So far I’m mostly been moving the sand in the box, seeing what I can build, what things can do, what happens if I kick the stuff around. And every time I find that the sand won’t do something, the admins come along and plug in a new shovel so I can make the sand do what I want it to do.

It’s time to start thinking seriously about mining data off the back end of the project.  (I’m just getting started.)

MySQL and Usage Stats

The back end of the web is, I must admit, one of the places where I’m still quite unfamiliar. It’s not a dark alley I fear so much as a space I haven’t explored enough – yet. I can upload stuff, make some mods to files on the server, etc. But manipulating and culling material out of a database is not my thing at all. And so I have some serious questions about what CAN and what CANNOT readily occur back there.

I’m thinking now about stats on my own usage of the AC over the last week or so.  Can the admin quickly pull that data to provide stats on such hypothetically meaningful information as:

• How many posts have I made in my blog?
• How many files have I uploaded?
• How many times have I added a page, made an edit, etc?
• How many comments have I made on others’ blogs?
• How many friends to I have?
• How many times have I logged into the AC?
• How many edits have I made to the Mediawiki install?
• How many widgets/plugins/etc have I activated? How many themes have I tried out?

Why does this matter?

I’m seeing that e-portfolio adoption is about much more than simply getting students to reflect on and upload work samples in their classes.   I think it is important to start to think about e-portfolio in terms of usage.  I’d like the system to be able to report, for example, that 250 students created 1200 pages, and commented on 400 posts in an academic year; that students, on average, friended 39 peers over the year; that students change their themes 7 times in a year on average, and that they add 4 widgets/plugins/etc.

MySQL and Portfolio Samples

At WordCampEd there was some serious discussion of assessing e-portfolio samples over time, with Bret Eynon raising perhaps the single most important assessment issue of the day:  If students can freely add to, subtract from, and edit their e-portfolio elements at any time, an effort to assess development over time likely to be highly problematic.  One possibility involves archiving e-portfolios each year, something that I think Joe suggested.

The possibilities here are fascinating.  Imagine not only evaluating a student’s writing and reflection at Year 1 and again at Year 4; imagine evaluating a student’s level of engagement with the platform, and level of engagement in the community.  NSSE surveys students about engagement on campus; this would potentially measure engagement on a whole different dimension.

My question for now is far simpler, while still being very important.  Is it a straightforward database query to pull, say, a random sample of 5% of the final papers from a freshman writing cohort, together with the students’ reflections on those samples?  (I imagine that the reflections would be in a Page, and the papers would be word docs or PDFs attached to the page.)