Ordered and Quantum Treemaps: Making Effective Use of 2D Space to Display Hierarchies. By Bederson, B.B., Shneiderman, B., and Wattenberg, M. ACM Transactions.

Slides:



Advertisements
Similar presentations
The Web Wizards Guide to HTML Chapter Six Tables.
Advertisements

Chapter 3 – Web Design Tables & Page Layout
Equality Join R X R.A=S.B S : : Relation R M PagesN Pages Relation S Pr records per page Ps records per page.
Order Statistics Sorted
Section 6.1 Write Web text Use a mission statement Generate and organize content ideas Section 6.2 Use page dimension guidelines Determine content placement.
6 Developing Content and Layout Section 6.1 Generate and organize content ideas Write and organize Web text Section 6.2 Identify page dimension guidelines.
Cushion Treemaps and Botanical Visualization Yimeng Dou
Jun 2, 2014 IAT Trees2 Chapter 3.2 of Spence ______________________________________________________________________________________ SCHOOL OF INTERACTIVE.
LifeLines:Visualizing Personal Histories Plaisant, Milash, Rose, Widoff, Shneiderman Presented by Girish Kumar and Rajiv Gandhi.
Copyright © 2014, 2008 Pearson Education, Inc. Publishing as Pearson Addison-Wesley Starting Out with C++ Early Objects Eighth Edition by Tony Gaddis,
Label Placement and graph drawing Imo Lieberwerth.
Chapter 9: Searching, Sorting, and Algorithm Analysis
CPSC 335 Computer Science University of Calgary Canada.
Microsoft ® Research Faculty Summit 2002 Building Scalable User Interfaces for Mobile Devices using.NET Technologies Benjamin B. Bederson Computer Science.
1 Working with Web Graphics – About Web Graphics. File Formats. Image Resolution, Image Size – Creating Web Graphics. Getting Started, Paint Shop Pro 5.
1 Learning to Detect Objects in Images via a Sparse, Part-Based Representation S. Agarwal, A. Awan and D. Roth IEEE Transactions on Pattern Analysis and.
© University of Minnesota Data Mining for the Discovery of Ocean Climate Indices 1 CSci 8980: Data Mining (Fall 2002) Vipin Kumar Army High Performance.
Tree-Maps: A Space-Filling Approach to the Visualization of Hierarchical Information Structures Brian Johnson Ben Shneiderman (HCIL TR 91-06) Steve Betten.
Selection Sort, Insertion Sort, Bubble, & Shellsort
Software Visualization Space Filling Approach & Semantic Zooming Siva Venkatachalam 03/23/2004.
GUI Layout Managers Arkadiusz Edward Komenda. Outline Components and Containers Layout Managers Flow Layout Grid Layout Border Layout Nested Containers.
E.G.M. PetrakisHashing1 Hashing on the Disk  Keys are stored in “disk pages” (“buckets”)  several records fit within one page  Retrieval:  find address.
Programming Logic and Design Fourth Edition, Comprehensive
Ordered and Quantum Treemaps: Making effective use of 2D space to display hierarchies B. Bederson, B. Shneiderman, M. Wattenberg ACM Trans. On Graphics.
Routing 2 Outline –Maze Routing –Line Probe Routing –Channel Routing Goal –Understand maze routing –Understand line probe routing.
Quark QuarkXPress 4 Foundation Level Course. What is QuarkXPress? This courseware teaches the fundamentals of QuarkXPress 4.1. It is a page layout application.
Using Charts in a Presentation Lesson 6. Software Orientation Charts can help your audience understand relationships among numerical values. The figure.
A tale of 2-dimensional (Guillotine) bin packing Nikhil Bansal (IBM) Andrea Lodi (Univ. of Bologna, Italy) Maxim Sviridenko (IBM)
Database Management Systems, R. Ramakrishnan and J. Gehrke1 Tree-Structured Indexes Chapter 9.
Tree Structures (Hierarchical Information) cs5764: Information Visualization Chris North.
Tree-Maps: A Space-Filling Approach to the Visualization of Hierarchical Information Structures Presented by: Daniel Loewus-Deitch.
Oct 7, 2013 IAT Trees2 Chapter 3.2 of Spence ______________________________________________________________________________________ SCHOOL OF INTERACTIVE.
CHAPTER 11 Tables. How Are Tables Used Data Display  Very tidy and very useful Better Text Alignment  Putting text in tables allows you to format indents.
Charts and Graphs V
Adding Images & Working with Images Unit 2. TITLE CORNELL NOTES TOPIC: NOTES: Name: Date:08/10/2009 Period : Summary: To display Art To display Photographs.
Self-Organizing Agents for Grid Load Balancing Junwei Cao Fifth IEEE/ACM International Workshop on Grid Computing (GRID'04)
ITP 104.  While you can do things like this:  Better to use styles instead:
Information Design and Visualization
CS654: Digital Image Analysis Lecture 3: Data Structure for Image Analysis.
S EGMENTATION FOR H ANDWRITTEN D OCUMENTS Omar Alaql Fab. 20, 2014.
CHAPTER 09 Compiled by: Dr. Mohammad Omar Alhawarat Sorting & Searching.
Chapter 9 – Classification and Regression Trees
Information Visualization –III Treemaps and Fisheye Views.
IAT 814 Trees Chapter 3.2 of Spence ______________________________________________________________________________________ SCHOOL OF INTERACTIVE ARTS +
Mr. Justin “JET” Turner CSCI 3000 – Fall 2015 CRN Section A – TR 9:30-10:45 CRN – Section B – TR 5:30-6:45.
A Case Study of Interaction Design. “Most people think it is a ludicrous idea to view Web pages on mobile phones because of the small screen and slow.
CSC 211 Data Structures Lecture 13
CMSC 341 B- Trees D. Frey with apologies to Tom Anastasio.
Chapter 4 Working with Frames. Align and distribute objects on a page Stack and layer objects Work with graphics frames Work with text frames Chapter.
Fall 2002CS/PSY Information Visualization 2 Case Study: Portraying Hierarchies Visualizing hierarchies  Variety of techniques Traditional tree views,
Duy & Piotr. How to reconstruct a high quality image with the least amount of samples per pixel the least amount of resources And preserving the image.
PMR: Point to Mesh Rendering, A Feature-Based Approach Tamal K. Dey and James Hudson
© 2011 Delmar, Cengage Learning Chapter 4 Working with Frames.
R-Trees: A Dynamic Index Structure For Spatial Searching Antonin Guttman.
Creating Tables LESSON 6 - #1.06 ESSENTIAL STANDARD #1 - UNDERSTAND WORD PROCESSING INDICATOR #6 – CREATING TABLES.
IA and the Small Screen Carolyn Watters & Bonnie MacKay Web Information Filtering Lab Dalhousie University.
?. When designing, you should analyze your target audience. Design you webpage around your target audience Target Audience.
Liang, Introduction to Java Programming, Sixth Edition, (c) 2007 Pearson Education, Inc. All rights reserved Chapter 23 Algorithm Efficiency.
October 1, 2013Computer Vision Lecture 9: From Edges to Contours 1 Canny Edge Detector However, usually there will still be noise in the array E[i, j],
Efficient Motion Updates for Delaunay Triangulations Daniel Russel Leonidas Guibas.
Algorithm Design Techniques, Greedy Method – Knapsack Problem, Job Sequencing, Divide and Conquer Method – Quick Sort, Finding Maximum and Minimum, Dynamic.
IAT 355 Trees2 ______________________________________________________________________________________.
Basic Principles of Layout
Sorting Algorithms Ellysa N. Kosinaya.
Information Design and Visualization
Lecture 3 / 4 Algorithm Analysis
Information Visualization 2 Case Study: Portraying Hierarchies
Information Visualization (Part 1)
Charts A chart is a graphic or visual representation of data
Donghui Zhang, Tian Xia Northeastern University
Presentation transcript:

Ordered and Quantum Treemaps: Making Effective Use of 2D Space to Display Hierarchies. By Bederson, B.B., Shneiderman, B., and Wattenberg, M. ACM Transactions on Graphics (TOG), October Layla Shahamat Kenny Weiss March 8, 2005

Outline  Motivation  Main Ideas  Related Work  Demo Treemap 4.0  Quantum Treemaps  Demo Photomesa  Concluding Remarks

Motivation What information do you get from this tree? How about more realistic bigger trees?

Motivation  Displaying large hierarchical information structures  First Spark! Solving 1990’s common problem of displaying filled hard disk in order to achieve goals such as Displaying the file system in a compact way Better utilization of available pixels Visually appealing Informative, easy to browse, easy to read

Treemap Development  Treemap - Novel idea and design of the algorithm was developed by

What is a treemap?  Space-filling visualization method to represent large hierarchical structures of quantitative data.  The Key idea is creating the nested rectangles that make up the layout.

Treemap Algorithm Overview  Each node in the tree hierarchy has a name, and an associated size.  Treemap is constructed via recursive subdivision of the initial rectangle  The direction of the subdivision alternates per level between horizontally and vertically.  The area of each rectangle in the treemap is proportional to the size of that particular node.  Aspect ratio of a rectangle is the maximum of width/height and height/width.

Different types of Treemaps Quantum

Slice and Dice Algorithm  Creates rectangles with high aspect ratio As a result hard to see skinny rectangles 6 4 Areas 6, 6, 4, 3, 2, 2, horizontal 16/1Vertical 36/1

Cluster (Map of Market)

Cluster Treemap & Squarified Treemap  Drawbacks Rapid Dramatic changes in the layout resulting from change in data  Second by second updates  Flickering Ignoring the order of the data (e.g. alphabetical)  Switching between vertical and horizontal squares  Harder to locate data or see patterns

Ordered Treemap  Algorithm change smoothly under dynamic updates.  Produces rectangles with low aspect ratio. Higher readability  Solves the problem of ordered data such as alphabetical indexed data

Ordered Treemap Algorithm RpRp R 1 (< R p ) R 2 (> R p ) (< R 3 ) R 3 (> R p ) (> R 2 ) R p = {Size, Middle, Split-Size}

Ordered Treemap Algorithm Pivot by size Algorithm  Inspired byQuickSort  If the number of items <= 4, lay them out in R, stop.  Let P, the pivot, be the item with the largest size.  Divide R into 4 rectangles, R1,Rp,R2,R3  Divide the items in the list, except p items, into 3 lists L1,L2,L3  All L1 indexes are less than p.  All the items in L2 have smaller indexes than the ones in L3 (L2.index < L3.index)  Repeat until items <= 4.

Different types of Ordered Treemaps  All these algorithm preserve the ordering of the index of the items.  Pivot-by-size O(n*n)  Pivot-by-middle O(nlogn) worst case  Pivot-by-split-size

Strip Treemap Overview  Modification of Squarified Treemap Alg.  Preserves Order.  Eliminating the final skinny strip by developing look ahead strip  Better readability than the ordered treemap Algorithm.  Average Run time O(sqrt(n))

Strip Treemap Algorithm  Creates an empty strip called current strip.  Adds a rectangle to the current strip  If average aspect ratio increases, remove the rectangle from the strip Create a new strip & add the rectangle  If average aspect ratio decreases or stays the same, add the rectangle to the current strip.

How Do We Compare?  Evaluation Metric Average aspect Ratio of a treemap layout  Average aspect ratio is arithmetic average of all aspect ratios of all leaf-node rectangles. The lowest is 1.Different calculation is possible. Layout Distance Change  Distance function measuring how much rectangles move as data gets updated Readability: how easy it is to locate an item  Measuring eye motion direction changes

Monte Carlo Trials Experiment Design  The data constantly gets updated.  For each experiment ran 100 trials of 100 steps each.  3 different collections of data

Monte Carlo Trials Experiment Results  Slice and Dice Method shows the tradeoff between aspect ratio and smooth updates.  Ordered and Strip treemaps fall in the middle.  Cluster and squarified treemaps show low aspect ratio, large changes.

Static Stock Market Experiment  Data: 535 publicly traded companies  High aspect ratios are due to the outliers in the data

User Study of Layout Readability  100 rectangles with random size & uniform distribution.  Measured the time it took the user to find a numerical ID.  Each subject performed 10 tasks for each 3 alg.  20 subjects, 20% female,80% male, 50% student  Squarified treemap longest time to locate an item lowest user preference

User Study of Layout Readability results  Subject preference was in the same direction as the readability metric  Strip users found the item 60% faster than when using squarified treemap.

Treemap Demo

Quantum Treemaps Problem  Input size of elements is fixed Similar Images  Pictures  Pages  Grouping Integral multiples of fixed input  Search Meaningful categories Ordered layout

Quantum Treemaps Procedure  Similar to other treemaps  Input: List of image groups  Number elements in each group Fixed aspect ratio  Output List of rectangles  Constraints Guaranteed to fit images Can have extra space Must fit contents into rows and columns

Fit images into rectangles  Scale to fit Long skinny rectangle Visually unattractive Slow to scan  Align content Global Grid Groups are integral multiples of quanta

Quantized Strip Treemap  Difference from normal striptree (ST) Rectangle Area = Integral multiples of quanta Ragged Edges  Distribute extra space throughout width  Same complexity as ST Up to a constant  Can use other treemap strategies Subtle changes may be necessary

Element Aspect Ratio  Aspect ratio (AR) doesn’t affect layout algorithm Can stretch out starting box by inverse of aspect ratio Visually has same effect

Ragged Edges  QST distribute space along width to fix ragged edge  General case distribute globally  Left and Bottom edges may be ragged

Horizontal and Vertical Growth  Grow to match rectangle to quantum ratio Experiments show best results when  Grow width in wide layouts  Grow height in vertical layouts x x x Change width Wide layout Change height

Comparison QT has better average aspect ratio, but wastes more space than ordinary treemap

Analysis  QT works better when groups have more elements More flexibility Wastes less space (proportionally)  1000 elements (30x34), (31x33), (32x32)  Each element is ~0.1%  5 elements (1x5), (2x3)  Each element is ~20%  Single global grid  Quantum size Usually, if domain requires constant size, other considerations are outweighed

Demo

Thoughts  Informative background information on various Treemap strategies  Experiments discuss dynamic ordered data Order preservation helps users orient themselves Categorical data not discussed  Photomesa Impressive overview of a lot of information very nice {overview, zoom, filter, details} GUI works well while loading images Zoom is abrupt and pixelated  Compare to Picasa and Photoshop Album  TreeMap Would be nice if “File Mapper” offered file previews