Why do introductory real analysis courses teach bottom up?, from math stack exchange.
On the distribution of time-to-proof of mathematical conjectures, paper by Ryohei Hisano and Didier Sornette.
The Diagram Prize shortlist is revealed; this is the prize for oddest book title of the year. No math books this year, unlike 2010 winner Crocheting Adventures in Hyperbolic Planes and 2009 nominee The Large Sieve and its Applications. (via Marginal Revolution)
Stefan Collini, The threat to our universities, from the Guardian (so it’s about British universities, but probably applies on both sides of the Atlantic).
David Smith of Revolution Analytics on The Uncanny Valley of Big Data, on how predictive technologies are getting good enough to creep people out; see that New York Times article from a couple weeks ago.
Robert Lucky asks is math still relevant (for engineers).
Jordan Ellenberg has a heart-shaped plot of the American electorate obtained by principal component analysis of some raw polling data.
Keith Devlin at Devlin’s Angle on the difference between teaching and instruction. On my more frustrated days as a teacher I find myself saying that what I’m doing is not teaching; Devlin argues that what we do in front of large groups, getting them to know how to solve specific types of problem, is better called “instruction”. He reserves “teaching” for a more interactive process, which is perhaps what I meant as well. (I do find what I do in office hours, working one-on-one with students, to be “teaching”, but my friends will all tell you that I lament that nobody comes to my office hours this semester.)
The gathering of math geeks who are taking over sports, an article on this weekend’s MIT Sloan Sports Analytics Conference. I wonder if I would have mocked this back when I was an undergrad at MIT. Sloan is the business school, and I hung out in decidedly anti-business-school crowds; on the other hand, sports statistics are fun. (Why haven’t I posted about them, you ask? Because I’m mostly a baseball fan, and it’s not baseball season.)
Kevin Lynch and the imageable Boston, from Bostonography.
A profile of Patrick Ball, “human rights statistician”. This New York Times piece from 2008 suggests what sounds like a capture-recapture approach to estimate the true number of crimes committed — if you have two lists of murders with little overlap, there are probably many you haven’t seen, but if your two lists have a lot of overlap you can bet they’re close to complete. (Via simply statistics.)
Becoming an expert statistician, by Ann Maria de Mars.
Readings for an honors liberal art math course, reader suggestions at MathOverflow.