Programming course for people who know some statistics already.
"Decide on a sample size in advance and wait until the experiment is over before you start believing the “chance of beating original” figures that the A/B testing software gives you."
Mostly about detecting cheating in chess through extremely crafty statistical methods. Fascinating stuff -- although you probably need to know more about chess than I do to really appreciate it.
Fancy statistics package for Python. (Not quite as fancy as R, but then again you also don't have to write in R to use it...)
With lots of handy guides for scientific data processing with Python. A good starting point.
Having had a play with this, I can see why everyone's so enthusiastic about it. It even has an XKCD mode.
"PyTables is a package for managing hierarchical datasets and designed to efficiently and easily cope with extremely large amounts of data." Might be overkill for my temperature sensors!
Neat trick: this uses some LLVM instrumentation to shuffle memory layout around in a program while it's running, to randomise the effects of layout on performance. As a result of the central limit theorem, this tends to normalise the distribution of timing errors too (provided your program runs long enough to have been thoroughly shuffled).
Review of "Ten ironic rules for non-statistical reviewers". Read the original paper first, since it's got some good points -- particularly on exactly what the limitations on normality are, and why you need to be careful about very large studies -- but it probably overstates its case a bit, as this review suggests.
Tomas and Richard are involved in this project for empirical measurement in CS. Their position paper would be sensible reading for students; it explains some of the common pitfalls of performance measurement.