Leaving

Morning

• I have been spending the last few weeks preparing for my iminent departure from CSIRO, as my contract ends today, which is why I haven’t posted here in awhile. It’s surprising how much work it takes to leave a place. So I am now officially looking for another job. A few promising positions to apply for so far. I will see how it goes. Finally signed up for LinkedIn, in case it makes a difference :)

• I discovered that my local WA beach (Scarborough Beach) has free wi-fi access! So I am thinking during my work hiatus, I will be spending some time catching up on manuscripts at the beach, watching the surfers, and enjoying the sea air. I am looking forward to it.

Afternoon

• Did some work on my phylogenetic habitat heterogeneity simulation project (#PGHG). Started writing out the models that underly it http://rdinnager.github.io/PGHG-Project/theModel.

• The new package based project system is working out well so far. A few issues I have run into using Carl Boettiger’s system of putting .Rmd scripts into a gh-pages branch, so that their ‘html’ outputs can be rendered by github. The main issue is when I use a script to make a cleaned up dataset, which I then put into the data directory of the project package. This works great, as long as I remember not to commit the data file into the gh-pages branch, and then switch to master before committing it. The issue comes when I have committed the data file to master but then I realize I need to update and rerun the data cleaning script. When I do so, the new data file created will overwrite the master version if I switch branches from gh-pages to master, and so git will not let me do it. My current workaround is to copy the new data file from data into the package root, then switch branches, and then copy it back into data, overwriting the old version. This is a bit cumbersome, but not too bad, as long I don’t have to do it too often.

• I am considering writing an R script to automatically knit all .Rmd files in a directory, adding a header, and generating an index page. This would mainly be to avoid cluttering all of my project’s gh-pages branches with Jekyll layout files, and also because I think it would be fun to write a static page generator in R. My greater knowledge of R compared with Ruby would allow me to more easily add interesting original features too.