Document Collections

1
Document Collections

• cs5984 Information Visualization
• Chris North

2
Where are we?

• Multi-D
• 1D
• 2D
• Hierarchies/Trees
• Networks/Graphs
• Document collections
• 3D

• Design Principles
• Empirical Evaluation
• Java Development
• Visual Overviews
• Multiple Views
• Peripheral Views

3
Structured Document Collections

• Multi-dimensional
• author, title, date, journal,
• Trees
• dewey decimal
• Networks
• web, citations

4
Envision

• Ed Fox, et al.
• Multi-D
• similar to Spotfire

5
Unstructured Document Collections

• Focus on Full Text
• Examples
• digital libraries, encyclopedia
• Web, homepages, photo collections
• Tasks
• search, keyword
• Browse
• Themes, subjects, topics, library coverage
• Size, distributions

6
Visualization Strategies

• Cluster Maps
• Keyword Query
• Relationships
• Reduced representation
• User controlled layout

today
today
7
Cluster Map

• Create a map of the document collection
• Similar documents near
• Dissimilar document far
• Grocery store concept

8
Document Vectors

• Doc1 Doc2 Doc3
• aardvark 1 2 0
• banana 2 1 0
• chris 0 0 3
• Similarity between pair of docs
• Layout documents in 2-D map by similarity
• similar to spring model for graph layout

9
Cluster Algorithms

• Partition clustering Partition into k subsets
• Pick k seeds
• Iteratively attract nearest neighbors
• Hierarchical clustering Dendrogram
• Group nearest-neighbor pair
• Iterate

10
Kohonen Maps

• Xia Lin, Document Space
• samal, ying
• http//faculty.cis.drexel.edu/sitemap/index.html

11
(No Transcript)
12
Themescapes, Cartia

• PNL
• Mountain height Cluster size

13
WebSOM

• http//websom.hut.fi/websom/

14
Map.net

• http//maps.map.net/start

15
Cluster Map

• Good
• Map of collection
• Major themes and sizes
• Relationships between themes
• Scales up

• Bad
• Where to locate documents with multiple themes?
• Both mountains, between mountains, ?
• Relationships between documents, within
  documents?
• Algorithm becomes (too) critical

16
Keyword Query

• Keyword query, Search engine
• Rank ordered list
• Information Retrieval

17
Tilebars

• Hearst, Tilebars
• reenal, xueqi
• http//elib.cs.berkeley.edu/tilebars/

18
VIBE

• Korfhage, http//www.pitt.edu/korfhage/interface
  s.html
• Documents located between query keywords using
  spring model

19
VR-VIBE
20
Keyword Query

• Good
• Reduces the browsing space
• Map according to users interests

• Bad
• What keywords do I use?
• What about other related documents that dont use
  these keywords?
• No initial overview
• Mega-hit, zero-hit problem

21
Assignment

• Thurs Document Collections
• Bederson, Image Browsing
• Rui, anusha
• Card, Web Book and Web Forager
• mrinmayee, ming
• Demo your hw3 tues or thurs

22
Next Week

• Tues 3-D data
• Kniss, Interactive Volume Rendering with Direct
  Manip
• xueqi, mahesh
• Thurs Workspaces
• Robertson, Task Gallery
• supriya, varun
• Upson, AVS
• christa, jun
• Thanksgiving break
• Tues 27 Debates
• Kobsa, Empirical comparison of comm infovis
  systems
• kunal, zhiping

23
Upcoming Sched

• Tues 3-D data
• Thurs Workspaces
• Thanksgiving break
• Tues 27 Debates
• Thurs 29 How (not) to lie with visualization
• Dec project presentations
• Dec 7 CHI 2-pagers due, student posters due