Skip to contents

Setting up a workflow in R

r-workflow.Rmd

Working directory

In one of the previous steps, we:

  • Created a folder for the project of translating a microsimulation EPL/devCEQaux.
  • Cloned example microsimulation in Stata from https://github.com/wbEPL/Example_FiscalSim to the folder EPL/devCEQaux/Example_FiscalSim/.

Now, we need to create a project for our analysis in R, which will subsequently become a place, where we develop our R Shiny application with the microsimulation.

Creating an RStudio project

In your root folder, where you perform translation of the analysis to R and execute Stata code, (in this example, this is EPL/devCEQaux folder) create an R Studio project.

  • Familiarize with RStudio project basics here if needed R4DS. Workflow: scripts and projects.
  • Use a short and concise name for the project. (It must not contain spaces and be short, up to 20 characters long).
    • For example, use the same name as the Stata project has and add “_app” to it. The result is: “Example_FiscalSim_app”. In this way you will discriminate where is the original Stata project and where is the R version of it.
  • The result of creating the RStudio project EPL/devCEQaux/Example_FiscalSim_app. is the following folder structure:
    • Note. In the R Studio project folder, there is a file Example_FiscalSim_app.Rproj. Double-clicking on this file will start R Studio for you.

The result is the following folder structure:

./EPL/devCEQaux
├── Example_FiscalSim
│   ├── 01.Data
│   │   └── ...
│   ├── 02.Dofile
│   │   └── ...
│   ├── 03.Tool
│   │   └── Example_FiscalSim.xlsx
│   └── README.md
└── Example_FiscalSim_app
    └── Example_FiscalSim_app.Rproj

Organizing files in the R Project.

The general principles of file organization in the R project are:

  • R folder is only used for functions, not scripts.

  • data-app folder for data that is used inside the shiny app (input structure data file, pre-simulation data, and baseline simulation results).

  • dev folder is the one, where we place our scripts. Here we usually develop such scripts as re-saving pre-simulation data, performing manual quality assurance routines, and developing and validating the simulation code in R.

  • inst/app folder exists for the auxiliary shiny-app related files, such as info-page.md markdown document of the shiny app landing page.

In the root directory of an R project, there are:

  • .gitignore file relevant to git.

  • project_name.proj is the file used to launch R Studio in the current project.

  • DESCRIPTION internal file relevant to the R package development.

  • app.R key r script, from where we launch our shiny app.

Now, create the folder structure as suggested above (not the files). The result should roughly be the following:

Creating and saving the first script in R

Now, let us create the first scrip in R and save it under the name example-script.R in the dev folder. Try to follow the steps from the animation.