R/SQL/scripting, oodles of data, a willing outlet for write-ups. Any takers? Do some good. Intern at Barron’s, the New York financial publication with a decades-long tradition of investigative journalism and a more recent commitment to data analytic exposure of fraud in finance, business and healthcare. Bring us your zeal and your data munging skills and
Blog Archives
How fast is JAGS?
From Martyn Plummer, on the JAGS news blog. Key graph below, showing a few outlying cases in which JAGS is substantially slower than OpenBUGS, but generally, JAGS performs quite favorably. Key point from Martyn: Incidentally, these figures are for JAGS with the glm module loaded. The glm module is not loaded by default. If you
the spatial concentration of Green support
I did some poking and prodding of the polling place data, looking at the way Green support is highly concentrated in the inner-capital cities (and in Melbourne and Sydney, in particular). There are a few interesting exceptions that will make sense to...
mapping the Australian election (2010 edition)
The AEC makes this reasonably easy, as do the authors of some very helpful R packages, the good people at Google Maps etc. Full description here (PDF); entire collection here; a sample here, showing Green 1st preferences, by polling places across metro...
Labor vote share across different types of balloting
We’re starting to get more divisions reporting two-candidate preferred numbers by vote type. The emerging picture (literally) is one in which Labor’s performance on the pre-polls and postals is lagging its performance in ordinary votes. On the other hand, Labor seems to be doing well among absentee voters (the regression line sitting above the 45
informality, the 2010 edition
For the last two cycles I’ve done some simple regression analysis of the informal vote. I saw Possum have his go at it, using a specification that is virtually the same as what I’ve run in the past (2007, 2004). The 2010 edition follows. As usual, electorate-level informality in House of Reps voting increases with
swing graph
I’m updating a swing dotplot PDF every 10 minutes as the count progresses (and the cool part is that the updates continue even as I’m flying Heathrow to SFO).
state-by-state pendulum
By popular demand (!), my state-by-state pendulum (pendula?) for 2010 is up (big PDF), just in time for the election. 550px wide JPG version is inline, below. This follows the same formatting I used in the 2007 edition. We start with the 2PP ALP vote shares recorded at the last election (incorporating changes from electoral
countrycode R package
I am digging this countrycode package in R I just found, written by a UM grad student Vincent Arel-Bundock. Very nice. Thanks! countrycode("2","cown","iso3c") "USA" It would be better if the package could handle failed matches a little more gracefully: e.g., > countrycode(c(2,730,818),"cown","iso3c") "USA" when the user was expecting to get back a vector
JAGS 2.0
Is out. On sourceforge. Along with a new rjags. Thanks Martyn. I’m looking forward to working my way through some of the improvements, which include block-updating of parameters in GLMs more use of R’s d/p/q family of functions for den...
Zero Inflated Models and Generalized Linear Mixed Models with R.
Zuur, Saveliev, Ieno (2012).
