mao lin huang ( 黃茂林) university of technology, sydney,

35
Mao Lin Huang ( 黃黃黃 ) University of Technology, Sydney, Information Visualization and Its Applications 黃黃黃黃黃黃黃黃黃

Upload: gala

Post on 19-Mar-2016

109 views

Category:

Documents


7 download

DESCRIPTION

Information Visualization and Its Applications 信息可视化及其应用. Mao Lin Huang ( 黃茂林) University of Technology, Sydney,. Visualization ( 可视化). Scientific Visualization (科学计算可视化). Information Visualization (信息可视化). None Graph Visualization (非图论型可视化). Graph Visualization (图论型可视化). - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

Mao Lin Huang (黃茂林 )University of Technology, Sydney,

Information Visualization and Its Applications 信息可视化及其应用

Page 2: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

2

Visualization Visualization (( 可视化 )

Information Visualization( 信息可视化 )

Scientific Visualization( 科学计算可视化 )

None Graph Visualization( 非图论型可视化 )

Graph Visualization( 图论型可视化 )

Graph G = (V, E)

Page 3: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

3

Information VisualizationInformation Visualization

Maps, diagrams, and PERT charts Maps, diagrams, and PERT charts are examples of using visual are examples of using visual representations to see things. representations to see things. A A good picture is worth ten thousand good picture is worth ten thousand words. words.

一个好的图画所能表达的东西胜一个好的图画所能表达的东西胜过十万个字过十万个字Today, computers help people to Today, computers help people to see and understand abstract data see and understand abstract data through pictures.through pictures.

Advances in science & Advances in science & technology have allowed technology have allowed people to people to see old things in new see old things in new ways.ways. Telescopes, microscopes Telescopes, microscopes and oscilloscopes are typical and oscilloscopes are typical instrument examples.instrument examples.

现代科学技术允许人们用新现代科学技术允许人们用新的方法来看旧的事物的方法来看旧的事物 . . 例如例如天文望远镜天文望远镜 , , 显微镜 …显微镜 … ..

Page 4: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

4

The Definition (The Definition ( 定义定义 ))

Information visualization: the use of interactive visual Information visualization: the use of interactive visual representations of abstract, non-physically based data to representations of abstract, non-physically based data to amplify cognition [CMS99].amplify cognition [CMS99].

信息可视化信息可视化 : : 用可交流的用可交流的 , , 视觉表达方式来表达抽象的视觉表达方式来表达抽象的 , , 非物理非物理的的 数据从而增强对数据的认识数据从而增强对数据的认识 ..

[CMS99] Stuart K. Card, Jock D. Mackinlay, and Ben Shneiderman. Readings in information visualization: using vision to think. Morgan Kaufmann Publishers, Inc., 1999.

Xerox Palo Alto Research Center (PARC)

Page 5: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

5

Information VisualizationInformation Visualization

The little image dots represent data records of the number of sun spots, from 1850 to 1993, zoomed in on a small area. (collected from GVU Center, Georgia I. T.)

An example of using SeeNet to view email data volumes generated by AT&T long distance network traffic. Edges represent email connections. Weigh and colors of edges represent volumes of email data.

Graphics-driven & Graph-driven

None-relational data & Relational data

Page 6: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

6

Graph-Driven Visualization of Relational DataGraph-Driven Visualization of Relational Data

An example of visualizing relational data. This is the visualization of a family tree (graph). Here each image node represents a person and the edges represent relationships among these people in a large family.

Graph VisualizationGraph Visualization

Page 7: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

7

The Model of the Relational DataThe Model of the Relational Data

关联数据模型关联数据模型Relational information (graph) visualization systems use graphs G = (V, E) to model relational structures: the entities are nodes, and the relationships are edges (sometimes called links).

For example, the structure of the World Wide Web can be modeled as a graph: the nodes are HTML documents, and a hyperlink from one document to another is represented as a directed edge.

Page 8: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

8

Challenges in GV

Graph layout Problem Scale Problem Scope Problem Navigation Problem

Readability, cognitive effort comprehension

Small window, large data setsContext display

Distributed huge data sets, which are partially unknown How to find particular data

items by visual manipulation?

Page 9: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

9

Classical Graph Layouts Link-node diagrams Layout algorithms (graph drawing) Geometric positioning of nodes & edges Small amount of nodes Avoid node overlaps Reduce edge crossings

hierarchical force-directed orthogonal

symmetric

radial layout

Page 10: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

10

The Graph Drawing (The Graph Drawing ( 画图画图 ))

A graph drawn using the original spring algorithm.

一个用弾璜算法来画的图 .

A drawing D(G) of a graph G = (V, E) consists of a location (x, y) for each v in V and a route ((x1,y1), (x2, y2)) for each edge e in E.

一个“图画”是一个“图”的几何表达方式 .

Page 11: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

11

Classical Tree Layouts A special type of graphs With no circles Structured hierarchically Inefficient use of display space Small amount of nodes

Classical hierarchical layout hyperbolic tree

radial layout

balloon layout

radial layout

Page 12: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

12

Space-Optimized Tree Layout Redefine the wedge (v) Efficient use of display space Connection+enclosure Large amount of nodes How to navigate? (Distortion, Zooming+Filtering

context+detail)

A large data set of approximately 50 000 nodes My Unix root with approx. 3700 directories and files

(We introduced this new method in 2001)

Page 13: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

13

Redefine the wedge(v) of radial drawings

-A region P() of a node is defined by the wedge wg(v) and one (or more) cutting edges (boundaries) cut by other regions that cross the line l in wg(v). -A wedge is defined as wg() = {, l, ()}, where is the father of and l is a straight line going through that determines two boundaries of P().

Page 14: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

14

The Problem of Viewing Large DataThe Problem of Viewing Large Data大型数据可视化的问题大型数据可视化的问题Traditional graph visualization assumes that the whole graph can reasonably be represented in a readable and understandable manner on the display medium.

• Amount of information we want to visualize becomes larger• A small modern file system (say with a 2GB drive on a PC) there are hundreds of nodes and links• Web graphs are much larger; even a small organization such as a University has many thousands of web documents.• No techniques can visualize the complete World Wide Web.

Classical visualization methods tend to be inadequate.

Page 15: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

15

Clustered Layout (nodes grouping)

Connection + Enclosure

Page 16: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

16

Clustered Layout High scalability Dynamic viewing Abridged context Open & close clusters Average readability

Page 17: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

17

Force-Directed Clustering (DA-TU)(we introduced this method in 2002 and it’s the first attempt of using force-directed method to draw clustered graphs)

Multi-level clustering Multiple spring forces High scalability Dynamic viewing Abridged context Open & close clusters Good readability

Page 18: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

18

Multi-Forces in Clustered Graphs

Page 19: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

19

The Online Graph Model 在线图模型 Proposed in 1997 Scope problem

is well addressed Scalability is

increased through dynamic viewing

Lost overall context Change frames Animation preserves

“mental map” Modified Spring Algorithm

Page 20: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

20

We proposed OFDAV in 1997 that provides a major departure from traditional methods. We visualise a tiny part (a “frame” Fi ) of a huge graph at time t. We change from Fi to Fi+1 by user interaction.

Online Navigational Visualization(Online Force-Directed Animated Visualization)

Page 21: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

21

Online Navigational VisualisationOnline Navigational Visualisation在线导航型可视化在线导航型可视化OFDAV provides a major departure from traditional methods. We visualise a tiny part (a “frame” Fi ) of a huge graph at time t. We change from Fi to Fi+1 by user interaction.

OFDAV does notneed to know the whole graph, it does not predefine the geometry (the user can navigate logically), and it is user-oriented.

Page 22: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

22

We incrementally calculate and maintain a small local visualization on-line. The graph is supplied to the system by a series of requests for neighbourhoods of focus nodes.

Hugegraph

new focus node v

neighbourhood of v

Small local graph

Scope Problem is Addressed

Page 23: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

23

The specific criteria for online drawing: The layout of logical frame must show the direction of the

exploration. Reduce the overlaps among the local regions. The sequence of drawing preserves the mental map.

The general criteria for graph drawing: Reduce the edge crossings. Avoid nodes overlaps

Online Graph Layout Problem

Page 24: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

24

In the spring model, each node is replaced by a steel ring, and edges are replaced by Hookes’s law springs ( 胡克定理 ). The rings have a gravitational repulsion ( 反向万有引力定理 ) acting between them, and we can find a drawing which minimizes the energy.

Spring model (Spring model ( 弹鐄模型弹鐄模型))

Page 25: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

25

In this frame, there are two focus nodes, x and y. The total force on node v is:

},{},,4,...,1{

)(yxu

uv

yxu

uvxv hgfvf

Modified Spring Algorithm (MSA)

Page 26: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

26

The layout of The layout of FFii must show the direction must show the direction of the exploration.of the exploration.

Spring model Modified spring model

Page 27: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

27

Application1:Application1:Visual Web browserVisual Web browser

WebOFDAV - mapping the entire Web,

Look at the whole of WWW as one graph; a huge and partially unknown graph.

Maintain and display a subset of this huge graph incrementally.

Reduce mouse-click rate

Maintain a 2D map & history of navigation

Page 28: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

28

The “lost in hyperspace” problem 迷失在超空间 Even in this small document, which could be read in

one hour, users experienced the ‘lost in hyperspace’ phenomenon as exemplified by the following user comment: ‘ I soon realized that if I did not read something when I stumbled across it, then I would not be able to find it later.’ Of the respondents, 56% agreed fully or partly with the statement, ‘When reading the report, I was often confused about where I was.’ [Nielson, 1990].

Page 29: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

29

Visual Web Browser addresses the problem of Visual Web Browser addresses the problem of “lost in hyperspace” with a sense of “space”.“lost in hyperspace” with a sense of “space”. Graphic Web Browser addresses the fundamental problem

of “lost in hyperspace” by displaying a sequence of logical visual frames with a graphic “history tail” to track the user’s current location and keep records of his previous locations in the huge information space.

The logical neighborhood of the focus nodes indicates the current location of the user, and the tail of history indicates the path of the past locations during the navigation.

Page 30: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

30

Application2:Application2:File ManagementFile Managementand Site Mappingand Site Mapping

An example of using Space-Optimized Tree Visualization for a small web site mapping (approximately 80 pages)- viewing techniques neededMapping to a Unix root with

approx. 3700 directories and files

Page 31: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

31

Application3: Application3: Web Reverse EngineeringWeb Reverse Engineering HWIT (Human Web Interface Tool) is able to reuse existing structures of web site by

visualizing and modifying the corresponding web graphs, and then re-generating a new site by save the modified web graphs.

The layout of an existing structure of a web site Enhancing the existing Web site by adding a sub-site

Page 32: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

32

Application4: Application4: B2C e-CommerceB2C e-Commerce VOS (Visual Online Shop) can be used for online grocery shopping,

shopping cart model. It is applicable to any e-commerce shopping application (dynamically navigate e-catalogs).

Page 33: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

33

Application5: Application5: Online Business Online Business Process ManagementProcess Management

WbIVC (Web-based Interactive Visual Component) is applied to a research project management system (RPMS) in universities.

A participant can review the details of a specific process element by clicking on the corresponding rectangle, and then selecting the “open a process element” in the popup menu.

A participant can also create a new artifact (a Java methods) to a research project by opening a edit window.

The output interface of the WbIVC in RPMS

The input interface of the WbIVC in RPMS

Page 34: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

34

Application6: Application6: Program UnderstandingProgram Understandingand Software Miningand Software Mining

JavaMiner is for non-linear visual browsing of huge java code for programming understanding.

textual data mining Visualize a variety of

relationships between terms in Java code, e.g. HAS, SUBCLASS, CALL and INTERFACE relationships.

Text documents, the lexicon, the neighborhood function

The input interface of the WbIVC in RPMS

Page 35: Mao Lin Huang  ( 黃茂林) University of Technology, Sydney,

35

Conclusion Conclusion The above talk gives an introduction to Information

Visualization and covers most of research works I have done in this field.

In the future, I will remain doing research across these three layers because I believe that to guarantee the quality of the research, it needs theory but to assess the value of the research, it needs applications.

Thank You !Thank You !