E xample of a bibliomining system: logs.library.cornell

Example of a bibliomining system:logs.library.cornell.edu

Adam ChandlerData Discussion on Library Data

Cornell University LibraryJune 1, 2012

logs.library.cornell.edu 2

What is it?

logs.library.cornell.edu 3

logs.library.cornell.edu 4

logs.library.cornell.edu 5

Live demo 1

logs.library.cornell.edu 6

That’s it? Why bother? Just use Google Analytics

logs.library.cornell.edu 7

logs.library.cornell.edu 8

Why not Google Analytics?

logs.library.cornell.edu:

1. Uses Cornell single sign for security and convenience2. gives us the freedom to export and use the data anyway we want for

our special reporting needs3. requires no changes to our websites. Google Analytics requires a section

of Javascript code that sends information about each request to Google where it is recorded. Repeated privacy violations from commercial sites such as Facebook are driving some users towards widgets such as ghostery (http://news.ghostery.com/) that block javascript based web tracking.

4. our flexible design allows us to store logs which cannot easily be tracked with javascript: examples: PURL, checkip, flickr

logs.library.cornell.edu 9

'bibliomining'

Nicholson, S. (2003) The Bibliomining Process: Data Warehousing and Data Mining for Library Decision-Making. Information Technology and Libraries 22 (4)

logs.library.cornell.edu 10

“The term 'bibliomining' was first used by Nicholson and Stanton (2003) in discussing data mining for libraries. In the research literature, most works that contain the terms 'library' and 'data mining' are not talking about traditional library data, but rather using library in the context of software libraries, as data mining is the application of techniques from a large library of tools. In order to make it more conducive for those concerned with data mining in a library setting to locate other works and other researchers, the term 'bibliomining' was created. The term pays homage to bibliometrics, which is the science of pattern discovery in scientific communication.”

logs.library.cornell.edu 11

logs.library.cornell.edu 12

logs.library.cornell.edu 13

“Bibliomining is the application of statistical and pattern-recognition tools to large amounts of data associated with library systems in order to aid decision-making or justify services.”

logs.library.cornell.edu 14

“The bibliomining process consists of

· determining areas of focus;· identifying internal and external data sources;· collecting, cleaning, and anonymizing the data into a data warehouse;· selecting appropriate analysis tools;· discovery of patterns through data mining and creation of reports with traditional analytical tools; and· analyzing and implementing the results.”

logs.library.cornell.edu 15

Nicholson, S. (2003) The Bibliomining Process: Data Warehousing and Data Mining for Library Decision-Making. Information Technology and Libraries 22 (4)

logs.library.cornell.edu 16

“The process is cyclical in nature: as patterns are discovered, more questions will be raised which will start the process again. As additional areas of the library are explored, the data warehouse will become more complete, which will make the exploration of other issues much easier.”

logs.library.cornell.edu 17

Apache Log

logs.library.cornell.edu 18

Apache Log

logs.library.cornell.edu 19

CU (Campus)

CU Lib (Staff)

CU Lib (Public)

CU (Weill) CU (Qatar)

Ithaca not CU

NY not Ithaca

USA not NYOverseas

CUL Logs IP Address Groups

logs.library.cornell.edu 20

logs.library.cornell.edu 21

Live demo 2

logs.library.cornell.edu 22

Who uses it and for what?

logs.library.cornell.edu 23

How do I get help?

logs.library.cornell.edu 24

How do I get help?

logs.library.cornell.edu 25

https://confluence.cornell.edu/display/culweblogstool/

logs.library.cornell.edu 26

Credits

logs.library.cornell.edu 27

Credits

System design Adam Chandler and Pete HoytSoftware Pete Hoyt User interface Adam Chandler and Nancy Solla

Documentation Adam Chandler, Glen Wiley, Nancy Solla

In-library IP address lookup tables

Linda Miller and Assessment, Pete Magnus and Desktop Services

logs.library.cornell.edu