Date: October 31, 2012, 10:30am
(This is the last session I have notes for. After this, the material was just not as relevant to my needs or the needs of my smaller college).
I. "Mining Library and University Data to Understand User Populations and Behavior."
- (Out of the three presentations in this batch, this was the one that interested me and seemed to have ideas I can adapt in my setting. Reminder to self: talk to the IR folks back home).
- This project was done at the University of Miami.
- In lieu of surveys, try to use existing campus data and data mining.
- Data mining uses large data sets or multiple sets to seek out patterns. Form partnerships with other campus organizations that use large data sets.
- Sample questions/things to ask:
- Does book use change over time?
- Who is using the library?
- You may need to secure IRB approval for some things. Collect data from registrar, library, HR, and student activities.
- Clean the data by replacing university ID's with project ID's.
- Data was put into a "restricted" vault area for privacy concerns.
- In seeing patterns while data mining, you may get more questions.
- For collection development, checked days between catalog date and date of last checkout for materials acquired in a year.
- Correlated library use to GPA. (Note they have a turnstile that scans IDs at their library, so this may not be as easy or realistic to do in other places). For their turnstile data, students do scan their IDs to enter the library, which allowed them to match ID data to registrar data.
- Use ILL data correlated to instruction, maybe via papers' analysis, to see what the students are borrowing.
- From the Q&A:
- Is GPA correlating problematic given issues of grade inflation? GPA is the prominent measure of student success (i.e. it may be problematic, but it is the measure we've got).