Information Visualization INFORMS Roundtable Ben Shneiderman bencs umd

Information Visualization INFORMS Roundtable Ben Shneiderman (ben@cs. umd. edu) Founding Director (1983 -2000), Human-Computer Interaction Lab Professor, Department of Computer Science Member, Institutes for Advanced Computer Studies & Systems Research

Human-Computer Interaction Laboratory Interdisciplinary research community - Computer Science & Psychology - Information Studies & Education www. cs. umd. edu/hcil

User Interface Design Goals § Cognitively comprehensible: Consistent, predictable & controllable § Affectively acceptable: Mastery, satisfaction & responsibility NOT: Adaptive, autonomous & anthropomorphic

Scientific Approach (beyond user friendly) § § Specify users and tasks Predict and measure • • § time to learn speed of performance rate of human errors human retention over time Assess subjective satisfaction (Questionnaire for User Interface Satisfaction) § § Accommodate individual differences Consider social, organizational & cultural context

Design Issues § Input devices & strategies • • • § Output devices & formats • • • § § Keyboards, pointing devices, voice Direct manipulation Menus, forms, commands Screens, windows, color, sound Text, tables, graphics Instructions, messages, help Collaboration & communities Manuals, tutorials, training www. awl. com/DTUI usableweb. com hcibib. org useit. com

Library of Congress § Scholars, Journalists, Citizens § Teachers, Students

Visible Human Explorer (NLM) § Doctors § Surgeons § Researchers § Students

NASA Environmental Data § Scientists § Farmers § Land planners § Students

Bureau of Census § Economists, Policy makers, Journalists § Teachers, Students

NSF Digital Government Initiative Find what you need § Understand what you Find § UMd & UNC www. ils. unc. edu/govstat/

Session 1: The Case for Information Visualization Seven types (1 -, 2 -, 3 -, multi-dimensional data, temporal, tree and network data) Seven user tasks (overview, zoom, filter, details-on-demand, relate, history, and extract) Direct manipulation Dynamic queries

Session 2: Structured data Multidimensional and multivariate data Temporal data visualization Hierarchical and tree structured data Network information visualization

Session 3: User controls Zooming interfaces Focus+Context vs Overview+Detail Large Screen High Resolutions Displays 2 D versus 3 D desktops & workspaces Coordination of visualizations Other Challenges

Information Visualization The eye… the window of the soul, is the principal means by which the central sense can most completely and abundantly appreciate the infinite works of nature. Leonardo da Vinci (1452 - 1519)

Information Visualization Compact graphical presentation AND user interface for manipulating large numbers of items (102 - 106), possibly extracted from far larger datasets. Enables users to make discoveries, decisions, or explanations about patterns (trend, cluster, gap, outlier. . . ), groups of items, or individual items.

Information Visualization: Using Vision to Think § Visual bandwidth is enormous • • § Human perceptual skills are remarkable • Trend, cluster, gap, outlier. . . • Color, size, shape, proximity. . . Human image storage is fast and vast Opportunities • • • Spatial layouts & coordination Information visualization Scientific visualization & simulation Telepresence & augmented reality Virtual environments

Information Visualization: US Research Centers § Xerox PARC • • § Univ. of Maryland • • § § 3 -D cone trees, perspective wall, spiral calendar table lens, hyperbolic trees, document lens, butterfly dynamic queries, range sliders, starfields, treemaps zoombars, tight coupling, dynamic pruning, lifelines IBM Yorktown, AT&T-Lucent Technologies Georgia Tech, MIT Media Lab Univ. of Wisconsin, Minnesota, Calif-Berkeley Pacific Northwest National Labs

Which of my high-spending customers are most profitable? Which customers should I target for cross-sell/up-sell?

Dynamic Queries: Home. Finder www. cs. umd. edu/hcil/spotfire

Information Visualization: Mantra § § § § § Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand Overview, zoom & filter, details-on-demand

Film. Finder www. cs. umd. edu/hcil/spotfire

www. spotfire. com

Dynamap: Choropleth maps www. cs. umd. edu/hcil/census

Dynamap: Choropleth maps

Dynamap: Choropleth maps

Influence Explorer Tweedie, Spence et al. CHI 96

Information Visualization: Data Types Document Lens, See. Soft, Info Mural, Value Bars § 1 -D Linear 2 -D Map 3 -D World Multi-Dim § Temporal Perspective Wall, Life. Lines, Lifestreams, Project Managers, Data. Spiral § Tree Network Cone/Cam/Hyperbolic, Tree. Browser, Treemap § § GIS, Arc. View, Page. Maker, Medical imagery CAD, Medical, Molecules, Architecture Parallel Coordinates, Spotfire, XGobi, Visage, Influence Explorer, Table. Lens, DEVise Netmap, net. Viz, See. Net, Butterfly, Multi-trees (Online Library of Information Visualization Environments) otal. umd. edu/Olive

Information Visualization: Tasks § § Overview Gain an overview of the entire collection Zoom in on items of interest Filter out uninteresting items Details-on-demand Select an item or group and get details when needed View relationships among items § Relate History § Extract Allow extraction of sub-collections and of the query parameters § Keep a history of actions to support undo, replay, and progressive refinement

Information Visualization: Design Guidelines Direct manipulation strategies • Visual presentation of query components • Visual presentation of results • • • Rapid, incremental and reversible actions Selection by pointing (not typing) Immediate and continuous feedback Reduces errors Encourages exploration

Visual Information Seeking: Design Principles § Dynamic queries • • • § Starfield display • • • § Visual query formulation and immediate output Rapid, incremental and reversible actions Sliders, buttons, selectors Complete overview: ordinal & categorical variables as axes Colored points of light reveal patterns Zoom bars to focus attention Tight coupling to preserve display invariants • • • No errors Output becomes input Details-on-demand

Session 2: Structured data Multidimensional and multivariate data Temporal data visualization Hierarchical and tree structured data Network information visualization

Parallel Coordinates § § § One vertical bar per dimension Each point becomes a set of connected lines True multidimensional technique Needs powerful interface for filtering, marking, coloring Powerful technique but long learning period

Parallel Coordinates (Parallax-Inselberg)

Table. Lens/Eureka and Infozoom § § § Two compact views of tables Learned easily Table. Lens: local enlargement of areas of interest, creation of subtables Infozoom: shows distributions, allows progressive filtering Different orientation (vertical/horizontal)

Table. Lens/Eureka www. inxight. com

Info. Zoom www. humanit. de

Info. Zoom

Temporal data visualization: Life. Lines § § Parallel lines color/size coded & grouped in categories Zooming or hierarchical browsing allows focus+context • • Examples • Youth histories & medical records • Personal resumes, student records & performance reviews Challenges • Aggregation & alerts Overview & detail views • Easy import & export Labeling (Plaisant et al. , CHI 96) www. cs. umd. edu/hcil/Life. Lines

Life. Lines www. cs. umd. edu/hcil/lifelines

Lifelines: Customer records Temporal data visualization Medical patient histories § Customer relationship management § Legal case histories §

Perspective wall (Xerox Parc) Mackinlay et al, CHI 91

Time. Searcher www. cs. umd. edu/hcil/timesearcher

Tree Visualizations § § § Hierarchical data Challenge: understand relationships without getting lost Explicit vs. implicit depictions of trees • § Connections vs. containments Size limitations – breadth and depth

Tree Visualization Toolkits www. ilog. com

Spacetree + Familiar & animated + Space limited + Focuses attention Requires some learning - www. cs. umd. edu/hcil/spacetree

Hyperbolic trees § § § § Visually appealing Space limited 2 -level look-ahead Easy affordances Hard to scan Poor screen usage Too volatile Lamping et al. CHI 95

Startree Toolkit www. inxight. com

Cam. Tree - Cone. Tree Xerox PARC

Treemap - view large trees with node values + + Space filling Space limited Color coding Size coding Requires learning - Tree. Viz (Mac, Johnson, 1992) NBA-Tree(Sun, Turo, 1993) Winsurfer (Teittinen, 1996) Diskmapper (Windows, Micrologic) Treemap 97 (Windows, UMd) Treemap 3. 0 (Java) www. cs. umd. edu/hcil/treemaps

Treemap - Stock market, clustered by industry www. smartmoney. com/marketmap

Treemap – Product catalogs www. hivegroup. com

Treemap – Monitoring www. hivegroup. com

Million-Item Treemap www. cs. umd. edu/hcil/millionvis

GRIDL – Hierarchical Axes www. cs. umd. edu/hcil/west-legal/

Network Visualization § Arbitrarily connected items • § Problems • • § Nodes-links-paths-clusters Layout as size grows Clutter vs clusters Net. Map External relationships • • Geography Taxonomies Entrieva Semio. Map

Web Browsing www. kartoo. com

Communication Networks www. netviz. com www. ilog. com

Enterprise Networks www. lumeta. com

Treemap – Directed, Acyclic Graphs

Session 3: User controls Zooming interfaces Focus+Context vs Overview+Detail Large Screen High Resolutions Displays 2 D versus 3 D desktops & workspaces Coordination of visualizations Other Challenges

Fisheye views & Zooming User Interfaces § § § Distortion to magnify areas of interest User-control, zoom factors of 3 -5 Multi-scale spaces Zoom in/out & Pan left/right Smooth zooming Semantic zooming Overviews + details-on-demand Stasko, GATech

Glass. Eye – Zooming Exploration (see Hochheiser paper www. cs. umd. edu/hcil)

Date. Lens www. cs. umd. edu/hcil/datelens

Photo. Mesa www. cs. umd. edu/hcil/photomesa

Photo. Finder www. cs. umd. edu/hcil/photolib

Snap-Together Visualization www. cs. umd. edu/hcil/snap • Allow coordination designers to create novel layouts that combine existing visualizations • Allow users to navigate large datasets conveniently

Snap in Java – with Builder snap. cs. vt. edu

Hierarchical Clustering Explorer www. cs. umd. edu/hcil/multi-cluster

High-Resolution, Wall-Size Displays graphics. stanford. edu/~francois/ www. cs. umd. edu/gvil/

Spectrum of 3 -D Visualizations § § Immersive Virtual Environment with head-mounted stereo display and head tracking Desktop 3 -D for 3 -D worlds • medical, architectural, scientific visualizations § Desktop 3 -D for artificial worlds • Bookhouse, file-cabinets, shopping malls § Desktop 3 -D for information visualization • cone/cam trees, perspective wall, web-book • SGI directories, Visible Decisions, Media Lab landscapes • XGobi scatterplots, Themescape, Visage § Chartjunk 3 -D: barcharts, piecharts, histograms

Themescape Wise et al. , 1995 - see also www. omniviz. com

Starlight Battelle – Pacific Northwest National Lab

Mineset

Web. Book-Web. Forager Card, Robertson, George and York, CHI 96

Microsoft: Task Gallery research. microsoft. com/ui/Task. Gallery/

Clockwise 3 d www. clockwise 3 d. com

Clockwise 3 d

Other challenges: Labeling Excentric Labeling www. cs. umd. edu/hcil/excentric

Other challenges: Universal Usability § Helping novice users get started § 508 / disabled users § Range of devices, network speed, etc. www. otal. umd. edu/uupractice

Human-Computer Interaction Laboratory 20 th Anniversary Open House May 29 -30, 2003 www. cs. umd. edu/hcil

For More Information § Visit the HCIL website for 300 papers & info on videos (www. cs. umd. edu/hcil) § See Chapter 15 on Info Visualization Shneiderman, B. , Designing the User Interface: Strategies for Effective Human-Computer Interaction: Third Edition (1998) (www. awl. com/DTUI) § Book of readings: Card, S. , Mackinlay, J. , and Shneiderman, B. Information Visualization: Using Vision to Think (1999)