1. 1/48
Visual search for supporting content
exploration in large document collections
Drahomira Herrmannova and Petr Knoth

2. 2/48
Contents
• What do we do
• Information Visualisations and Visual Search
Interfaces
• Our approach
• Conclusion

3. 3/48
Contents
• What do we do
• Information Visualisations and Visual Search
Interfaces
• Our approach
• Conclusion

4. 4/48
What do we do
• Improve search in (large) document collections
• Examples of collections:
– News articles
– Cultural heritage collection
– Collection of scientific papers
• Current search engines:
– Support for lookup
– Much less support for exploration

5. 5/48
Search tasks (Rose and Levinson, 2004)
• Undirected (or exploratory) queries – significant
portion of all searches (Rose and Levinson, 2004)

6. 6/48
Exploratory search (Marchionini, 2006)

7. 7/48
How to support exploratory search
• One possible solution – information
visualisation
• Why?
– Easier to communicate structure, organisation and
relations in content
– Visually appealing

8. 8/48
Contents
• What do we do
• Information Visualisations and Visual Search
Interfaces
• Our approach
• Conclusion

9. 9/48
Information Visualisation (1/2)
• Division according to granularity of
information
– Collection level
– Document level
– Intra-document level

10. 10/48
Collection level visualisations
• Visualise attributes of the collection
• Typically aim at providing a general overview
of the collection content
• Examples

11. 11/48
Tag clouds (Montero and Solana, 2006)

12. 12/48
TIARA (Wei et al., 2010)

13. 13/48
GRIDL (Schneiderman et al., 2000)

14. 14/48
Document level visualisations
• Visualise attributes of the collection items
• Mutual links and relations of collection items
• Examples

15. 15/48
Hopara (Milne and Witten, 2011)

16. 16/48
Wivi (Lehmann et al., 2010)

17. 17/48
Apolo (Chau et al., 2011)

18. 18/48
Intra-document level visualisations
• Visualise the internal structure of a document
• Example

19. 19/48
TileBars (Hirst, 1995)

20. 20/48
Information Visualisation (2/2)
• Division according to the “starting point” of
the visualisation
– Browsing focused
– Query focused

21. 21/48
Browsing focused
• Exploration starts at a specific point in the
collection from which the user navigates
through the collection
• Usually the same starting point is used every
time

22. 22/48
InfoSky (Granitzer et al., 2004)

23. 23/48
Query focused
• Starts with a query
• The query determines the entry point from
which the exploration starts

24. 24/48
ThinkPedia (Hirsch et al., 2009)

25. 25/48
Our approach
• Document level information
• Query focused browsing

26. 26/48
Design principles (1/2)
• For visual search interfaces
• Should be considered when designing the
interface
• Related studies:
– Chen and Yu, 2000
– Sebrechts et al., 1999

27. 27/48
Design principles (2/2)
1. Added value
2. Simplicity
3. Visual legibility
4. Use of colours
5. Dimension
6. Fixed spatial location

28. 28/48
Contents
• What do we do
• Information Visualisations and Visual Search
Interfaces
• Our approach
• Conclusion

29. 29/48
Considered types of collections
• Every document in a collection defined
according to a set of dimensions
• Dimensions typically of different types
• Document = set of properties expressing
values of dimensions
• Dimensions always present
• Examples

30. 30/48
News articles collection
• Dimensions:
– Time
– Themes
– Locations
– Relations to other articles

31. 31/48
Cultural heritage artifacts
• Dimensions:
– Artifact type
– Historical period
– Style
– Material

32. 32/48
Scientific papers
• Dimensions:
– Citations
– Authors
– Concepts
– Similarities with other articles

33. 33/48
The visualisation

34. 34/48
The visualisation

35. 35/48
The visualisation

36. 36/48
The visualisation

37. 37/48
Discovering connections

38. 38/48
Comparing and contrasting documents

39. 39/48
Limitations
• In theory not restricted, the limitations might
be:
– the size and resolution of the screen
– the limitations of human perception

40. 40/48
Contents
• What do we do
• Information Visualisations and Visual Search
Interfaces
• Our approach
• Conclusion

41. 41/48
Conclusion (1/2)
• Motivation:
1. Provide better support for exploratory search
than current textual interfaces
2. Interface that is conceptually applicable in any
document collection regardless of its type
3. Provide an added value by assisting in the
discovery of interesting connections that would
otherwise remain hidden

42. 42/48
Conclusion (2/2)
• Results:
1. Support for comparing and contrasting content.
2. Support for exploration across dimensions.
3. Universal approach to the visualised dimensions.

43. 43/48
Future plans
• Planned release end of June
• Integration with CORE system
• Evaluation

44. 44/48
References (1/4)
• G. Marchionini. Exploratory search: from finding to understanding.
Communications of the ACM - Supporting exploratory search. 2006.
• D. Rose & D. Levinson. Understanding user goals in web search.
Proceedings of the 13th conference on World Wide Web. 2004.
• Yusef Hassan-Montero and Victor Herrero-Solana. Improving tag-clouds as
visual information retrieval interfaces. In MERIDA, INSCIT2006
CONFERENCE. 2006.
• Furu Wei, Shixia Liu, Yangqiu Song, Shimei Pan, Michelle X. Zhou, Weihong
Qian, Lei Shi, Li Tan, and Qiang Zhang. Tiara: a visual exploratory text an-
alytic system. In Proceedings of the 16th ACMSIGKDD international
conference on Knowledge discovery and data mining. 2010.

45. 45/48
References (2/4)
• Ben Shneiderman, David Feldman, Anne Rose, and Xavier Ferré Grau.
Visualizing digital librarysearch results with categorical and hierarchical
axes. In Proceedings of the fifth ACM conference on Digital libraries. 2000.
• Marti A. Hearst. TileBars: Visualization of Term Distribution Information in
Full Text Information Access. In the Proceedings of the ACM SIGCHI
Conference on Human Factors in Computing Systems. 1995.
• David Milne, Ian Witten. A link-based visual search engine for Wikipedia.
Proceeding of the 11th annual international ACM/IEEE joint conference on
Digital libraries. 2011.
• Simon Lehmann, Ulrich Schwanecke, and Rolf Dorner. Interactive
visualization for opportunistic exploration of large document collections.
Information Systems. 2010.

46. 46/48
References (3/4)
• Duen Horng Chau, Aniket Kittur, Jason I. Hong, and Christos Faloutsos.
Apolo: making sense of large network data by combining rich user
interaction and machine learning. In Proceedings of the 2011 annual
conference on Human factors in computing systems. 2011.
• Michael Granitzer, Wolfgang Kienreich, Vedran Sabol, Keith Andrews, and
Werner Klieber. Evaluating a system for interactive exploration of large,
hierarchically structured document repositories. In Proceedings of the IEEE
Symposium on Information Visualization. 2004.
• Christian Hirsch, John Hosking, and John Grundy. Interactive visualization
tools for exploring the semantic graph of large knowledge spaces.
Interfaces. 2009.

47. 47/48
References (4/4)
• Chaomei Chen and Yue Yu. Empirical studies of information visualization: a
meta-analysis. Int. J. Hum.- Comput. Stud. 2000.
• Marc M. Sebrechts, John V. Cugini, Sharon J. Laskowski, Joanna Vasilakis,
and Michael S. Miller. Visualization of search results: a comparative
evaluation of text, 2d, and 3d interfaces. In Proceedings of the 22nd
annual international ACM SIGIR conference on Research and development
in information retrieval. 1999.

48. 48/48
Thanks for listening!
Questions?