style.md

title	R Style
editor_options	chunk_output_type inline

In terms of the functions to use, attempt to use the functions and idioms from the tidyverse where at all possible

• ggplot2 for graphics
• tibbles instead of data frames
• dplyr and tidyr for data frame wrangling
• purrr for apply-type functionals
• forcats for factors
• lubridate for dates and times
• stringr for strings
• readr, haven, readxl or rio for reading from data formats

Use ggplot2, not base or lattice. The only exception is to use the plot() method of an object when convenient. But do not create new graphics using non-ggplot2 formats.

Some rules:

• do not use setwd() in scripts EVER.
• do not use install.packages in scripts.
• Do not use subset or transform, use filter and mutate
• Do not use aggregate or by, use summarise
• Do not use merge, use the various dplyr *_join functions
• Do not use order or sort (when sorting data frames, it's okay for vectors), use arrange.
• Do not use unique or duplicated with data frames. Use distinct()
• Do not use sapply, vapply, tapply, use purrr map functions.
• It is okay to use apply as the current purrr functions are annoying. However, they should only be used for matrices. Never on data frames.
• use haven::read_dta instead of foreign::read.dta
• use readr::read.csv instead of read.csv
• use haven::read_spss intead of foreign::read.spss

It is okay to create new columns usine $<- or [[<-. But do not filter for using logicals within [. Use filter().

In writing code, follow the tidyverse Style Guide.