To run the code in the book and complete the questions, you'll need to download R,
rstan and the following R packages (using install.packages()):
coda,mvtnorm,devtools,dagitty, ellipse, and ape.
You also need to install McElreath's R package rethinking by running (from within R):
library(devtools)
devtools::install_github('rmcelreath/rethinking')
Of those steps, installing rstan is the trickest as you'll need to make sure you have a C++ compiler (and configure it for R).
Alternatively, you can run everything in a programming environment called a jupyter notebook via a piece of software called docker (which one can use to configure a lightweight virtual software environment that is standardized between operating systems).
I've set up a docker environments for R and also another for edward2 and tensorflow-probability using python if you'd prefer to use something other than R (you could also use pymc3 for another python alternative).
The first step is to download docker.
Once you have downloaded and configured docker you need to get an image (the virtualisation that has all the software installed).
You can get this by typing the following into your terminal ("Terminal" in mac/linux, "PowerShell" in windows):
docker pull joshuarchristie/statistical-rethinking:r
(If you want the tensorflow version, use docker pull joshuarchristie/statistical-rethinking:tfp instead.)
If you get a permission error then try sudo docker pull joshuarchristie/statistical-rethinking:r and enter your password when prompted.
You can then run a container (a specific instance of the image you downloaded) by typing into the terminal (if mac/linux):
docker run --rm -p 8888:8888 -e JUPYTER_ENABLE_LAB=yes -v "$PWD":/home/jovyan/work joshuarchristie/statistical-rethinking:r
or if on windows:
docker run --rm -p 8888:8888 -e JUPYTER_ENABLE_LAB=yes -v ${pwd}:/home/jovyan/work joshuarchristie/statistical-rethinking:r
and then click on (or copy and paste) the underlined link down the bottom that says something like "or http://127.0.0.1:8888/...". (again use sudo if you have permission issues).
That will launch a jupyter notebook in your default browser where clicking on the work directory from within jupyter will show your current working directory
(i.e. the directory given if you enter pwd in the terminal).
Click on the R button to open a new notebook using the configured R environment.
We can go through how to set this up when we meet if these instructions aren't very clear or if you run into any trouble.
To run the tensorflow image you need to use the following instead:
docker run --gpus all --rm -p 8888:8888 -e JUPYTER_ENABLE_LAB=yes -v "$PWD":/tf/notebooks/ joshuarchristie/statistical-rethinking:tfp
In the r/ directory I've added the code from chapter 2 of the book to chapter2_code.ipynb and the questions to chapter2_questions.ipynb.
If you fork the github repository and start docker from the r/ directory then you can work on these files by navigating into the
work directory in the jupyter notebook.