为具有小平面的图设置不同的Y轴范围（比例=“免费”似乎无效）

Question

I am trying to set different Y-axis ranges for different facets. I read many answers on the web and I tried to apply them on my script, but it seems not working.

I am using scales = "free" to make the axis appear on all the facets and I am using geom_blank to set different axis limits depending on the variable Variable of the somePoints dataframe

Here is a reproducible example of the data:

library(ggplot2)

dummy <- expand.grid(Year = NA, Season = c("Winter","Spring","Summer","Fall"),
                     Variable = c("P","T"), stringsAsFactors=FALSE)

dummy$ymin = NA
dummy$ymax = NA
dummy[dummy$Variable=="P","ymin"] = 0
dummy[dummy$Variable=="P","ymax"] = 400
dummy[dummy$Variable=="T","ymin"] = -10
dummy[dummy$Variable=="T","ymax"] = 30

somePoints = expand.grid(Year = NA, Season = c("Winter","Spring","Summer","Fall"),
                         Variable = c("P","T"), stringsAsFactors=FALSE)

somePoints$Value = NA
somePoints[somePoints$Variable=="P","Year"] = 1940
somePoints[somePoints$Variable=="T","Year"] = 1940
somePoints[somePoints$Variable=="P","Value"] = c(350,365,200,150)
somePoints[somePoints$Variable=="T","Value"] = c(-5,8,25,7)

Basically what I want is having the same Y-axis for P Variable ranging from 0 to 400 and the same Y-axis for T Variable ranging from -10 to 30.

Here are the ggplot code lines I am using to obtain what I want:

ggplot()+

  facet_grid(Season ~ Variable, scales = "free_y")+

  geom_blank(data = dummy,aes(y=ymin)) +
  geom_blank(data = dummy,aes(y=ymax)) +

  geom_point(data = somePoints, aes(Year,Value), size = 5) +
  geom_point(data = somePoints, aes(Year,Value), size = 5)

What I notice is that setting scales = "free_y" in the facet_grid function seems not working.

What am I doing wrong? I feel I am missing something stupid.

Thank you.

Answer 1

If I'm understanding correctly, your issue is more with geom_blank . The purpose of geom_blank is to ensure common scales between different plots. You should use the expand argument instead. I suggest trying the following:

library(ggplot2)
ggplot()+

  facet_grid(Season ~ Variable, scales = "free_y")+

  geom_point(data = somePoints, aes(Year,Value), size = 5) +
  geom_point(data = somePoints, aes(Year,Value), size = 5) +

  scale_y_continuous(expand = c(.2,0))

expand is just buffering your edges.

Alternatively, you can use facet_wrap , rather than facet_grid so that the axes are displayed on each plot. We'll just specify number of columns ( ncol = 2 ) and vertical direction ( vir = "v" ):

ggplot()+

  facet_wrap(Variable ~ Season, scales = "free_y",
             dir = "v", ncol = 2)+

  geom_blank(data = dummy,aes(y=ymin)) +
  geom_blank(data = dummy,aes(y=ymax)) +

  geom_point(data = somePoints, aes(Year,Value), size = 5) +
  geom_point(data = somePoints, aes(Year,Value), size = 5) +

  scale_y_continuous(expand = c(.2,0))

If you don't like the labels, you can pass labeller = labeller(.multi_line = FALSE) to facet_grid