setup.qmd

---
title: Setup
---

You will need to have a running R environment - R can be downloaded for 
Linux, Mac, and Windows from [here](https://cran.r-project.org/).

We use [renv](https://rstudio.github.io/renv/index.html) to manage package 
dependencies. Installing `renv` and running `renv::restore()` should provide you with an 
environment suitable for running this model.

```{r}
#| eval: false
install.packages('renv')
renv::restore()
```

We use the [targets](https://docs.ropensci.org/targets/) package to manage the 
computational pipeline. Targets essentially ensures that computations are
only repeated when downstream dependencies or code are changed. 
For an introduction take a look at the [targets manual](https://books.ropensci.org/targets/)

You may also need to install `cmdstan` - it will be installed via
`renv::restore` but there it needs an additional step to install the 
toolchain.

```{r}
#| eval: false
cmdstanr::install_cmdstan()
```

Once `cmdstan` is installed you can run the pipeline

```r
library(targets)
tar_make()
```

This will take 15 or so minutes - the modelling portion will spit out a lot of warnings.

When the pipeline is complete you can load individual _targets_ using `tar_load()`, e.g.

```r
tar_load(polls)
tar_load(model2023_summary_model2023)
```

All the targets are defined on the [pipeline page](pipeline.qmd).

We also make use of the [stantargets](https://docs.ropensci.org/stantargets/index.html). 
The [`tar_stan_mcmc` function](https://docs.ropensci.org/stantargets/reference/tar_stan_mcmc.html)
creates a few different automatically named targets - including `model2023_summary_model2023`.