Publications by Bluecology blog
R course in data wrangling and spatial analysis notes are online
R course in data wrangling and spatial analysis notes are online I have posted the notes to our new R course in data wrangling and spatial analysis notes on this webpage. The course is aimed at biological scientists who have some basic knowledge of R, but want to learn to use it for a broader variety of tasks. It is aimed to show how you can use ...
1662 sym 2 img
Diary of becoming a Bayesian
Diary of becoming a Bayesian Here’s my timeline of going from frequentist stats to becoming a fully fledged Bayesian eco-statistician. 2006 Read the Ecological Detective (Hilborn and Walters), the philosophy makes sense, the history of falsification is interesting but, geez, the maths looks hard. 2008 Started PhD. Peers also interested in Bayes...
3661 sym
R courses February 2020
R courses February 2020 Our next series of R courses for professionals and graduate students is now open for registration. The courses include 1-day for absolute beginners, 3 days for intermediate material and 3 days for the advanced class. They are run 18-21 February 2020 at the University of Queensland in St Lucia, Queensland. The courses are t...
1487 sym
GLM tip getting non-linear with splines
GLM tips: get non-linear with splines This tip is great for a quick non-linear test, before you go all the way with a GAM or parametric non-linear model. You’ll need the splines library, which comes shipped with R anyway. First, let’s make up a bit of count data. The underlying ‘true’ model will be poisson (think count data) with a log li...
3706 sym R (1251 sym/4 pcs) 6 img
Informing priors for a time-series models to detect declines in tiger sharks
Informing priors for a time-series models to detect declines in tiger sharks Our new research has revealed a 71% decline in tiger sharks across Queensland’s coastline (open-access version). In this blog I will explore the time-series modelling techniques we used to estimate this trend. (I’ve written about the significance of this decline else...
5486 sym 4 img
Make a map of your study site with tmap
Make a map of your study site with tmap I’m loving the tmap package right now. Here’s why. And we’ll look at how to make a clean map of a study site as we go. We’ll use the data from our study of pollution and coral reefs in the Solmon Islands (repo is here). It runs smoothly with sf sf (‘Simple Features’) is the new spatial data stan...
2288 sym R (1258 sym/8 pcs) 4 img
Using R and power analysis to inform experimental design
Using R and power analysis to inform experimental design A collaborator once asked me: “can you do that thing where you take a little bit of data and model it up to make lots more data?” Does it sound suspicious to fake a lot of data from a little bit of data? It depends on the context. Making up data is totally appropriate if you want to do ...
8054 sym R (1855 sym/9 pcs) 4 img
Three ways to calculate distances in R
Three ways to calculate distances in R Calculating a distance on a map sounds straightforward, but it can be confusing how many different ways there are to do this in R. This complexity arises because there are different ways of defining ‘distance’ on the Earth’s surface. The Earth is spherical. So do you want to calculate distances around ...
4562 sym R (1516 sym/11 pcs) 10 img
Spatial predictions with GAMs and rasters
Spatial predictions with GAMs and rasters One powerful use of GAMs is for interpolating to unsampled locations. We can combine GAMs with raster package to conveniently predict a GAM model to places we have not got data. Simulate some spatial data We’ll simulate some spatial data based on rasters. There are two spatial covariates, x1 and x2. We ...
3504 sym R (3773 sym/10 pcs) 14 img
A case against pipes in R and what to do instead
A case against pipes in R and what to do instead Pipes (%>%) are great for improving readibility of lengthy data processing scripts, but I’m beggining to learn they have some weaknesses when it comes to large and complex data processing. We are running a number of projects at the moment that require managing and wrangling large and complex data...
5060 sym R (2931 sym/16 pcs)