This is my workflow when working on an R project. It is written in the first person because it is a post to instruct me what to do (I sometimes forget what the next step is…).

Prerequisites

This requires the following software to already be installed:

R
RStudio
Git for Windows (well, just Git, but I typically use Windows, so…)

The installation and configuration of Git will be addressed in another blog post. Once you set it up, it works.

Setting up Git on Windows

I can use RStudio and the usethis package to setup Git after it is installed. Just make sure usethis is installed in RStudio and then type usethis::edit_git_config() to edit your gitconfig file.

Process

I typically use version control so if that is the case here, I need to create a repository for this in GitHub. I can create that repository first, or the create project process can create it if I am creating a new directory. Let’s let RStudio create it (less work for me).

If the directory already exists, I will need to create the corresponding GitHub repository. Again, this can be done on GitHub or RStudio can do it and, again, let’s let RStudio create it (it is still less work for me).

Create a project in R

Create project in R using File -> New Project.... Select one of the directory options, depending on whether I already have the directory created.

If I already have the directory, then go to creating in an existing directory. Otherwise, go on to the next section.

New Directory

If I am creating the folder, the most likely scenario, the select New Directory. I then need to select the project type, which is typically just a new project. Then complete the prompts in the dialog for the directory name, the location where this directory should be created, and then be sure to select Create a git repository but do not select the Use renv with this project. The first option makes sure we have source control set up for this project. The second option makes sure the packages are isolated to this project, but we are going to need some additional configuration before we create the Renv environment.

When you click the Create Project button, the new project is created and RStudio opens it.

Now go finish setting up the GitHub connection.

Existing Directory

If the project folder already exists (typically because I received some data and created a folder for it), then select Existing Directory. Use the Browse button to select the existing folder.

When you click the Create Project button, the new project is created and RStudio opens it.

After the project is opened, I need to create the Git repository for version control. To do this, enter usethis::use_git() in the Console. This will create and initialize the local Git repository, allow you to commit the existing files, and restart RStudio to start using the Git repository.

Setting up the GitHub connection

Even though I have a Git repository, it is not synchronized with an online repository in GitHub. To do that, we need to enter usethis::use_github(private=TRUE) to create a private repository in GitHub that is linked to this local repository. I prefer my projects to be private since I am working with campus data. :) The repository in GitHub that is created will have the same name as the project.

Now I can proceed to setting up the Renv environment.

Setting up the Renv environment

Before setting up Renv, I need to edit the project’s .Renviron to set some variables used by Renv. In the Consle, execute usethis::edit_r_environ('project') and enter the following data into this file:

RENV_CONFIG_PAK_ENABLED = TRUE
RENV_CONFIG_MRAN_ENABLED = FALSE

If I need to ensure all the packages are local, then I need to add RENV_CONFIG_CACHE_SYMLINKS = FALSE to the .Renviron.

After the .Renviron file is saved, restart the R session.

Now I can setup the Renv environment by executing the following code in the Console.

# You can change the date to whatever date you want for
# the checkpoint. You would typically use today's date.

today <- Sys.Date()
pak::repo_add(CRAN = glue::glue("MRAN@{today}"))
renv::init()

# If you are developing a package for use in Power BI,
# then you need to use Microsoft R Open 3.4.4 and 
# an MRAN date of 2018-04-01.
# pak::repo_add(CRAN = "MRAN@2018-04-01")

Now we are ready to code.

Setting up the code environment

I need to install the necessary packages for my coding environment. There are a few that are common across all projects, and some are dependent on what I am trying to do or what kind of data I have.

The base packages can be installed by executing the following commands in the Console:

renv::install("tidyverse")
renv::install("data.table")
renv::install("dtplyr")

I probably also need the local package to access the campus database. I can install that with renv::install("haywood-ierg/haywoodcc").

Working with files

If I am working with files, I will need the here (renv::install("here")) and fs (renv::install("fs")) packages. If using here, I will need to create a UUID by calling uuid::UUIDgenerate() and passing that as a parameter to here::am_i(uuid = UUID_GOES_HERE). The here package allows me to reference files based on the root of the project no matter where I reference them. For example, here(path("data","data.csv")) refers to <project_dir>/data/data.csv no matter where the source file lives within the project.

Working with Excel files

If I am working with Excel files, I will need the rio package (rev::install(rio)). After install rio, execute rio::install_formats() in the Console.

A collection of library calls

Here is the full collection of library calls. I will need to uncomment the ones I need for this project. I can delete the rest.

library(data.table)
library(dtplyr)
library(tidyverse) # This loads ggplot2, dplyr, tidyr, readr, purrr, tibble, stringr, and forcats 
library(dplyr, warn.conflicts=FALSE)
library(magrittr)
library(glue)

# Load this if working with CCDW_HIST database
library(haywoodcc)

# Load this if working with dates
# library(lubridate)

# To work with files on the file system, load these
# library(here)
# library(fs)
# Then you can load a CSV file called data.csv in the data folder by using the following
# data_csv <- read_csv(here(path("data","data.csv")))

# To work with Excel files, load this
# library(rio)