R: Making multiple categorized scatterplots from one dataframe

Question

I'm trying to create scatterplots using data from B and C, however I'd like to create a difference scatterplot for each category in A.

A           B   C
Monday      2   4
Tuesday     4   2
Monday      3   5
Wednesday   3   10
Friday      7   9

This is the code I currently have is to just make a normal scatterplot with my data. Is there an addition or something that I can use to automatically make scatterplots based on category?

attach(data)
plot(C, B, main="scatterplots",xlab="C", ylab="B", pch=10) 
abline(lm(C~B), col="red")

I suspect this is a toy example, but I feel compelled to point out that `attach()` is not recommended in R, & you'll need more than 1 B C pair per day for scatterplots to be meaningful, perhaps a different plot is more appropriate. — gung - Reinstate Monica, Sep 10 '13 at 16:19
@gung; this is not my real data set, but I wanted to illustrate my point. I've seen attach() in a lot of sample scripts I've been looking at for graphing, do you have a reccommended way instead of that? — Mengll, Sep 10 '13 at 16:52
check out this thread: [in-r-do-you-use-attach-or-call-variables-by-name-or-slicing?](http://stackoverflow.com/questions/1310247/) — gung - Reinstate Monica, Sep 10 '13 at 17:15
If you use "Formula Notation" notation of `plot`, you can use both `data` and `subset` arguments, e.g. `df <- data.frame(x = 1:10, y = rnorm(10)); plot(y ~ x, data = df, subset = x > 5)` — Henrik, Sep 10 '13 at 17:18

score 3 · Accepted Answer · answered Sep 10 '13 at 16:03

3

This could be a solution:

par(mfrow=c(1, length(levels(A))))
for (day in levels(A)){
  subs <- subset(data, A==day)
  plot(subs$C, subs$B, main=day)
  abline(lm(C~B), col="red")  
}

Note that you have to adjust your xlim and ylim values in case you want to see the red lines for the fit in each plot. Does this help?

answered Sep 10 '13 at 16:03

user1981275

13,002
8
72
101

Thank you! I've created the scatterplots I need. However, they're super scrunched up. I suppose this is what you mean by adjusting the lim values? – Mengll Sep 10 '13 at 16:15
This is because the code plots all four plots in one row. You can resize the plotting window manually or create the plots in different windows using for instance the `x11` command. When you save the plots with `jpeg` or `pdf`, you can specify width and height as well. – user1981275 Sep 10 '13 at 17:05

score 3 · Answer 2 · answered Sep 10 '13 at 16:53

3

Showing a plot conditioned on another variable is what the lattice package was designed to do. In your case it may be as simple as:

library(lattice)
xyplot(B~C|A, data=data, type=c('p','r'))

The ggplot2 package also does this using faceting:

library(ggplot2)
qplot( C, B, data=data, facets= A ~ .) + geom_smooth(method='lm')

answered Sep 10 '13 at 16:53

Greg Snow

48,497
6
83
110

This works fine and the graph itself looks really good! Thank you for the code. – Mengll Sep 10 '13 at 16:57

R: Making multiple categorized scatterplots from one dataframe

2 Answers2