Hierarchy Hierarchy VisualizationVisualization

ByBy

Christian ChitaChristian Chita

Papers Surveyed:Papers Surveyed:

1. Cone Trees: Animated 3D Visualizations of Hierarchical Information. George G. Robertson, Jock D. Mackinlay, Stuart K. Card, SIGCHI 1991

2. Multitrees: Enriching and reusing hierarchical structures. George W. Furnas and Jeff Zacks, SIGCHI 1994 , pp 330-336.

3. Animated Visualization of Multiple Intersecting Hierarchies. George G. Robertson, Kim Cameron, Mary Czerwinski, and Daniel Robbins. Information Visualization, 1(1), p.50-65, 2002

Table of Contents:Table of Contents:1. ForEach (Paper) Do {

a) Problem Addressed and Knowledge Gapb) Key Issuesc) Implementationd) The Goode) The Not So Good}

2. Synthesis: ForEach (Paper) Do {a) Assumptions behind eachb) Did they solve the problem?}

I. ConeTrees paper:I. ConeTrees paper:

►Problem Problem AddressedAddressed:: ManagingManaging and and

AccessingAccessing large large information spacesinformation spaces

Once information is Once information is displayed, how are displayed, how are the various parts the various parts related? related?

►Knowledge GapKnowledge Gap:: Cognitive load of Cognitive load of

understanding the understanding the displayed displayed structure is not structure is not addressedaddressed

How to alleviate How to alleviate the currently the currently highhigh cognitive loadcognitive load ??

I. ConeTrees paper I. ConeTrees paper Key Key IssuesIssues::

► IssueIssue:: 2D layouts of complex 2D layouts of complex

structures will not fit structures will not fit onto the screenonto the screen

2D invariably leads to 2D invariably leads to scrolling/zoom-outscrolling/zoom-out

► SolutionSolution:: Use 3DUse 3D Use animation to Use animation to

reduce Cognitive Load reduce Cognitive Load

2D: Branching

Factor == 3

2D: Branching

Factor == 2

Levels Displayed Correctly

Aspect Ratio == (widthOfBase)/

(numLevels)

3D_ConeTree: fixed to fit

room

I. ConeTrees paper: I. ConeTrees paper: ImplementationImplementation

►Uses an Uses an Information Information

VisualizerVisualizer as engine as engine Supports: Supports:

►Continuous Continuous rotation for rotation for structure analysisstructure analysis

►Smooth interactive Smooth interactive animationanimation

►Mechanisms for 3D Mechanisms for 3D navigationnavigation

I. ConeTrees paper: I. ConeTrees paper: The GoodThe Good►No need for special equipmentNo need for special equipment►Fish eye view by defaultFish eye view by default►Shadow provides added structure info Shadow provides added structure info

without the user even without the user even noticing/focusing itnoticing/focusing it

►Prune and Grow opsPrune and Grow ops►Search handled by other process Search handled by other process

(allows user to continue work)(allows user to continue work)►Bottom line: get all of the above Bottom line: get all of the above ++

reduction in cognitive loadreduction in cognitive load

I. I. ConeTrees paper: ConeTrees paper: The Not So The Not So GoodGood

► Criticizes previous work, but Criticizes previous work, but input data is input data is different in the two cases different in the two cases they solve they solve different problemsdifferent problems

► Questionable structure-segment partition: Questionable structure-segment partition:

too much focus on too much focus on symmetrysymmetry:: Is this what the users want?Is this what the users want?

► Contradictions with self(?):Contradictions with self(?):

User is allowed to continue workUser is allowed to continue work BUTBUT“when a search starts, all nodes are made “when a search starts, all nodes are made invisible”invisible”

► Problem AddressedProblem Addressed:: DAGs have DAGs have

shortcomings:shortcomings:► Edge crossing even Edge crossing even

for small for small neighbourhoodsneighbourhoods

II. MultiTrees Paper:II. MultiTrees Paper:► Problem AddressedProblem Addressed::

Common trees have Common trees have shortcomings:shortcomings:

► Only one way to go Only one way to go from node_A to from node_A to node_Bnode_B

► No multiple No multiple organizing contextsorganizing contexts

II. MultiTrees Paper:II. MultiTrees Paper:

► Problem AddressedProblem Addressed:: Hierarchical Hierarchical

structure structure aggregate aggregate scalescale

► Problem AddressedProblem Addressed:: Hierarchical Hierarchical

structure structure reusereuse► Current approach not Current approach not

satisfactorysatisfactory

Tree Unordered CollectionMonolith Structure

II. MultiTrees Paper: II. MultiTrees Paper: Knowledge Knowledge GapGap

Tree DAG

Need a new type of structure to represent info: a multitreeA MultiTree ISA

hierarchy with shared subtrees

Professor ALibOfCongres

s

II. MultiTrees Paper: II. MultiTrees Paper: Key IssueKey Issue

► Focused on following Focused on following facts:facts: From any node:From any node:

► ifif((lookUPlookUP))seesee (diverse (diverse hierarchical hierarchical contextcontext –– a tree of a tree of contexts contexts))

► ifif((lookDownlookDown))see (see (contentcontent under a under a node node –– a tree ofa tree of contentscontents))

Professor B

LibOfCongress

II. II. MultiTrees paper: MultiTrees paper: ImplementationImplementation

View Center

Descendants Tree

Ancestors Tree

II. II. MultiTrees paper: MultiTrees paper: The The GoodGood

Excellent theoretical background and analysis of proposed solution

II. II. MultiTrees paper: MultiTrees paper: The The GoodGood

►Starts from real life problem: situation Starts from real life problem: situation actually occurring in authors' companyactually occurring in authors' company

II. II. MultiTrees paper: MultiTrees paper: The Not So The Not So GoodGood

►How many roots can we fit in one view?How many roots can we fit in one view?

►Allows reuse out of context(?)Allows reuse out of context(?)

►Must be constructed by handMust be constructed by hand

►No user testingNo user testing►However, all pointed out by the authors However, all pointed out by the authors

themselves (except last themselves (except last ))

III. III. Polyarchies paper:Polyarchies paper:

► ProblemProblem:: Understand Understand

relationships behind relationships behind multiple data basesmultiple data bases

How to viz a How to viz a metadirectory?metadirectory?

► Knowledge GapKnowledge Gap:: Current viz techniques do Current viz techniques do

not allow simultaneous not allow simultaneous view of relationships view of relationships along >= 1 dimensionalong >= 1 dimension

Employees:Benefits

Employees:Contact Info

Employees:Salaries

Employees:Access privileges

Bob

Bob

Bob

Bob

III. III. Polyarchies paper Polyarchies paper Key Key issuesissues::

►IssueIssue::

How to view How to view inter-relation inter-relation between between separate separate entities?entities?

►SolutionSolution::

Only show Only show parts of the parts of the parent parent hierarchy (not hierarchy (not global global relationship)relationship)

Query Term == Bishop

Lots of Bishops returned

AddTo btn

Desired Hierarchy View

== mgm

QuerySet: “Bishop, Scott” +

ppl Bishop Relates To

Matches

RClick here Pivot to other hierarchy view

LClick here becomes new

Pivot Point

III. III. Polyarchies paper Polyarchies paper Key Key issuesissues::

►ProblemProblem::

How to move How to move from one from one hierarchy to hierarchy to another another without without loosing loosing context?context?

►SolutionSolution::

Use Use animationanimation to to reduce reduce Cognitive Cognitive LoadLoad

III. III. Polyarchies paper: Polyarchies paper: ImplementationImplementation

III. III. Polyarchies paper: Polyarchies paper: The The GoodGood

► Used a Flash prototype Used a Flash prototype firstfirst

► Excellent formal user Excellent formal user

study (5 of them)study (5 of them)

► Allows users to choose Allows users to choose

animation speedsanimation speeds

► Good survey of Good survey of

previous work previous work But it confirms their own But it confirms their own

findingsfindings

III. III. Polyarchies paper: Polyarchies paper: The Not So The Not So GoodGood

► Quotes “”:Quotes “”:1.1. MultiTrees are multiple hierarchies with MultiTrees are multiple hierarchies with

shared subtrees.shared subtrees.

2.2. But Polyarchies are multiple But Polyarchies are multiple intersectingintersecting hierarchies, sharing at hierarchies, sharing at least one node rather than sharing least one node rather than sharing subtrees.subtrees.

3.3. Hence, MultiTrees are a subset of Hence, MultiTrees are a subset of polyarchies.polyarchies.

4.4. The added complexity requires a new The added complexity requires a new approach as described in this paper.approach as described in this paper.

Part 2: Part 2: SynthesisSynthesis

► Foreach (Paper) Do {Foreach (Paper) Do {

a)a) Assumptions behind each Assumptions behind each implementationimplementation

b)b) Did they solve the problem?Did they solve the problem?

}}

I. I. AssumptionsAssumptions Behind ConeTrees paper: Behind ConeTrees paper:

►Only visualizes Only visualizes hierarchical hierarchical information information structuresstructures; not ; not arbitrary arbitrary graphsgraphs i.e. No structure, No Vizi.e. No structure, No Viz

►Future work will solve:Future work will solve: Gains of 3D layoutGains of 3D layout If 3D maximizes use of screen spaceIf 3D maximizes use of screen space What other organizations can be usefully What other organizations can be usefully

displayed by Cone Treesdisplayed by Cone Trees What graphs can or cannot be displayed What graphs can or cannot be displayed

by Cone Treesby Cone Trees

I. I. Did They Solve the ProblemDid They Solve the Problem??[ConeTrees paper:][ConeTrees paper:]

► What problem are we talking about?What problem are we talking about? Problem(s) solved: Problem(s) solved:

1.1. ““Show the entire _structure_ of a complex Show the entire _structure_ of a complex organization in one viz”organization in one viz”

2.2. Shift part of cognitive load to perceptual Shift part of cognitive load to perceptual systemsystem

Paper quote: “[…] this is the first time Paper quote: “[…] this is the first time the organization chart could be seen in the organization chart could be seen in one visualization.” one visualization.” (Xerox Corp 650 executives – requires (Xerox Corp 650 executives – requires 80 pages)80 pages)

I. I. Did They Solve the ProblemDid They Solve the Problem??[ConeTrees paper:][ConeTrees paper:]

► Future Work leftovers:Future Work leftovers: From 10 refs, 5 are to selfFrom 10 refs, 5 are to self

►Earliest 1986Earliest 1986

►Latest (the wall) 1991Latest (the wall) 1991

►Progress: '86, '89, '90, '90, '91Progress: '86, '89, '90, '90, '91

► Bottom line: Bottom line: (plenty of time to do formal user (plenty of time to do formal user testing)testing)

ANDAND

(plenty of time to infer ecologically (plenty of time to infer ecologically valid task)valid task)

II. II. AssumptionsAssumptions Behind MultiTrees Behind MultiTrees paper:paper:

► No diamondsNo diamonds► BUTBUT::

People will want to People will want to store same node in store same node in more than one more than one structure structure OR OR

Two paths bellow Two paths bellow one nodeone node

II. II. Did They Solve the Problem?Did They Solve the Problem?[MultiTrees paper:][MultiTrees paper:]

►What problem are we talking about?What problem are we talking about?

They solved the problem they started up They solved the problem they started up to solve to solve

►How well was the Viz doneHow well was the Viz done??

In doing so, they inferred a new data In doing so, they inferred a new data structurestructure

Bottom line: WW research community Bottom line: WW research community benefits from their workbenefits from their work

III. III. AssumptionsAssumptions Behind Polyarchy Viz Behind Polyarchy Viz paper:paper:

►Designer Designer decides and has decides and has complete control complete control over:over: Which database to includeWhich database to include Which hierarchies to exposeWhich hierarchies to expose Candidate search attributesCandidate search attributes

►MS-only hardware/software:MS-only hardware/software: Uses PQLUses PQL Uses Polyarchy Query ServerUses Polyarchy Query Server Uses MS Metadirectory ServicesUses MS Metadirectory Services

III. III. Did They Solve the Problem?Did They Solve the Problem?[Polyarchy Viz paper:][Polyarchy Viz paper:]

►What problem are we talking about?What problem are we talking about? Problem(s) solved:Problem(s) solved:

►Allow user to see relationships between Allow user to see relationships between hierarchies in the context of selected hierarchies in the context of selected nodesnodes

►Allow user to see relationships between Allow user to see relationships between multiple entities within a hierarchymultiple entities within a hierarchy

III. III. Did They Solve the Problem?Did They Solve the Problem?[Polyarchy Viz paper:][Polyarchy Viz paper:]

►Why only MS, HR data set?Why only MS, HR data set? How about: How about:

►Stock MarketsStock Markets►Human GenomeHuman Genome►Biomed data in generalBiomed data in general

►Why so much self-praise for PQL?Why so much self-praise for PQL? ““rich QL, allowing enormous flexibility for rich QL, allowing enormous flexibility for

exploration”exploration”

►Perhaps a slightly self-centered Perhaps a slightly self-centered approach?approach?

Questions?Questions?