Why?

• Keeps all related files together in one folder

• Sets working directory so file paths can be relative to project

• Facilitates use of RStudio Git integration

How to create an RStudio project

• In RStudio, go to File -> New Project…

• Create a new project or add a project to an existing directory

Opening an RStudio Project

Folder structure

• Where possible, store all files related to your project within your RStudio Project directory

  my-project/
      ├── code/
      ├── functions/
      ├── data/
      ├── lookups/
      ├── outputs/
      └── my-project.Rproj

Why?

• Loading all packages in one place makes it easy to see what is required to run the code

• Defining all variables / parameters in one place reduces the number of places in the code you need to edit to make a change

Example

# Read in data for current and previous year

schools <- readr::read_csv(
   here::here("data", "2023-03-31_schools.csv")
)

library(readr)

schools_prev <- read_csv(
   here::here("data", "2022-03-31_schools.csv")
)

Example - Setup script

00_setup.R

# Load packages ----

library(here)
library(readr)
library(lubridate)

# Set parameters ----

month_end <- ymd("2023-03-31")

prev_year <- month_end - years(1)
  

Example - Sourcing setup script

01_read-data.R

# Run setup script ----

source(here::here("code", "00_setup.R"))

# Read in data for current and previous year ----

schools <- read_csv(
   here("data", paste0(month_end, "_schools.csv"))
)

schools_prev <- read_csv(
   here("data", paste0(prev_year, "_schools.csv"))
)
  

Why?

• Makes code easier to write

• Makes code easier to understand

• Makes code easier to debug

Tidyverse style guide

• https://style.tidyverse.org/

  • Naming files

  • Naming objects

  • Spacing

  • Using pipes (%>%)

  • and lots more

Example

• Bad practice:

do_something_very_complicated(something="that",requires=many,arguments="some of which may be long")

lintr

lintr runs style checks on existing code and reports back any issues (but doesn’t change any of your code).

code.R

do_something_very_complicated(something="that",requires=many,arguments="some of which may be long")

lintr::lint("code.R")

[infix_spaces_linter] Put spaces around all infix operators.
[commas_linter] Commas should always have a space after.
[infix_spaces_linter] Put spaces around all infix operators.
[commas_linter] Commas should always have a space after.
[infix_spaces_linter] Put spaces around all infix operators.
[line_length_linter] Lines should not be more than 80 characters.

styler

styler re-styles existing code following the tidyverse style guide

code.R

do_something_very_complicated(
  something = "that",
  requires = many,
  arguments = "some of which may be long"
)

Why?

• Help somebody else understand the code you’ve written

• Help future you understand the code you’ve written

• Provide context

• Section your code to provide structure and easier navigation

How?

• Start a comment line with a single # followed by a space

  # My first comment

Why?

• Minimises repetition

• Simplifies code

• Ensures consistent methods across code

• Easier to maintain

Example

• Calculating financial year of a date

functions/fin_year.R

fin_year <- function(date) {
  m <- lubridate::month(date)
  y <- lubridate::year(date)

  ifelse(
    m <= 3,
    paste0(y - 1, "/", y),
    paste0(y, "/", y + 1)
  )
}

Using the function

lubridate::today()

[1] "2023-12-07"

fin_year(lubridate::today())

[1] "2023/2024"