Plotting specific graphs given in additional input (ggplot)

Question

I have the problem which I solved but I think my solution is to involved and unproductive when dealing with more variables. I want to write a function which combine three graphs :

Numbers from standard normal distribution
Numbers from standard uniform distribution
Numbers from standard exponential distribution

The length of the samples can be specified. Also in output I want to have one additional input named plot_types. By default it's combining all graphs. But user can specify also exactly which plots should be included. For example if plot_types==c('Norm.plot','Unif.plot') R should only plot numbers from normal and uniform distribution omitting numbers from exponential distribution.

My work so far

I wrote code following :

library(ggplot2)
library(gridExtra)

visual=function(num,plot_types='all'){
  Norm.plot <- ggplot()+
    geom_line(aes(x=1:num,y=rnorm(num)))+
    labs(title='Norm plot',x=NULL,y=NULL)+ 
    theme(plot.title = element_text(hjust = 0.5))
  
  Unif.plot <- ggplot()+
    geom_line(aes(x=1:num,y=runif(num)))+
    labs(title='Unif plot',x=NULL,y=NULL)+ 
    theme(plot.title = element_text(hjust = 0.5))
  
  Exp.plot <- ggplot()+
    geom_line(aes(x=1:num,y=rexp(num)))+
    labs(title='Exp plot',x=NULL,y=NULL)+ 
    theme(plot.title = element_text(hjust = 0.5))
  
  if (plot_types=='all'){grid.arrange(Norm.plot,Exp.plot, Unif.plot,ncol=2)}
  else if (plot_types==c('Norm.plot','Unif.plot') || 
           plot_types==c('Unif.plot','Norm.plot')){grid.arrange(Norm.plot, Unif.plot)}
  else if (plot_types==c('Norm.plot','Exp.plot') ||
           plot_types==c('Exp.plot','Norm.plot')){grid.arrange(Norm.plot, Exp.plot)}
  else if (plot_types==c('Unif.plot','Exp.plot') || 
           plot_types==c('Exp.plot','Unif.plot')){grid.arrange(Exp.plot, Unif.plot)}
}

visual(50,plot_types = c('Norm.plot','Unif.plot'))

The code above has several problems. The first one is that it has a lot of loops so it's very unproductive. Also it would be very problematic when trying to extend it to higher numbers of variables.

Do you have any ideas how can I omit using so many loops ?

Z.Lin · Accepted Answer · 2020-09-14 11:32:26Z

See if this suits your needs

visual.new <- function(num, plot_types = 'all') {
  # define a data frame for all the results
  data <- data.frame(x = seq(1, num),
                     y.norm = rnorm(num),
                     y.unif = runif(num),
                     y.exp = rexp(num))

  # define a base ggplot object
  gg <- ggplot(data, aes(x = x)) + 
    geom_line() +
    theme(plot.title = element_text(hjust = 0.5),
          axis.title = element_blank())

  # define a list of plots
  plot.list <- list(gg + aes(y = y.norm) + ggtitle("Norm plot"),
                    gg + aes(y = y.unif) + ggtitle("Unif plot"),
                    gg + aes(y = y.exp) + ggtitle("Exp plot"))

  # initial default: do not show any plot
  show.plot <- c(FALSE, FALSE, FALSE)

  # determine whether to show all / any plot based on plot_types value
  if('all' %in% plot_types) {
    show.plot <- c(TRUE, TRUE, TRUE)
  } else {
    if('Norm.plot' %in% plot_types) show.plot[1] <- TRUE
    if('Unif.plot' %in% plot_types) show.plot[2] <- TRUE
    if('Exp.plot' %in% plot_types) show.plot[3] <- TRUE
  }
  
  if(sum(show.plot) == 0) {
    message("Invalid plot_types parameter entered. No plot shown.")
  } else {
    cowplot::plot_grid(plotlist = plot.list[show.plot])
  }
}

Test

visual.new(50) # 3 plots
visual.new(30, c('Norm.plot', 'Exp.plot')) # 2 plots
visual.new(80, c('Exp.plot', "Unif.plot", "Norm.plot")) # 3 plots, in norm-unif-exp order
visual.new(25, 'Unif.plot') # 1 plot
visual.new(40, "something") # no plot, because plot_types value is invalid

(I used plot_grid from the cowplot package, rather than gridExtra's grid.arrange, because the former can accept a list of ggplot objects, which is useful here.)

Duck · Accepted Answer · 2020-09-14 12:24:46Z

2

Pointing out to the fantastic solution of @Z.Lin you can also use wrap_plots() function from patchwork. I took the smart function from amazing @Z.Lin and added a slight change. Here the code:

#Code
visual.new <- function(num, plot_types = 'all') {
  # define a data frame for all the results
  data <- data.frame(x = seq(1, num),
                     y.norm = rnorm(num),
                     y.unif = runif(num),
                     y.exp = rexp(num))
  
  # define a base ggplot object
  gg <- ggplot(data, aes(x = x)) + 
    geom_line() +
    theme(plot.title = element_text(hjust = 0.5),
          axis.title = element_blank())
  
  # define a list of plots
  plot.list <- list(gg + aes(y = y.norm) + ggtitle("Norm plot"),
                    gg + aes(y = y.unif) + ggtitle("Unif plot"),
                    gg + aes(y = y.exp) + ggtitle("Exp plot"))
  
  # initial default: do not show any plot
  show.plot <- c(FALSE, FALSE, FALSE)
  
  # determine whether to show all / any plot based on plot_types value
  if('all' %in% plot_types) {
    show.plot <- c(TRUE, TRUE, TRUE)
  } else {
    if('Norm.plot' %in% plot_types) show.plot[1] <- TRUE
    if('Unif.plot' %in% plot_types) show.plot[2] <- TRUE
    if('Exp.plot' %in% plot_types) show.plot[3] <- TRUE
  }
  
  if(sum(show.plot) == 0) {
    message("Invalid plot_types parameter entered. No plot shown.")
  } else {
    patchwork::wrap_plots(plot.list[show.plot])
  }
}

Some test (Using fantastic code from @Z.Lin):

visual.new(80, c('Exp.plot', "Unif.plot", "Norm.plot"))

Output:

answered Sep 14, 2020 at 12:24

Duck

39.6k14 gold badges47 silver badges89 bronze badges

2 Comments

John Over a year ago

Why do you reckon that wrap_plots is better than plot_grid ? ;))

Duck Over a year ago

@John Hi John, I am just proposing as an option. wrap_plots() is like an improvement for plot_grid() and it has more advanced features. So the OP can have different options in case of any modification :)

Collectives™ on Stack Overflow

Plotting specific graphs given in additional input (ggplot)

2 Answers 2

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related