Some advice for my SEB113 students who may struggle with the workload of first semester university comes from the sagest of equines, @horse_ebooks.Hugh Possingham‘s talking on Monday about the mathematics and economics of conservation as part of the BrisScience seminar series. I’ve been meaning to make it to a Possingham talk for a while.

On an unrelated note, another UPTECH paper has been published. This one’s looking at spatial variation of particle number concentration in the school environment. Congratulations to Farhad Salimi, the first author of this paper, on the publication of his first paper. Farhad’s one of the PhD students on the UPTECH project and is due to finish his thesis later this year. I’ve worked with him on two of his papers (this one and another which has been submitted) and he’s really thrown himself into learning how to use R. This has not only made it easier for me to collaborate with him but it’s also made his analysis possible.

Ever wondered what it’s like to spend two months on an ice-breaker?

The lecture slides from Harvard’s Stat 221 course are all online now.

Commander Chris Hadfield’s farewell from the ISS is, fittingly, a cover of David Bowie’s “Space Oddity”.

I’ve been excited about the PLoS journals since their launch and I believe QUT is a subscribing member, which means our publication fees are covered. It’s one of the best Open Access journal groups around and doesn’t appear to be a cash grab like some other publishers who are attempting to use Open Access as a business model to increase profits rather than because they believe in the free dissemination of research.

UPTECH collected fungi and endotoxin data at the 25 schools, and we’re about to submit the fungi paper (which means work must continue on the endotoxin paper). I was considering whether we should submit to PLoS One (IF 2011: 4.092) and then had a look at what other journals they have which may be an appropriate home. I really think once we get the clinical data from our Southern collaborators we should aim to do the best statistical modelling we can. I’m heartened by the fact that the head of the clinical group we’re working with has a strong background in stats and a desire to learn more Bayesian statistics. I don’t know if we can pull it off, but the prospect of having something investigating the role of fungi and endotoxin on child health published in PLoS Pathogens (IF 2011: 9.172) is exhilarating.

I’ve managed to pick up a few little statistical novelties along the way such as nonparametric Bayes, hierarchical linear models and some of the theory behind GMRFs and Gaussian Processes but I feel like I’m lagging behind where I want to be. This could be a consequence of being based in a group with no other statisticians. Were I doing my postdoc in a statistics group I’d be more deeply immersed in a group with a culture of doing statistics research rather than doing scientific research which requires statistics I already know.

Randall Munroe made a chart that helps you figure out whether you’re spending more time working on your time-saving technical trick than you’d actually save.

SAMSI hosted a talk about the public health impact of climate change. Wish I could have been there.

A few (somewhat) recent posts from Andrew Gelman regarding p values

• reporting probabilities instead of p values is very Bayesian
• p values and statistical practice

I find myself agreeing with a lot of what Gelman says on p values. I think they put the focus of statistics on testing rather than modelling, which is part of why we end up with scientists and others with such poor statistical skills and next to no creativity in terms of the models they fit.

Further information can be found in the three documents below.

• CAR 2013 Scholarships_Information
• PhD Scholarships CAR 2013 Call for Applications
• PhD application form CAR 2013.docx

There’s been a bit of buzz recently about how two economists really dropped the ball when they published a paper showing that high levels of national debt as a percentage of GDP slows economic growth. I won’t write any more about it because it’s been done by better writers and statisticians, but I’d like to bring your attention to the following articles:

• The original working paper (National Bureau of Economic Research)
• Replicating the results (Political Economy Research Institute) was only possible by making huge methodological and data selection mistakes (Next New Deal)
• This may have led to massive avoidable unemployment around the world (The Guardian)
• The authors should gracefully retract, admit they were wrong and move on (Andrew Gelman’s blog)
• And to use R (Phil at Gelman’s blog) rather than Excel (Andrew Gelman)
• There are many reasons to not use Excel, such as it providing the wrong answer for paired t-tests and being terrible at generating Normal random numbers.

I wrote this in LaTeX, which generates pdf files. Therefore, I do not have a Word document for this. Apologies for any inconvenience this may cause.

I’ve been working with this collaborator to get them using Github, R and LaTeX . There was the denial, the anger, bargaining, depression and ultimately acceptance. Watching this collaborator send this email I felt like a proud father watching their child graduate.

As I got further and further into my PhD I saw extensions for R-INLA being written thanks to a few visits from, and email correspondence with, Dr Daniel Simpson, and the help list on the R-INLA site where Dan, Håvard Rue and Finn Lindgren are very quick with a reply.

A few days ago I got an email from Rue telling me he’d been made aware of one of my thesis papers and if I wouldn’t mind having a look at running it with the new testing distribution of R-INLA. It’s the first time I’ve looked at the code again since submitting the paper for publication and it seems that an awful lot of work has been put into internal optimisation. The code for running my model requires less manual tuning now and I’m excited about using it in follow-up papers where I’ll be looking at more of the UPTECH data.

There’s also a new tutorial for spatial modelling with INLA, written by Elias T. Krainski, which covers a number of topics such as a simple spatial regression, a spatial model with misalignment and non-stationary spatial models (which I’ve seen talked about a few times but there’s very little documentation about them).

I think R-INLA, particularly the spatial modelling, has really come a long way over the last few years and it’s encouraging to see it being taken up at QUT where students would probably have used WinBUGS in the past. While there are some limitations in terms of the flexibility of the classes of models that can be fit in BUGS versus R-INLA I’d much rather do any spatial, spatio-temporal or non-parametric smoothing in R-INLA.