CODEWITHSUNDEEP
rplotscatter-plot
Slim Shady
Slim Shady

Reputation: 220

Different plotting symbols in r scatter plot

    bweight  gestwks            hyp  sex
1   2974    38.5200004577637    0   female
2   3270    NA                  0   male
3   2620    38.150001525878899  0   female
4   3751    39.799999237060497  0   male
5   3200    38.889999389648402  1   male
6   3673    40.970001220703097  0   female

I would like to plot Baby Weight (bweight) against gestation period (gestwks) .

Here is my code for that:

plot(courseworkData1$bweight ~ courseworkData1$gestwks, xlab="Gestation Period", ylab="Baby weight")

But then, I have the "hyp" variable. It is denoted in short for maternal hypertension. I would like to make a plot which would plot the baby weights with hypertension with a different symbol, like a square or something else maybe. How do I do that?

Upvotes: 2

Views: 3820

Answers (2)

Edward
Edward

Reputation: 19134

"If i'd like to change the symbols for example from plain circle to dark filled circle and squares to say triangles, how would you do that?"

R has 25 easily accessible plotting characters (pch) based on the integers 1-25. There are others, but these are the commonly used ones. You can see them all from these commands:

# windows(7,4)
plot(-1:25, rep(0,27), pch=-1:25, cex=2, col="blue", bg="red", yaxt="n")
text(-1:25, rep(0,27), labels=-1:25, pos=1)
abline(v=c(14.5, 20.5))
mtext(text=c("Hollow symbols", "Filled symbols", "Filled symbols\nwith borders"), 
      side=3, at=c(6, 17.5, 23), line=-2)

enter image description here

For pch -1:14 you get hollow symbols – they cannot be filled in with colour or background. Only symbols 15 to 20 can be filled in completely with colour using the col graphical parameter (the default is black). For symbols 21 to 25, the colour is specified using the bg graphical parameter (the default is none) and additional borders can be added with their colour specified using col (the default is black) and width using lwd (the default is 1). Symbols 15 to 18 do not have a border, so the filled circle given by pch=16 will be smaller than the solid circle given by pch=19, by an amount which depends on the size of the border (lwd) and the size of the symbol (cex).

So, to obtain a dark filled circle you could choose pch=16 or pch=19 and for a (hollow) triangle you could choose pch=2 or pch=6 (inverted).

Upvotes: 1

dario
dario

Reputation: 6483

We can use the pch parameter to plot to specify a variable to be used for symbols:

Edit:

Adapted for your data:

Create MRE:

courseworkData1 <- read.table(text="ID   bweight  gestwks            hyp  sex
1   2974    38.5200004577637    0   female
2   3270    NA                  0   male
3   2620    38.150001525878899  0   female
4   3751    39.799999237060497  0   male
5   3200    38.889999389648402  1   male
6   3673    40.970001220703097  0   female", header=TRUE, stringsAsFactors=FALSE)

With ?pch we get the list of available symbols. Then we just have to add a new variable to encode the desired key. For example if we want to use squares (key = 0) and triangles (key = 2):

courseworkData1$symbol_key <- ifelse(courseworkData1$hyp == 0, 0, 2)
plot(bweight ~ gestwks, data=courseworkData1, 
     xlab="Gestation Period", ylab="Baby weight", pch=courseworkData1$symbol_key)
legend("bottomright", title="hyp", legend = c("0", "1"), pch = c(0,2))

enter image description here

Upvotes: 2

Related Questions