2

I would like to create multiple plots at one page, where I am iteration over different Y variables over the same X. = i.e. I want one plot per one Y. Usually, I would copy and paste my ggplot with just changed Y value, store individual plots as p.y1, p.y2 and plot all of them using grid.arrange(p.y1, p.y2) like here:

enter image description here

This approach is not very fun when I have 10 different Y variables, and I want to plot all of them. I wonder how to make the process more efficient?

I thought that I can simply create a vector of Y (colnames of df) and then loop through them to create multiple plots. But, it seems that my output plots are not correct to pass to grid.arrange(), and I can not plot them neither.

How can I loop through multiple Ys, and then arrange all plots on one page? As I do not have multiple factors, I probably cannot use facet_grid nor facet_wrap.

Here is my dummy example for two Ys: y1 and y2

set.seed(5)
df <- data.frame(x = rep(c(1:5), 2),
                 y1 = rnorm(10)*3+2,
                 y2 = rnorm(10),
                 group = rep(c("a", "b"), each = 5))  

# Example of simple line ggplot
ggplot(df, aes(x = x,
              y = y2,   # here I can set either y1, y2...
              group = group,
              color = group)) +
  geom_line()

Now, iterate over the vectors of Ys and store output plots in a list:

# create vector of my Ys
my.s<-c("y1", "y2")

# Loop over list of y to create different plots
outPlots<- list()

for (i in my.s) {
      print(i)
      my.plot <- 
        ggplot(df, aes_string(x = "x",
                              y = i,
                              group = "group",
                              color = "group")) +
        geom_line()
     # print(plot)
      outPlots <- append(outPlots, my.plot)
    }

Intendent plotting of multiple graphs on one page: does not work because of Error in gList(list(data.x1 = 1L, data.x2 = 2L, data.x3 = 3L, data.x4 = 4L, : only 'grobs' allowed in "gList"

grid.arrange(outPlots)
maycca
  • 3,848
  • 5
  • 36
  • 67

2 Answers2

1

You could try this. I hope this helps.

library(reshape2)

Melted <- melt(df,id.vars = c('x','group')) 

#Plot
ggplot(Melted,aes(x=x,y=value,group=group,color=group))+
  geom_line()+
  facet_wrap(~variable,ncol = 1,scales = 'free')+theme_bw()

enter image description here

Duck
  • 39,058
  • 13
  • 42
  • 84
  • This is nice approach, it makes it more versatile if needed to add later another factor or modify the `ylim()` etc. Thank you @Duck! – maycca Jul 09 '20 at 07:44
1

I propose another solution based on this post.

Plotfunction <- function(y){my.plot <- 
    ggplot(df, aes_string(x = "x",
                          y = y,
                          group = "group",
                          color = "group")) +
    geom_line()}

n <- ceiling(sqrt(length(my.s)))
do.call("grid.arrange",
        c(lapply(my.s, Plotfunction), ncol = n, nrow = n))
Rémi Coulaud
  • 1,684
  • 1
  • 8
  • 19
  • Thank you @RemiCoulaud this is exactly what I could not figure out. To make it even more versatile, I suggest to replace `ncol = 2` by `ncol = length(my.s)` or something similar, to adjust to different number of Ys. Thanks again! – maycca Jul 09 '20 at 06:36