title	author	date	output
README	Mhairi McNeill	28 May 2016	html_document

Ordered

I got tired of fiddling around with manually setting factor levels, so I made this set of convenience functions to set factor levels. This is particularity useful when you are working in ggplot and want your categories in the right order.

Installation

# install.packages("devtools")
devtools::install_github("mhairi/ordered")

library(ordered)

Functions

There's currently three functions in the package.

order_by

First argument is the factor you want to order, second is the order of the elements of your factor.

order_by(letters[c(1,1,2,3,3)], c(1,1,2,3,3))

## [1] a a b c c
## Levels: a < b < c

Won't let you order if you don't have a 1-1 relationship between the order and the elements. (Each element can only have one order).

order_by_summary

Possibly more useful. First argument is the factor you want to order, the second is some attribute that you summarise to find the order.

factor = c('a', 'a', 'b', 'b')
order = c(10, 12, 4, 5)
order_by_summary(factor, order, mean)

## [1] a a b b
## Levels: b < a

order_by_number

I often find myself working with numbered factor variables, which have a natural order in the numbers in them.

f <- c('group_1', 'group_2', 'group_3')
order_by_number(f)

## [1] group_1 group_2 group_3
## Levels: group_1 < group_2 < group_3

When is this useful?

I mainly use this for making bar graphs in ggplot2!

By default ggplot2 orders your factors in alphabetic order. Which is sometimes fine.

library(ordered)
library(ggplot2)

df <- data.frame(x = letters[1:10], y = rpois(10, 5))

ggplot(df) +
  aes(x = x, y = y) +
  geom_bar(stat = 'identity')

But quite often I find it clearer to have them ordered by the size of the bar.

df$x <- order_by(df$x, df$y, desc = TRUE)

ggplot(df) +
  aes(x = x, y = y) +
  geom_bar(stat = 'identity')