Almost forgot to blog on this: a rare (!) departure off the 19s the other night. There was foul weather at SFO (hail and wind), and so this unusual configuration was in place. We got airborne quickly, just after the 28/10 intersections, and then turned left pretty much immediately for the cross-country flight.

And I really wish UAL skippers would be less paranoid about leaving channel 9 on for us.

I’m solo parenting this week, doing housework while listening to NPR discuss a report they co-sponsored on the distressingly high variation in the quality of death investigations around the United States. Bottom lines: huge state and local-level variation in legislative mandates, oversight, scope, funding, professionalism, training, remuneration of staff, leading to woeful lapses, sloth and even malfeasance in some places.

The phrase “patchwork system” comes up all the time as the journalists are talking, a “jumble”, vastly different laws from place to place…

As a political scientist I’m like (a) “duh“; (b) sounds so much the state of voter registration and election administration.

The “genius” of American federalism, and all that…

The Kaggle people are presenting at Stanford on Wednesday afternoon:

*** EXTRA SEMINAR ANNOUNCEMENT ***

Wednesday, February 9, 2:30pm: Computational Biology Seminar, Clark Center Room S363
(entrance through Peet’s Coffee and Tea)
Anthony Goldbloom, Founder and CEO of Kaggle
“Making Data Science a Sport”

Ahh, today was that special moment in the linear modeling class when I show them the beauty of the orthogonal projector H = X (X’X)^-1 X’ and of course M = I – H. There is something so intuitive about the model fit being the “shadow” of the y vector in the column space of X, and hence that the residual vector is orthogonal to Hy, etc, even I am literally flapping my arms, making vectors in R^3, asking them to imagine the floor as the column space of X. Moreover, understood this way, it is easy to see how HHy = Hy etc… Least squares as an application of the Pythagorean Theorem is a nice teaching moment too.

I own Michael Wichura’s book The Coordinate-Free Approach to Linear Models (Cambridge 2006), but tend to think it is way too tough (abstract/mathy) for an audience of PoliSci PhD and MA students. I point stronger students towards Seber and Lee, say for a reasonably rigorous treatment of where the F test comes from etc.

Off-scale low readings from the BOM’s wind measurement equipment at Willis Island; except to see more of this as the cyclone comes on-shore. Gusts at 185 km/h before the instrument died…or was blown away…