Next to last class. Thinking ahead ten years, there’s no point trying to squeeze in any more “material”.

This is what I suggested in email to the class:

One good way to review for the final: find one or two problems (maybe from homework) that you are on the edge of understanding (you can almost get right). Ask about them in class.

Note “on the edge” – it’s too late to try to learn something you’re absolutely clueless about, and a waste of study/class time to think about stuff you know well.

Let’s see what they come up with.

I should have graded/returned papers today, but forgot and did the xword/sudoku on the T instead. Sorry.

Worked the regression problem on US energy consumption. The class went well or badly depending on your feelings about watching the teacher make mistakes and recover from them. The recovery is an interesting process that should be made explicit, but boy is it confusing for the students.

In this case we prefaced constructing the regression line with a discussion of why energy usage increased in the last half century – population growth and more gadgets. To understand the former we computed energy use in Twh/person. The first attempt gave something like 100 kwh per person per year – which was (to me) obviously ridiculously small. It’s just about enough to keep a 100 watt bulb on for 1000 hours (3 hours per day). So I went looking for three more zeroes. First had to check the Tera was 10^12, not 10^15. Then found the zeroes in misreading a `’` for a `.` in the data. Then the correct calculation gave an answer that was several orders of magnitude too big – the answer is probably in the meaning of “domestic” energy consumpion. That might be “household” but more likely is “United States” and counting all forms of energy, converted to Twh.

Thinking these things through is hard – and really only possible if you have a kind of confidence that the students don’t – for perfectly good reasons.

Finished with useful Excel work drawing the trendline. Boring by comparison.

blog home page