I'm at the early contemplation/design stage of a potential big-data project. The idea is to use the data stored deep inside WW to get quantitative insights into the degree of difficulty of my problems, and/or the work habits of my students.
Before tackling this myself, I should check: has someone else undertaken something similar before? If so, any pointers would be most welcome. If not, here below are two questions on which I would appreciate expert opinion.
Background: The WW server at my University is run by strangers in IT Services, who see me as just another "Instructor"-type user of the usual web front-end (also with LMS integration). But I can ask them for reasonable favours.
I already know how to harvest each student's score on each question, and the number of incorrect answers they have submitted, using the "Scoring Tools" feature of the web interface. This already gives me some clues about which questions are easy and which are hard.
Is there any way to get more granular info? E.g., I know that WW keeps a (timestamp,answer)-pair for each and every student submission. I'd like a way to get these -- ideally from the web interface. Is that possible? If not, is the whole idea futile without admin rights on the server? Or, is there some simple intermediate alternative like, "Ask your IT shop for the sqlite database file with the following standard path and name, and apply sqlitebrowser"?
As suggested in the preamble, I'm not sure what will come of this. If you have relevant info at your fingertips, please do share; but please don't let my half-formed plans distract you for more than a few minutes.
Thanks and warm regards, Philip