R grouped bar plot

Question

I'm quite new to R studio, so I apologise in advance. I need help to create a grouped barplot. I have three variables: "Time": converted to a continuous variable "Treatment": "Con", "Hya" "Trial": "T1", "T2", "T3" I want to produce something like this:

There should be three groups of three columns stacked beside each other. Time on the Y-axis; Trial (1,2 & 3)on the X-axis; Treatment corresponding to coloured columns (Hya=grey, Con=white) with a legend explaining Treatment colour.

Here is the structure of my data:

'data.frame':   102 obs. of  3 variables:
 $ Trial    : int  1 1 1 1 1 1 1 1 1 1 ...
 $ Treatment: $ Trial    : int  1 1 1 1 1 1 1 1 1 1 ...
 $ Treatment: Factor w/ 2 levels "Control","Hyaluronan": 1 1 1 1 1 1 1 1 1 1 ...
 $ Time     : num  11 7 7.68 7.7 7 3 5 5.48 4 6 ...

I get this error message:

> barplot(table(Biopsy$Time, Biopsy$Treatment, Biopsy$Trial))
Error in barplot.default(table(Biopsy$Time, Biopsy$Treatment, Biopsy$Trial)) : 
  'height' must be a vector or a matrix

Please if anyone is able to help I would so appreciate it, I've been trying for so long :(

Do you want to aggregate the Time for observations where Trial and Treatment are identical? — Harald Gliebe, Sep 09 '17 at 07:47
I don't think so? Just three groups of 2 columns (Hya and Con) corresponding to Trial 1, 2 and 3. Sorry, my question wasn't clear. — Ashlee Hutchinson, Sep 09 '17 at 08:26
But what should the time value for a specific column be? Say for (Trial 1, Control) , there are multiple Time values for this combination, do you want to use the mean() of these values? If you have a unique Time value per Trial/Treatment combination you could plot them like this: http://www.r-fiddle.org/#/fiddle?id=LF739Zwg&version=1 — Harald Gliebe, Sep 09 '17 at 08:50
That looks really good! Thankyou so much! Yes, they are unique time points per trial/treatment. When I plot this though, the Y axis stops at 10, and the columns extend beyond this... Thankyou so much! — Ashlee Hutchinson, Sep 09 '17 at 08:58

score 1 · Answer 1 · edited Sep 09 '17 at 20:40

1

I think it is useful to mention the "ggplot2" package here. With this package the creation of a stacked bar plot is quite easy. I was not sure about the data frame you are working with since you are only providing a snapshot of your data structure, but I hope the data frame I created as an example will help to show you the basic function used to create such a plot. (You can just copy-paste in RStudio and run the code. Make sure to install the ggplot2 [install.packages("ggplot2")] package before running the library() function.)

trial <- c(1,1,1,1,2,2,2,2,3,3,3,3)
treatment <- c("Hya","Hya","Con","Con","Hya","Hya","Con","Con","Hya","Hya","Con","Con")
time <- c(1,7,1,7,2,8,2,8,3,9,3,9)

df <- data.frame(trial,treatment,time)

library(ggplot2)
ggplot(df, aes(y = time,
               x = trial,
               group = treatment)) +
  geom_bar(stat = "identity", position = "dodge", aes(fill = treatment))

The resulting plot can be found here.

The above code will create a data frame and a bar plot. The grouping is done with the argument "group", the color set with "fill". Of course you can modify the coloring etc. Since you are new RStudio/R I recommend you check out the documentation of ggplot.

I hope this example helps...

edited Sep 09 '17 at 20:40

CPak

13,260
3
30
48

answered Sep 09 '17 at 19:39

Ben

99
6

Thankyou. Where do the numbers come from in time? – Ashlee Hutchinson Sep 09 '17 at 22:46
time <- c(1,7,1,7,2,8,2,8,3,9,3,9) – Ashlee Hutchinson Sep 09 '17 at 22:46
Another person has used different numbers. Are they just a sample, and I need to put in my own numbers? Which numbers should I put in? – Ashlee Hutchinson Sep 09 '17 at 22:47
Yes. All the data/numbers in the data frame I created are just example numbers. You should have your data frame already if I am right. If you have problems with your data frame, meaning if you have problems to simply take your data frame and use the ggplot function I provided, the reason might be the structure of your data frame. – Ben Sep 09 '17 at 23:23
Thankyou that's really helpful – Ashlee Hutchinson Sep 10 '17 at 04:56

R grouped bar plot

1 Answers1