Making Sense of Search Results: The Presentation of Search Results for Professional and Desktop Sear

1
Making Sense of Search Results The Presentation
of Search Results for Professional and Desktop
Searchers

• Benjamin B. Bederson
• Computer Science Department/UMIACS
• Human-Computer Interaction Lab
• University of Maryland

2
Information Visualization

• Interactive visual display of abstract
  information to help users
• Find patterns, outliers and trends
• Explore data to build intuition
• Develop specific questions to be asked of more
  traditional mechanisms
• Visuals help us think
• Provide a frame of reference, a temporary storage
  area
• Effectively increase short term memory

3
Excellent Pattern Recognition

• How many 3s?

1281736875613897654698450698560498286762 980985845
3822450985645894509845098096585 909103020990595959
5772564675050678904567 884578980982167765487266490
8560912949686
1281736875613897654698450698560498286762 980985845
3822450985645894509845098096585 909103020990595959
5772564675050678904567 884578980982167765487266490
8560912949686
4
Pitfalls

• Complex navigation and occlusion
• Inappropriate use of 3D
• Meaningless spatial dimensions
• Self-Organizing Maps
• Unhelpful animation
• PowerPoint, Visual Thesaurus
• Inefficient and confusing use of screen space
• Chart Junk

Inventive, Imaginative, Ingenious, Fanciful!
gt But is it useful?
5
Strategy

• Show more than fits on the screen
• Scrolling
• 3D
• Dense information design
• Overview detail
• Techniques
• Zoomable User Interfaces (ZUIs)
• Transitional animations
• Work closely with users

Goal Support users to stay in the flow. NY
Times Feb 10, 2005
Bederson Shneiderman 2003 - Craft
6
PhotoMesa Zoomable Image Browser
Demo

• Browse large numbers of images
• See relationships among images
• Fast preview / detail
• Annotate Search
• Cluster results
• Query preview tells you where the data is
• Commercialized at www.photomesa.com

Suh, Ling, Bederson Jacobs - UIST
2003 Bederson - UIST 2001 Combs Bederson -
DL 1999
7
International Childrens Digital Library

• Largest freely available online collection of
  childrens books
• 28 languages, 592 books
• Over 40,000 unique visitors / month
• Children as design partners
• 7-11 years old
• Work at lab 2 afternoons a week, and 2 weeks
  over the summer
• Children generate design directions, suggest
  feedback, and help implement new technologies
• Funded by NSF, IMLS
• In collaboration with Profs. Allison Druin, Ann
  Weeks

www.icdlbooks.org
8
Differences in Searching and Browsing

• Children can search but
• They often prefer browsing
• Because children have a natural tendency to
  explore
• Because they dont plan out searches and react
  more to results
• Because spelling and typing are hard
• Generating keywords is hard
• Boolean is hard

Demo
9
ICDL In Use

• 600,000 unique visitors, 2 million book pages
  served
• 200 countries (2/3 of access outside US)
• 592 books in 28 languages, 25,000 pages
• 20 public domain, 60 1990 or later
• lt 10 search by keyword!
• Top category searches
• Language (Spanish, Chinese)
• Age (6-9)
• Physical attributes (short, rainbow)
• Type (picture books, fiction, fairytale, animal
  characters)

10
PaperLens - Non-Graph Network Visualization

• Conference Paper Analysis
• Show relationships dynamically
• Visual overview of data
• Multiple ways to filter, search and see
  relationships

Demo
Winner of InfoVis 2004 Contest
11
SpaceTree - See Hierarchies in Context
Demo

• Explore large hierarchies
• Gain understanding of relationships among data
• Integrate search/browse

Grosjean, Plaisant Bederson - InfoVis 2002
12
Conclusions

• Does zooming work?
• Is visualization and animation helpful?
• gt Clearly yes (sometimes)
• Good small representations needed
• Animation to help maintain object constancy best
• Meaningful spatial dimensions
• Integrate search/browse
• Understanding of domain and tasks crucial
• Good visualizations remain
• Hard to design
• Hard to deploy