data points as bars, according to their statistical value on a map
visualization of line data
- represent as line segments between pairs of endpoints specified by longitude and latitude
mapping of line data
- standard: data parameters mapped to line width, line pattern, line color, and line labelling
- start, end, and intersection points can be mapped to visual parameters (size, shape, color, labelling)
- lines do not need to be straight
Choropleth maps
values of an attribute or statistical variable are encoded as coloured or shaded regions on the map
Choropleth maps: Issues
- most interesting values often concentrated in densely populated areas with small polygons
- assumption that in one region values are uniformly distributed
- size of regions impact perception of importance (color)
- people can be color blind
Dasymetric maps
- variable to be shown forms areas independent of original regions
- attribute has different distribution than the partitioning
- ancillary information acquired -> cartographer steps statistical data according to extra information collected within boundary
Filters
Context filters before dimension filters
Context filters -> dimension filters -> measure filters
Data blending
- combines data from multiple data sources into single view
- sends separate queries to the separate data sources and aggregates results to common level in Tableu
-> usually join data at row level
The query pipeline
- Extract filters
- Data Source Filters
- Context Filters
-> Top N, Fixed LOD - Dimension Filters
-> Include/Exclude LOD, aggr, Data Blending - Measure Filters
-> Table Calcs, Clustering - Table Calc Filters
-> Trend lines, reference lines, pages
Elements for designing a dashboard
- identify audience
- define analytical questions
- built data viz that will be components of the dashboard
- design layout
- design useful tooltips
- define interactions
- experiment and refine
Isarithmic maps
- shows the contours of some continuous phenomena
-> isometric: if contours determined from real data points (e.g. temperatures)
-> isopleth: if data measured for a certain region and centroid considered as data point
main task: interpolation of data to obtain smooth contours (e.g. triangulation, inverse distance mapping)
Cartograms
- generalisation of ordinary thematic maps
- distorting geography: size of regions scaled to reflect statistical variable
- can’t fully satisfy shape or area objective -> search compromise between shape and area preservation
Noncontinuous cartograms
- keeps map, but fills correct size inside of the map
- doesn’t preserve input map’s topology
- scaled polygons drawn inside the original regions
Noncontiguous cartograms
- scale polygons to their target size
- satisfy area objective
- perfect are adjustment
- lose map’s global shape and topology
Circular cartograms
- completely ignore input’s polygon’s shape
- each polygon is a circle
- lose map’s global shape and topology
Continuous cartograms
- retains map’s topology perfectly
- not use area and shape constraints
Issues for spatial data mapping
- class separation (bins)
- normalisation (absolute vs. relative),
- spatial aggregation (area definition)
has severe impacts on visualization result
Map generalization
- process of selecting and abstracting information on a map
- used when small-scale map is derived from large-scale map
- application- and task-dependent
-> emphasise elements that are most important for task
Why use interactions?
- static visualisations belong to the past
- static fail to answer unpredicted questions
- interactions help users ask their own questions and answer them while exploring data
Types of interactivity
is a mechanism for modifying WHAT users see and HOW they see it
- Find relevant data (in sea of other data)
- reveal more data (drill down to enter details)
- change views or context (to better answer questions)
Interaction operators
navigation
altering position of the camera and for scaling the view
e.g. panning, rotating, zooming
Interaction operators
selection
- identify an object, collection of objects, or regions of interests
e.g. highlighting and modifying
Interaction operators
filtering
reduce size of data being mapped on the screen
e.g. eliminating records, dimensions, or both
Interaction operators
reconfiguring
- changing way data is mapped to graphical entities or attributes
e.g. reordering the data or layouts -> diff. way of viewing a data subset
Interaction operators
encoding
- changing graphical attributes to reveal different features
e.g. point size, line color
Interaction operators
connecting
linking different views or objects to show related items
Interaction operators
abstracting/elaborating
modifying the level of detail
Interaction operators
hybrid
- combining several of the above in one technique
Operand
space upon which operator is applied
- Screen space (pixels)
- Data Value space
- Data Structure Space
- Attribute Space
- Object space
- Visualization Structure space
Main interactivity tools in Tableau
- selection
- highlight
- filtering
- parameters
- sets
- tooltips
- URL actions
Highlighting suitable for
- finding data of interest in the same context
- show other marks that share attributes
- find data on another sheet that is related
Filtering suitable for
- focus only on the data we want to analyze, reducing cognitive load
- control the context of data
- remove unnecessary data and show only relevant one
Parameters used to
- explore what-if scenarios
- customize view
- make dashboard more flexible
- since defined globally, can be used across all data sources
Sets and way of creating sets
- are custom fields that define a subset of data based on some conditions
Ways of creating sets:
- from marks in a vis
- from a computation
- combining sets
Different types of sorting data
- sorting on axis
- sort by labels
- sorting by pill and toolbar sort button
- custom sort
- clear sorts
Default sort
All variables take on their default sort
Comparative sort
When variables are sorted by e.g. SUM(Shipping Cost)
sorted by the highest value of SUM
Nested sort
Each sub-category is independently sorted, e.g. after highest value
Manual sort
Variables were manually sorted
