So, my data wasn't all broken - about 60% of it was good. However, to get good values for the broken bits easily, I had to change the testing framework, which made it possible that the new data would be incompatible with the old data.
So, after a ton of bug fixing and batch file writing (and about 4 and a half hours of waiting), followed by a little more debugging, some batch file debugging and another half hour of waiting, I have all new test data - which gives me very similar results to the last time. Oh well - such is life. At least the new data is correct.
Overall, getting the data I need has taken me 3 days - which puts me a little behind. I'm doing fairly well, 'though - 4800 words down, which means I've approximately 5200 to go. I may end up a little short of that, but that's ok, so long as I've said what I need to.
On the positive side, I can afford to get the document bound, and I was assured that if I take it in on tuesday afternoon it will be ready for wednesday. I'm going to get two of the cheap version, for submitting, and 2-4 of the expensive version, for relatives.