visualizing crosstab tables with a plot in R

Question

I saw a plot generated in excel and I was wondering if R could also do it. this picture is essentially a visualization of a crosstab table comparing the days of the week to preffered meals on that day and counting the number of people that fall within those categories

This picture is essentially a visualization of a crosstab table comparing the days of the week to preferred meals on that day and counting the number of people that fall within those categories.

I've read up on some R bubble charts but I have yet to see one like this. Can someone point me to a package or a website that explains how I can make a plot like this one?

Perhaps [my answer here](http://stats.stackexchange.com/a/34778/13474) might help you. — A5C1D2H2I1M1N2O1R2T1, Dec 19 '13 at 05:18
I am interested in how that chart was made in Excel to start with. Any clues? — Faustin Gashakamba, Dec 20 '21 at 07:07

Tommy O'Dell · Accepted Answer · 2013-12-19T08:56:03.473

26

Using Hadley Wickham's ggplot2:

library(ggplot2)                           

# Set up the vectors                           
days <- c("Mon","Tues","Wed","Thurs","Fri")
slots <- c("Coffee/Breakfast","Lunch","Happy Hour","Dinner")

# Create the data frame
df <- expand.grid(days, slots)
df$value <- c(1,1,1,1,2,1,1,NA,NA,1,4,4,7,4,1,5,6,14,5,1)    

#Plot the Data
g <- ggplot(df, aes(Var1, Var2)) + geom_point(aes(size = value), colour = "green") + theme_bw() + xlab("") + ylab("")
g + scale_size_continuous(range=c(10,30)) + geom_text(aes(label = value))

enter image description here

Do you care that the axis lines go through the circles? Also, the greens are slightly different and the label text is black instead of white.

edited Dec 19 '13 at 08:56

answered Dec 19 '13 at 05:23

Tommy O'Dell

7,019
13
56
69

it's perfect this way, thank you. just one question, is there a way to change the shapes, the circles look a bit "ruffled" around the edges. Can the function make it a smooth circle? If not then that's okay – CadisEtRama Dec 19 '13 at 06:46
1

With `ggsave` you can set the dots per inch. The default is 72 dpi, with `ggsave(file="filename.png", dpi=300)` you can set the value for dpi to a higher value. This might solve your problem. – Jaap Dec 19 '13 at 15:35

Nick Vence · Answer 2 · 2021-03-05T22:47:51.523

The ggmosaic package extends ggplot to provide an alternative.

library(ggplot2)
library(ggmosaic)

# Set up the vectors                           
days <- c("Mon","Tues","Wed","Thurs","Fri")
slots <- c("Coffee/Breakfast","Lunch","Happy Hour","Dinner")

# Create the comporessed data frame
df <- expand.grid(days, slots, stringsAsFactors = TRUE)
df$value <- c(1,1,1,1,2,1,1,0,0,1,4,4,7,4,1,5,6,14,5,1)    
df.expanded <- df[rep(row.names(df), df$value), 1:2]

#Plot the Data
ggplot(data = df.expanded) +
  geom_mosaic(aes(x = product(Var2,Var1), fill = Var2)) + 
  ggsave("mosaic.png")

score 0 · Answer 3 · answered Jun 13 '21 at 02:21

Just adding an alternative approach that makes use of the excellent ggpubr package:

# Load ggpubr package
library(ggpubr)

## Borrowed from Tommy O'Dell's answer
# Set up the vectors                           
days <- c("Mon", "Tues", "Wed", "Thurs", "Fri")
slots <- c("Coffee/Breakfast", "Lunch", "Happy Hour", "Dinner")

# Create the data frame
df <- expand.grid(days, slots)
df$value <- c(1, 1, 1, 1, 2, 1, 1, NA, NA, 1, 4, 4, 7, 4, 1, 5, 6, 14, 5, 1)

## Plot the data (my contribution)
ggballoonplot(
  data = df, 
  x = "Var1", 
  y = "Var2",
  size = "value",
  size.range = c(10, 20), 
  fill = "green",
  show.label = TRUE, 
  rotate.x.text = FALSE, 
  legend = "none"
)

^{Created on 2021-06-12 by the reprex package (v2.0.0)}

visualizing crosstab tables with a plot in R

3 Answers3

Linked