Table Lens - Personal Web Pages

Download Report

Transcript Table Lens - Personal Web Pages

Table Lens
From papers 1 and 2
By Tichomir Tenev, Ramana Rao, and
Stuart K. Card
Overview



Uses focus+Context apporach
Context elements are represented graphically
Focus elements have text and graphic
display
Advantages



Increases viewable portion of table by 100
times
Ease of Navigation
Ease of Exploration
Table Lens Focal Technique


Mutates layout of table
Does not bend any rows or columns
Distortion Function Framework


DOI function: item -> value. Value indicates
level of interest
DOI function controls how available space is
allocated among items
DOI in Table lens


DOI maps cell address to interest level
2 of them, one for each dimension
Manipulation of Focus Operations



Zoom- changes amount of space to focal
area
Adjust- changes amount of contents viewed
within focus area
Slide- changes location of focus area within
the context
User manipulation



Clicking at upper left corner- zooms all cells
Touching any region in context will slide
current focus to that location
Grasping focus slides focus to that location
Results


Apply data to baseball stats of 323 rows by
23 columns (7429 cells)
Display whole table on screen at one time
Paper #2 Design 1 Nesting Focal Levels




Space allocated to each element is
dependent on the focal level of element
2 foci, Primary focus always inside region of
secondary focus
2ndary focus used for coarse navigation
Primary used for finer navigation
Design II Controlling focal spans



Space allocated per data element dependent
on focus level and parameter specified by
user
Primary focus elements may vary in size
Spatial map at any time depends on History
of user interaction
Conclusion

Felt design 2 was the better design.
Disadvantage


Works only for data tables which have have
<= number of entries as pixel rows and each
column has enough pixels wide to
accommodate variables.
Paper #2 discusses how to improve it
Polaris: A System for Query,
Analysis and Visualization of
Multi-dimensional Relational
Databases
Chris Stolte and Pat Hanrahan
Standford University
Polaris



Interactive exploration of large multidimensional databases
Expressive set of graphical displays
Uses tables to organize multiple graphs on a
display
Relational databases



Each row in table = basic entity (tuple)
Each column represents a field
Fields can be ordinal, or quantitative
Visual Specification


Is the configuration of the fields of the tables
on shelves
User does this by dragging and dropping
fields onto shelves
Visual Specification






Mapping of data sources to layers
# of rows, columns, and layers, and relative
order
Selection of tuples from the database
Grouping of data within a pane
Type of graphic displayed in each pane
Mapping of data fields with retinal properties
Table Algebra



Used to specify table configurations.
Dragging and dropping implicitly does it
Operands are the names of the ordinal and
quantitative fields of database
Operators (concatenation, cross, nest)
Types of Graphics (Ordinal- Ordinal)

Axis variables are independent of each other
R represents the fields encoded in the retinal properties of the marks
Following slide shows sales and margin as a function of product type,
month and state for items sold by coffee chain
Ordinal-Quantitative Graphics


Bar charts, dot plots, Gantt chart
Quantitative variable is dependent of ordinal
variable
Figure 6c shows a case where a matrix of bar charts is used to study
several functions of the independent variables product and month
Quantitative-Quantitative Graphics

Discover causal relationships between the
two quantitative variables.
Figure 3e shows how flight scheduling varies with the region of the
country the flight originated.
Visual mappings



Encoding different fields of the data to retinal
properties
Shape, Size, Orientation, Color
Used in the ordinal to ordinal example
Generating Database Queries

1. Selecting the Records
Generating Database Queries


2. Partitioning the records into pains
Putting retrieved records in their corresponding
pane
Generating Database Queries


3. Transforming Records within the Panes
If aggregation, it is done here
Results



Cut expenses for a national coffee store
Create table of scatterplots showing
relationship between marketing costs and
profit (Figure 6a)
Notice trend; certain products have high
marketing costs with no or little profit
Results


Used linked displays to determine that in New
York several products are offering very little
return despite high costs
Creates bar chart for products in New York
Future Work


Exploring interaction techniques for
navigating hierarchical structures of mulit-dim
databases
Use selected mark in one display as the data
input to another