Author Archives: phewson

Motorcycle fatalities in the GB (excluding Wales)

Posted on February 2, 2016 by phewson

(Sorry Wales, I just haven’t figured a way of sorting out the population histories given the local government reorganizations that went with Devolution)     These two figures show Lexis diagrams for (a) the number of reported motorcycle fatalities in … Continue reading →

Posted in Uncategorized | Leave a comment

Lexis Diagrams for Road Injuries

Posted on February 2, 2016 by phewson

This is based on the crude road injury rate (all injuries), and a reconstructed population up to 90 for England and Scotland from 1985 to the end of 2014.   The colors show the injury rate (per capita), for all … Continue reading →

Posted in Uncategorized | Leave a comment

Barnard’s views on statistical education (from

Posted on September 23, 2015 by phewson

I really enjoy reading Deborah Mayo’s blog.   I don’t really understand enough about the underlying philosophy – and even gently reading about ideas such as Popper asking for severe challenges (rather than out and out falsifiability) bring this home. … Continue reading →

Posted in Uncategorized | Leave a comment

Regression notation

Posted on October 2, 2013 by phewson

Aaargh.   A A R G H. I like David Freedman’s book on Statistical Modelling.   I like it a lot, and I particularly like the clarity.   It’s very clear that is a population parameter, a fixed but unknown constant (OK, so … Continue reading →

Posted in Theory I should know about | Tagged , , | Leave a comment

Developing code

Posted on September 14, 2013 by phewson

The wonder of the blogosphere is that I rediscovered a paper I had been meaning to think about for a long time: the paper reports Cook, Gelman and Rubin’s proposals that we: Specify a model and then simulate parameters from … Continue reading →

Posted in pymc, Uncategorized | Leave a comment

Emerging thoughts on the Poisson as a limit of the Binomial

Posted on September 14, 2013 by phewson

I am trying to set down some thoughts on the relationship between the Binomial and the Poisson. The specific problem emerges from modelling data where I assume and . The specific problem is that I assume that  and  are correlated in … Continue reading →

Posted in Theory I should know about | Tagged , , , | Leave a comment

Python Notebook

Posted on June 17, 2013 by phewson

It has been a busy year, for various reasons. However, one loose end that urgently needs tidying up are an assortment of bits and pieces relating to the use of python as a computing language. Somewhere during the year the … Continue reading →

Posted in iPython notebook, McMC | Leave a comment

My Mean Square Error Confusions

Posted on September 18, 2012 by phewson

Mean square error is defined as . To be defined, it requires that is fixed, constant (albeit unknown). In other words, it’s a very frequentist thing, but it’s such a dominant way of evaluating the performance of estimators that it … Continue reading →

Posted in Uncategorized | Leave a comment

Mark Glickman’s Music Page

Posted on July 4, 2012 by phewson

I’d really like to go to “Valencia” sometime, if nothing else just for the music.   There are some great song parodies here:  Mark Glickman’s Music Page.   Seems Bayesians have all the fun here.

Posted in Bayesian philosophy | Leave a comment

Multilevel Ordinal Logistic Regression in pymc

Posted on June 11, 2012 by phewson

I think I’m finally getting closer to having the code to fit a proportional odds logistic regression model in pymc.   There are still a few little places where the code could be tidied up (I’ve yet to figure out how … Continue reading →

Posted in McMC, pymc | Tagged , | Leave a comment