Anaheim, Calif. —Amazon. Netflix. Google. All personalize recommendations based on what they know about users.
A new project, unveiled at the Educause conference here today, plans to provide college students a similar experience on academic Web sites.
It’s called Sherpa, like the guides who lead climbers up Mount Everest. The goal of the software, developed by the South Orange County Community College District, is to mine data about students to guide them to courses, information, and services.
That’s a change from what students experience starting and finishing classes on the Blackboard course-management system, said Robert S. Bramucci, South Orange’s vice chancellor for technology and learning services.
“It’s as if Blackboard is somebody with hippocampal damage, that has severe amnesia,” he said. “It’s never seen you before, other than knowing that you have an account in the system. The systems outside learn about you. But the systems typically in academia do not.”
So how will Sherpa use its knowledge? A new student might get a link to the online orientation. A student with a high grade-point average might get a link to the honors program. A student with low grades might be pointed to tutoring services.
With more information about students, the suggestions could get much more specific. Jim Gaston, South Orange’s associate director for IT, academic systems, and special projects, gave the example of a tip he hopes to be able to send to a student who hasn’t yet registered for class:
“Your classes are filling fast. We looked at your academic plan and saw that you plan on transferring to UC Berkeley as a biology major. We searched the class schedule. We found a set of courses you said you were interested in. Based on the pattern of classes you’ve taken in the past, here are the four classes we think you’re going to be most interested in. We’ve already screened them for pre-recs. They don’t have a time conflict with when you said you were going to work. And one of them is your favorite instructor.”
Signs suggest that this kind of data-mining will grow in academe. Universities like Purdue have used computer analysis to predict dropouts, for example. And learning analytics are one of the thrusts of a new grant program announced this week by the Bill & Melinda Gates Foundation.
6 Responses to Like Netflix, New College Software Seeks to Personalize Recommendations
arrive2__net - October 14, 2010 at 12:49 am
I think having a system that makes useful suggestions, and points out possible risks of missing deadlines, could benefit the users … as long as it does not become obtrusive, and there is a open and consistent privacy policy so the user is aware of how widely the information could be distributed. The student/user should be enabled to opt out of Sherpa.
Bernard Schuster
Arrive2.net
jacobi1804 - October 14, 2010 at 8:14 am
I like it. It might be better received if it weren’t as anthropomorphic as the tip describes. Facebook and Gmail users seem to not care that their data is mined for the purposes of serving ads. But when Amazon says “Since you liked X, you’ll probably like Y” people get creeped out.
runningtune - October 14, 2010 at 10:42 am
As long as it does not replace advisors, sounds like a good idea.
josephofoley - October 14, 2010 at 11:26 am
This appears to be an excellent approach to reducing the clerical burdens on advisers while providing useful information to students. Rightly used, it should promote humane interactions among humans.
tpul2014 - October 14, 2010 at 12:51 pm
Netflix is TERRIBLE at recommending films based upon those that I have viewed in the past.
abcde1234 - October 14, 2010 at 1:12 pm
Excellent! Excellent! More progress in subbing out functions that used to be performed by faculty. Less and less need for faculty to interact with students! Now they can just go about their business increasing grant revenues for the university! They may never have to see a student again! And I have every confidence that the administration will pass off the savings from this efficiency increase to the students, by lowering tuition and feed. Rather than, say, building a new research building with an increased overhead.
More like this!