Advisor: Dr. Michele C. Weigle. Committee: Dr. Michael L. Nelson, Dr. Ravi Mukkamala Slide 32-34 contain embedded video which has been embedded as youtube video in this slideshare Direct links to videos: Treemap: http://www.youtube.com/watch?v=BJDrxQEEFYM Timecloud: http://www.youtube.com/watch?v=YYkI6aBO0to Bubble Chart, Image Plot and Timeline: http://www.youtube.com/watch?v=j94clxqKQk8 Abstract: Archive-It, a subscription service from the Internet Archive, allows users to create, maintain, and view digital collections of web resources. The current interface of Archive-It is largely text-based, supporting drill-down navigation using lists of URIs. While this interface provides good searching capabilities, it is not very efficient for browsing. In the absence of keywords, a user has to spend large amount of time trying to locate a webpage of interest. In order to provide a better visual experience to the user, we have studied the underlying characteristics of Archive-It collections and implemented six different visualizations (treemap, time cloud, bubble chart, image plot, timeline and wordle), each highlighting one or more of the underlying characteristics of the collection. Archive-It supports grouping of webpages into categories, however, it does not enforce its usage. As a result there are many collections with missing or improper grouping. For such collections, we present a method of grouping webpages based on a set of pre-defined rules.

1. Visualizing Digital Collections at
Archive-It
Kalpesh Padia
Director: Michele C. Weigle
Committee: Michael L. Nelson
Ravi Mukkamala
7/20/2012 MS Thesis - August 2012 1

2. Agenda
Introduction
Motivation
Related Work
Collection Retrieval and Processing
Visualizations
Case Studies
Future Work
Conclusion
7/20/2012 MS Thesis - August 2012 2

3. INTRODUCTION AND
MOTIVATION
7/20/2012 MS Thesis - August 2012 3

4. Digital Archives
http://www.loc.gov/index.html
http://digitalcollections.library.yale.edu/
7/20/2012 MS Thesis - August 2012 4

5. Archive-It
http://archive-it.org/
7/20/2012 MS Thesis - August 2012 5

6. Archive-It Collection Hierarchy
Collection
Root Title
Level 1 Category 1 Category n
Level 2 Web page 1 Web page n
Level 3 (Leaf Archived Archived
Nodes) Version 1 Version n
7/20/2012 MS Thesis - August 2012 6

7. Exploring Archive-It Collections
http://archive-it.org/collections/1068
7/20/2012 MS Thesis - August 2012 7

8. Exploring Archive-It Collections
http://archive-it.org/collections/1068
7/20/2012 MS Thesis - August 2012 8

9. Exploring Archive-It Collections
http://archive-it.org/collections/1068
7/20/2012 MS Thesis - August 2012 9

10. Exploring Archive-It Collections
http://wayback.archive-it.org/1068/*/http://acda.co/
7/20/2012 MS Thesis - August 2012 10

11. Exploring Archive-It Collections
http://archive-it.org/collections/1068
7/20/2012 MS Thesis - August 2012 11

12. Exploring Archive-It Collections
http://archive-it.org/collections/2836
7/20/2012 MS Thesis - August 2012 12

13. Drawbacks
No visual feedback
Discovering individual pages is difficult
Optional metadata and categorization
Collection structure known only to curator
7/20/2012 MS Thesis - August 2012 13

14. Contribution
Interactive visualizations
 Treemap
 Time cloud
 Bubble chart
 Image plot
 Wordle
 Timeline
Temporal exploration of collections
Uncover collection structure
7/20/2012 MS Thesis - August 2012 14

15. RELATED WORK
7/20/2012 MS Thesis - August 2012 15

16. Microsoft Pivot
http://www.microsoft.com/silverlight/pivotviewer/
7/20/2012 MS Thesis - August 2012 16

17. Page History Explorer
A. Jatowt, Y. Kawai, and K.
Tanaka, “Visualizing
Historical Content of Web
Pages,” in Proceedings of the
17th international conference
on World Wide Web,2008.
7/20/2012 MS Thesis - August 2012 17

18. 3D Wall
http://www.webarchive.org.uk/ukwa/wall/Blogs
7/20/2012 MS Thesis - August 2012 18

19. Treemap
Johnson and Shneiderman, “Space-Filling Approach to the Visualization of Hierarchical Information Structures” in
proceedings of the 2nd conference on Visualization '91
7/20/2012 MS Thesis - August 2012 19

20. Series Browser
M. Whitelaw, “Visualising Archival Collections: The Visible Archive
Project,” in Archives and Manuscripts, vol. 37, Issue 2, 2009.
7/20/2012 MS Thesis - August 2012 20

21. A1 Explorer
M. Whitelaw, “Visualising Archival Collections: The Visible Archive
Project,” in Archives and Manuscripts, vol. 37, Issue 2, 2009.
7/20/2012 MS Thesis - August 2012 21

22. EASY
Scharnhorst et.al. “Looking at a digital research data archive Visual
interfaces to EASY,” in CORR, 2012, http://arxiv.org/abs/1204.3200
7/20/2012 MS Thesis - August 2012 22

23. Wordle
. Jonathan Feinberg, http://wordle.net/ , Dogear
7/20/2012 MS Thesis - August 2012 23

24. DATA RETRIEVAL AND
PROCESSING
7/20/2012 MS Thesis - August 2012 24

25. 11 Collections, 2K+ Web pages, 70K+ Mementos
7/20/2012 MS Thesis - August 2012 25

26. Data Retrieval & Processing
Retrieval:
 Screen scrape
 Copy collection hierarchy
 Store page content
Processing:
 Calculate TF and TF-IDF
 Generate screenshots
 Generate wordles
 Rule-based categorization
 Construct JSON
7/20/2012 MS Thesis - August 2012 26

27. No categorization
7/20/2012 MS Thesis - August 2012 27
http://www.archive-it.org/collections/2836

28. Improper Categorization
7/20/2012 MS Thesis - August 2012 28
http://www.archive-it.org/collections/2323

29. Rule based categorization
News Web pages
Blogs
Social Media
Videos
7/20/2012 MS Thesis - August 2012 29
http://www.archive-it.org/collections/2836

30. Special URI and TLD based categorization
Pakistani news web
pages
7/20/2012 MS Thesis - August 2012 30
http://www.archive-it.org/collections/2836

31. VISUALIZATIONS
7/20/2012 MS Thesis - August 2012 31

32. Treemap
7/20/2012 MS Thesis - August 2012 32

33. Time Cloud
7/20/2012 MS Thesis - August 2012 33

34. Bubble Chart, Image Plot & Timeline
7/20/2012 MS Thesis - August 2012 34

35. CASE STUDIES
7/20/2012 MS Thesis - August 2012 35

36. 1. Collection Building and Growth
7/20/2012 MS Thesis - August 2012 36

37. 2. Re-Categorization
(Pakistan Flood: no categorization)
7/20/2012 MS Thesis - August 2012 37

38. 2. Re-Categorization
(Pakistan Flood: after categorization)
7/20/2012 MS Thesis - August 2012 38

39. 3. Collection Synopsis
7/20/2012 MS Thesis - August 2012 39

40. 3. Collection Synopsis
7/20/2012 MS Thesis - August 2012 40

41. 3. Collection Synopsis
7/20/2012 MS Thesis - August 2012 41

42. 3. Collection Synopsis
7/20/2012 MS Thesis - August 2012 42

43. 3. Collection Synopsis
7/20/2012 MS Thesis - August 2012 43

44. 4. Theme Tracking
7/20/2012 MS Thesis - August 2012 44

45. 4. Theme Tracking
7/20/2012 MS Thesis - August 2012 45

46. 4. Theme Tracking
7/20/2012 MS Thesis - August 2012 46

47. 4. Theme Tracking
7/20/2012 MS Thesis - August 2012 47

48. Informal User Evaluation
Alex Thurman, Columbia University Libraries
Feedback on
 ease of browsing and obtaining information
 user-friendliness of the interface
 whether they prefer textual or graphical
interface
 most effective visualization
 effectiveness of the rule-based categorization
in exploring archives
7/20/2012 MS Thesis - August 2012 48

49. Feedback
Effective visualizations:
 Treemap – color coding useful for identifying newer
additions
 Image plot – screenshots with mouse-over wordles
allow for good navigation
 Timeline – useful for visualizing development of
groups in collection
Suggestions
 Broader timescale for treemaps
 Include stop words from other languages
7/20/2012 MS Thesis - August 2012 49

50. FUTURE WORK AND
CONCLUSION
7/20/2012 MS Thesis - August 2012 50

51. Future Work
N-Gram wordles
Term expansion
Krovetz stemmer (dictionary based stemmer)
Integration with Archive-It
Detailed user evaluation
Implementation for other archives
7/20/2012 MS Thesis - August 2012 51

52. Conclusion
Identified metrics for collections
7/20/2012 MS Thesis - August 2012 52

53. Conclusion
Identified metrics for collections
Visualizations
 Treemap
7/20/2012 MS Thesis - August 2012 53

54. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
7/20/2012 MS Thesis - August 2012 54

55. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
 Bubble chart
7/20/2012 MS Thesis - August 2012 55

56. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
 Bubble chart
 Image plot
7/20/2012 MS Thesis - August 2012 56

57. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
 Bubble chart
 Image plot
 Wordle
7/20/2012 MS Thesis - August 2012 57

58. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
 Bubble chart
 Image plot
 Wordle
 Timeline
7/20/2012 MS Thesis - August 2012 58

59. Conclusion
Identified metrics for collections
Visualizations
 Treemap
 Time cloud
 Bubble chart
 Image plot
 Wordle
 Timeline
Rule – based categorization
7/20/2012 MS Thesis - August 2012 59

60. BACKUP
7/20/2012 MS Thesis - August 2012 60

61. Time Span
Small 1 Day - 2 Weeks
Time span Medium 2 Weeks - 4 Months
Large > 4 Months
http://wayback.archive-it.org/1068/*/http://amigosdemujeres.org/
7/20/2012 MS Thesis - August 2012 61

62. Groups
Small 1
Groups Medium 2-5
Large >5
http://www.archive-it.org/collections/1068
7/20/2012 MS Thesis - August 2012 62

63. URI Domains
Small 1 - 10
URI Domains Medium 11 - 20
Large > 20
http://www.archive-it.org/collections/2836
7/20/2012 MS Thesis - August 2012 63

64. Number of Web Pages
Small 1 - 10
# of Web Pages Medium 11 - 99
Large > 99
http://www.archive-it.org/collections/2836
7/20/2012 MS Thesis - August 2012 64

65. Jigsaw
Stasko et.al., IEEE VAST 2007
7/20/2012 MS Thesis - August 2012 65

66. Themeriver
Wei et.al. in SIGKDD, 2010
7/20/2012 MS Thesis - August 2012 66

67. Time Cloud
7/20/2012 MS Thesis - August 2012 68

68. Bubble Chart
7/20/2012 MS Thesis - August 2012 69
http://tmix.cs.odu.edu:8080/project/test.php?coll_id=1068

69. Image Plot with Wordle
7/20/2012 MS Thesis - August 2012 70
http://tmix.cs.odu.edu:8080/project/test.php?coll_id=1068

70. Timeline
7/20/2012 MS Thesis - August 2012 71
http://tmix.cs.odu.edu:8080/project/test.php?coll_id=1068