Multiscale Visualization Using Data Cubes

1
Multiscale Visualization Using Data Cubes

• Chris Stolte, Diane Tang, Pat Hanrahan
• Stanford University
• Information Visualization
• October 2002
• Boston, MA

2
Motivation

• Large multidimensional databases have become very
  common
• Need techniques for exploration and analysis
• Overview first, zoom and filter, then
  details-on-demand

3
Multiscale Visualization

• Visual representation changes as user pans and
  zooms
• Overview, lots of data ? highly abstracted
• Zoom, data density decreases ? detailed
  information shown
• Visual and data abstraction
• Visual abstraction different representation/same
  data
• Data abstractiontransformations to reduce data
  set size

4
Existing Multiscale Visualizations

• Cartography
• Multiscale information visualization
• Pad alternate desktops
• DataSplash
• XmdvTool
• ADVIZOR
• Main limitations
• One zoom path
• Primarily visual abstraction

5
Contributions

• Multiscale visualization with both visual and
  data abstraction using generalized mechanisms
• Data Abstraction ? Data Cubes
• Visual Abstraction ? Polaris
• Design Patterns

6
Data Cubes
7
Data Warehouses

• Store data for analysis (OLAP)
• Fact table contains measures categorized by
  dimensions

Fact table
State Month Product Name Profit Sales Payroll Mar
keting Inventory Margin ...
Ordinal fields (categorical dimensions)
Coffee chain (courtesy Visual Insights)
Quantitative fields (measures)
8
Hierarchical Structure

• Data warehouses are very largeneed to summarize
• Add hierarchical structure to warehouse

Dimension tables
Fact table
Time Year Quarter Month
Location Market State
State Month Product Name Profit Sales Payroll Mar
keting Inventory Margin ...
Products Product Type Product Name
9
Hierarchical Dimensions

• Each dimension table describes a tree
• Each level describes a level-of-detail
• Meaningful basis for aggregation
• Create summaries of fact tablefor each
  level-of-detail asData Cubes

Time Year Quarter Month
10
Data Cube

• Create cube for each level-of-detail combination
• Summary of fact table

Cube for (Quarter, Product Type, Market)
Each cell aggregatesall measures for those
dimensions
Each cube axis corresponds to a dimension in the
relation at a level-of-detail
11
Hierarchies Data Cubes

• Hierarchies define a lattice of cubes

Least detailed
Each cube is defined by a level-of-detail in each
dimension
Data abstraction
Most detailed
12
Projecting Data Cubes

• Can further abstract a cube by projection

Data abstraction
13
Data Cube Summary

• Industry standard for storing analytic data
• Provide summaries of data at meaningful levels of
  detail
• To perform data abstraction
• Design a hierarchical schema
• Choose a cube in the lattice of cubes
• Project to relevant dimensions
• Identifying a projection corresponds to
  specifying the desired data abstraction

14
Polaris
15
Exploring Data Cubes using Polaris

• Polaris is
• A UI for exploration, analysis of data warehouses
• A formal language for specifying queries
  visualizations
• An interpreter for compiling specification into
  queries/drawing commands
• Demo!

16
Polaris Formalism

• Visualization described using visual
  specifications that define
• Table configuration (algebra)
• Type of graphic in each pane
• Encoding of data as visual properties of marks
  (encoding system)
• Data transformations and queries
• Each specification corresponds to a projection of
  the data cube

17
Path of Exploration

• Can think of an analysis as path of specifications

18
Path of Exploration
Visual abstraction
19
Path of Exploration
This is a multiscale visualization!
Dataabstraction
20
Graphical Notation
21
Graphical Notation Templates
Instance
Template
22
Specifying Multiscale Visualizations

• Specify multiscale visualization using a graph of
  Polaris specifications a Zoom Graph
• Paper describes how to implement using Polaris

?Polaris Specification
Zooming
?Possible zoom
23
Specifying Multiscale Visualizations

• Can specify a zooming pattern by using templates

24
Specifying Multiscale Visualizations

• Independent zooming on different dimensions is
  described as a graph

y-axis zoom
x-axis zoom
25
Design Patterns
26
Design Patterns

• Zoom graphs simplify specifying and implementing
  multiscale visualizations
• Design is still very hard
• Design patterns (a la Gamma et al.)
• Capture zoom structures that have been used
  effectively reuse in new designs
• We present four such patterns
• Formal way to discuss multiscale visualization

27
Thematic Maps
28
Thematic Maps
29
Thematic Maps
30
Thematic Maps
31
Chart Stacks
32
Chart Stacks
33
Chart Stacks
34
Chart Stacks
35
Matrices
36
Matrices
37
Matrices
38
Matrices
39
Dependent QQ Plots
40
Summary
41
Summary

• Multiscale visualization with both visual and
  data abstraction using generalized mechanisms
• Data Abstraction ? Data Cubes
• Visual Abstraction ? Polaris
• Zoom Graphs for specifying and implementing
  multiscale visualizations
• Design Patterns

42
Future Work

• Designing new patterns
• Transitions between levels-of-detail
• Communicate parent-child relationships
• Non-uniform branching
• Animation/dissolve/fade?
• Data management
• Prefetching and caching of large data sets