In very rough ordering.
Andrew Gelman’s blog often sparks up interesting discussions in the comments. Some theoretical, some applied, with the latter usually being in political science, as that is Gelman’s main work topic.
John Cook’s blog, The Endeavour, is another source of interesting asides. He also has some nice charts, such as a diagram of common conjugate prior relationships.
Normal Deviate, a discontinued blog by Larry Wassermann that focuses on the interface between Statistics and Machine Learning, two subjects with the same goal, but widely differing methodology. Good posts include an informative look at Simpson’s Paradox, non-informative priors being a lost cause, and a summary of some of Stone’s papers about flat improper priors. One of the examples used in those papers is one I disagree with, for reasons I might expand on at some point, but a comment in one of the later papers suggests other had already suggested this to Stone.
Christan Robert touches mostly on Bayesian and computational topics. One of the more prominent figures in my field.
Matt Asher: “In Monte Carlo We Trust”. Mostly posts of what’s happened that week in statistics. Asher recently put up a test episode for a probability podcast discussing the Problem of Points from the early history of probability theory, hopefully this will become a regular feature.
Philosophy of Statistics
Deborah Mayo, “the frequentist in exile”. Philosophy of statistics with an aggressively classical outlook.
William M. Briggs, “Statistician to the Stars!” Reading Briggs can be infuriating for a lot of people, because he puts his beliefs in plain sight, particularly those on climatology. He also has a rather rambling style of writing, so sometimes his original point can get a bit lost. The original point behind his Statistics posts, however, is often sound. Briggs writes about a range of things, but his most common theme is simply that the way statistics is often done leads to far too much over-certainty. The short series Selling Fear is a Risky Business (I,II,III) is probably a good place to start, as are a lot of the posts where he
Michael Flynn is a former statistician, now a hard science fiction author. He occasionally touches on Statistics – often from the point of view of classical Quality Control theory – or on the history of science. A good example of the latter is The Great Ptolemaic Smackdown, his series on the transition from egocentrism to heliocentrism in astronomy.
John Baez recently became interested in ecology and climate, but has a strong background in Mathematical Physics and Category Theory. This makes a lot of his writing beyond my comprehension, but what I can understand is beautiful. Also well-known for his old column, This Week’s Finds in Mathematical Physics.
Brent Yorgey focuses on simple, but beautiful, mathematics, and is rather deft with the graphical side of things. Noteworthy for recently creating a large poster that draws numbers according to their prime factorisations, that’s better seen than described.
The Renaissance Mathematicus is a blog on the history of science. Two common themes are 1. stop portraying famous scientific figures as lone geniuses, and 2. stop asking scientists for the history of their subject, instead of asking historians. And certainly don’t ask comedians about the history of maths. Engages in a lot of taking people to task for writing science as a mythology full of lone heroes. Galileo wasn’t a poster child for Science versus Religion, Newton didn’t suddenly understand gravity because an apple fell on his head, Ada Lovelace wasn’t the first computer programmer, and so on.
Matt Might is a Computer Scientist, but writes a lot of articles useful for anyone who is in academia, or needs to grapple with programming.
My Gravatar image is from a paper on nomograms, hosted here. Nomograms are static diagrams, that were used for quick calculations before computers rendered them all but obsolete. Aside from the Dead Reckonings site associated with the above site, an overview of nomograms, as well as a lot of other stuff, can be found at Winchell Chung’s website.