Empowering the library search experience: Endeca at NCSU Libraries

1
Empowering the library search experienceEndeca
at NCSU Libraries

• Kristin Antelman
• Maryland Library Association
• May 3, 2007

2
The Context
3
Themes

• Catalogs are hard to use

• Valuable metadata is buried

• What is the catalog of the future?

4
Catalogs are hard to use

• Poor subject access
• keyword gives too many or too few results
• authority searching is under-utilized

• Relevance is system sort order (in some ILSs)

• Unforgiving on spelling errors, stemming

5
Valuable metadata is buried

• Subject headings are not leveraged in keyword
  searching
• they should be browsed or linked from, not
  searched

• Data from the item record is not leveraged
• should be able to easily filter result sets
  using item type, location, circulation status

6
Endeca _at_ NCSU
7
What is Endeca?

• Software company based in Cambridge, MA
• Provide search software formany ecommerce sites

8
What is Faceted Navigation?
9
Technical overview
Information Access Platform
NCSU exports and reformats
Data Foundry
MDEX Engine
Parse text files
Raw MARC data
Indices
Flat text files
HTTP
HTTP
NCSU Web Application
10
Data
11
Catalogs are hard to useImproving the search
experience

• relevance ranking

12
Relevance ranking with Endeca

• Based on a locally managed, customizable
  algorithm
• Current strategy includes seven modules,
• 5 rank results dynamically on things like
  phrase, field in which term appears, weighted
  frequency
• 2 rules provide static ordering based on
  publication date and aggregate circulation totals

13
Catalogs are hard to useImproving the search
experience

• relevance ranking
• search enhancements

14
Search enhancements

• Automatic spell correction

• Automatic stemming

• Bookmark-ability

15
Still significant issues

• topical keyword search causes of the
  revolutionary war

• LCSH headingUNITED STATE--HISTORY--REVOLUTION,
  1775-1783--CAUSES

16
Keyword search
17
Authority search
18
Outcomes
19
Usage statistics
20
Search assistance

• Automatic spell correction
• 2.5

• Did you mean suggestions
• 1.9
• of those, 6 are selected

21
Usage statistics
22
Usage statistics
23
Usability testing

• Tasks completed more easily
• 55 easy or medium in old catalog
• 71 easy or medium in new catalog

• Tasks completed more quickly (36)

• Need more data and better methodology on how
  to measure catalog success

24
What is the catalog of the future? Opening Doors
25
Creating an open platform

• CatalogWS -gt web services
• RSS
• OpenSearch
• QuickSearch integration
• and more

26
RSS
27
RSS
28
OpenSearch
http//mycroft.mozdev.org/download.html?namencsu
librariessherlockyesopensearchsubmitformSea
rch
29
QuickSearch
30
More doors to open

• Search TRLN
• Unutilized features
• shopping cart
• Enriched results
• book jackets
• LibraryThing community features (?)
• FRBRized view, i.e., a work display

31
Starting a conversation
32
Thanks

• NCSU project site
• http//www.lib.ncsu.edu/endeca