Clean way to show year once and group months within?

Question

Tough to explain using words, so please see the image below. I know I can facet by year, but that does not accomplish what I want.

Is this possible?
clean year-month

library(tidyverse)
library(lubridate)

set.seed(88)
start <- ymd("2009/11/01")
data <- data_frame(date = start + months(0:11)
              , donuts = rnorm(12, mean = 30, sd = 5) %>% round(., 0)
              , year = year(date)
              , month = month(date, label =T, abbr =T)
              )

ggplot(data = data, aes(x = date, y = donuts)) + geom_line()

This would be a good feature in ggplot2 for sure! – George Feb 10 '17 at 11:18 — George, Feb 10 '17 at 11:18

score 1 · Answer 1 · edited Jul 23 '21 at 18:16

You can get pretty close with facetting,

dummy <- mutate(data[3,], date = date - 1L, year = year(date)) 
data <- rbind(data, dummy)

ggplot(data = data, aes(x = date, y = donuts, group=1)) + geom_line() +
  facet_grid(.~year, scale="free", space="free", switch = "x") +
  scale_x_date(date_breaks = "1 month", date_labels = "%b", expand = c(0,0)) +
  theme_minimal() + theme(panel.spacing.x = unit(0,"line"), 
                          strip.placement = "outside")

score 0 · Answer 2 · answered Feb 10 '17 at 06:29

If someone doesn't know a slicker way to do this, you could do this with a bit of a work around. Make the months an ordered factor, starting with Nov 2009 through Oct 2010. Then you could increase the space beneath the x-axis lables and add 2009 and 2010 as text underneath them. I think you can then set the x-axis tick length manually with axis.ticks.length within your theme, making the ones after Dec 2009 and Oct 2010 longer...

Clean way to show year once and group months within?

2 Answers2