Transcript The Harmony of Peers - Mining Peer Selection Patterns to Improve

The Harmony of Peers Mining Peer Selection Patterns
to Improve Consortial
Engineering Collections
Orion Pozo – NCSU Libraries
[email protected]
Engineering Libraries Division Poster Session
ASEE Annual Meeting, Portland OR
Monday, June 13, 2005
The Harmony of Peers
1
Summary
 Triangle Research Libraries Network (TRLN) is a
consortium of libraries at four North Carolina
universities
 TRLN shares a common book vendor, YBP Library
Services
 YBP provides Consortial & Peer Reports
 Goal:
– Use Consortium and Peer Data to Increase quality
and quantity of titles available to the consortium
The Harmony of Peers
2
The Triangle Research Libraries Network:
North Carolina State University
University of North Carolina, Chapel Hill
Duke University
North Carolina Central University
The Harmony of Peers
3
Peer Institutions*












Hong Kong Polytechnic Univ
RMIT
Monash University
Univ of Cincinnati
Iowa State Univ
MIT
Penn State Univ
Purdue Univ
Massey Univ
Chinese Univ of Hong Kong
Cornell Univ
Stanford Univ











Ohio State Univ
UC San Diego
Hong Kong Univ of Sci & Tech
Carnegie Mellon Univ
Univ of Technology, Sydney
Clemson Univ
Cal State Poly Univ, Pomona
Rensselaer Polytech Inst
Univ of Wisconsin, Madison
Univ of Calgary
Duke Univ
* Peers were chosen based on the number of book selections in YBP’s Gobi2
database in the subjects of technology and computer science
The Harmony of Peers
4
Goals -Increase quality of titles available to the consortium
-Reduce duplication of titles in the consortium
Methodology – For 2004-2005 use Consortial Reports to reduce
duplication
– At the same time use Peer Reports to identify quality
titles not in consortium
– Compare the selection of duplicates and unique titles
in the Consortium from 2003-2004 and 2004-2005
The Harmony of Peers
5
Project Outline:
At one TRLN institution
 Use Consortial Reports to reduce duplication & identify
titles not held
 Use Peer Reports to identify quality books not held
 Compare Consortial reports with Peer reports to
purchase quality items not in the consortium
 Compare duplication and number of titles for 2003-2004
(before Peer Reports) and 2004-2005 (after Peer
Reports)
The Harmony of Peers
6
Details
Limited to:
– Books ordered from one vendor
– Subjects:
• Computer Science (QA75-77)
• Engineering & Technology (T-TX)
– Only two years of data available from vendor
– One library (other consortial libraries
purchased as always)
The Harmony of Peers
7
YBP Consortial Report Form
The Harmony of Peers
8
Sample YBP Record


GobiTween shows slips were sent to 4 Consortium Libraries
PeerGroup shows 10 Peer Libraries Selected this title
The Harmony of Peers
9
Peer Activity Data for Single Record
Lists 10 Libraries that Selected this Book
The Harmony of Peers
10
Number of Titles Purchased
 From 2003-2004 to 2004-2005 TRLN Consortium purchasing of books
from this vendor in these subject areas decreased 23%
 Number of books purchased decreased from 5,357 to 4,122
Number of books purchased by libraries in TRLN Consortium
YEAR
Subject Group
TRLN LIBRARY
NORTH CAROLINA STATE/RALEIGH
UNIV OF NC/CHAPEL HILL
DUKE UNIVERSITY
UNIV OF NC/CHAPEL HILL/UL
DUKE UNIV SCHOOL OF LAW
UNIV OF NC/CHAPEL HILL HEALTH
UNIV OF NC/CHAPEL HILL-LAW
FUQUA SCHOOL OF BUSINESS/LIB
UNIV OF NC/CHAPEL HILL/NCC
UNC-CHAPEL HILL/STONE
NORTH CAROLINA CENTRAL UNIV
Total Consortium Purchases
2003-2004
QA75-77
802
400
46
14
2
1
1
0
0
0
0
1266
T
2372
778
833
19
14
40
1
17
11
3
3
4091
2004-2005
Combined
3174
1178
879
33
16
41
2
17
11
3
3
5357
Difference between 2004 & 2005
The Harmony of Peers
QA75-77
T
Combined
653
327
28
13
0
0
0
0
0
0
0
1021
1790
654
567
15
17
38
3
10
5
2
0
3101
2443
981
595
28
17
38
3
10
5
2
0
4122
-245
-990
-1235 -23%
11
Number of Libraries Selecting Titles
•While the number of books went down 23%, the number of titles was down
only 18%
•Number of titles held by a single library in the Consortium went up 7.8%
•Almost 200 more titles were purchased using this methodology than would
have been bought before it was available
Number of TRLN Libraries Selecting Titles
YEAR
2003-2004
Subject Group
QA75-77
T
No. of Libraries
1
2
3
4
5
Total No. of Titles
Total No. of Items (# Titles X # of Libraries)
905
137
26
1
1
1070
1266
1831
763
220
19
0
2833
4093
Percent
Change
2004-2005
Percent
Combined of Total
2736
900
246
20
1
3903
5359
51.1%
33.6%
13.8%
1.5%
0.1%
The Harmony of Peers
QA75-77
T
860
52
19
0
0
931
1021
1564
587
105
12
0
2268
3101
Percent 2003-04 to
Combined of Total 2004-05
2424
639
124
12
0
3199
4122
58.8%
31.0%
9.0%
1.2%
0.0%
7.8%
-2.6%
-4.7%
-0.3%
-0.1%
-18.0%
-23.1%
12
Conclusions
 Consortial Reports can be used effectively to
reduce duplication in book selection within a
consortium
 Peer Selection Data can be effectively used to
find high quality materials that are not owned
by consortium members
 Consortial and Peer data can be used
harmoniously to build richer collections
The Harmony of Peers
13
Recommendations for future studies
 Improvements in YBP’s Gobi2 database to allow
sorting by and exporting the number of Peers
selecting items would help data mining
 Cooperative use of this methodology by
multiple members of a consortium should
enhance desirable outcomes
 Replicating this study with multiple selectors in
larger consortia
The Harmony of Peers
14