Script Conventions

Conventions I try to follow

Scripts

• one script = one purpose
• keep scripts short

Example layout of script folder:

R/
├── 01_data_processing.R   # scripts for cleaning and processing data
├── 02_linear_models.R     # scripts for running linear models 
├── utilities.R           # scripts for functions that are used across multiple analyses
├── functions_data_processing.R  # scripts for functions that are specific to data processing

Functions

• any repeated logic gets moved to functions
• if repeated across different tasks -> utilities.R
• if task-specific -> functions_01_data_cleaning.R

Style conventions

• Use pipes: |>
• one step per line

df |>
    dplyr::filter(age > 18) |>  
    dplyr::mutate(age_group = ifelse(age < 30, "young", "old")) |>
    dplyr::group_by(age_group) |>
    dplyr::summarise(mean_income = mean(income))

Naming Conventions

• use snake_case for everything
• objects start with noun (e.g., df_* or mod_*)
• functions start with verb (e.g., clean_, fit_)

Ordering and Flow

if structure begins to exist (e.g., order of scripts) use targets to remove manual ordering/execution

list(
  tar_target(raw, get_data()),
  tar_target(cleaned, clean_data(raw)),
  tar_target(model, fit_linear_model(cleaned)),
  tar_target(plot, create_plot(model))
)