R - how to use contents of one vector as the symbol in a plot?
I have a two vectors of numbers of equal length. How do I plot the first vector while using the corresponding eleme开发者_如何学编程nt in the second vector as the printing character? (Background: I sorted the first column and the second column holds the original indices. I want to use the indices as the printable character so that I can see which data points are outliers, since each number represents one run of data).
> x
$x
[1] 25 29 30 34 38 572 700 733 870 879 899 934 982 1054 1135 1258
[17] 1315 1491 1685 1700 2069 2131 2284 3498 3506 4467 4656 5633 6642 8348
$ix
[1] 23 3 18 30 13 8 4 14 11 17 12 29 9 15 19 16 7 1 20 2 6 28 21 10 5 22 24 26
[29] 25 27
First vector is x$x, second vector is x$ix (results of calling sort with index.return = TRUE)
I've tried plot(x$x, pch=str(x$ix)) but that treats x$ix numerically. If this were Python I would do something like strings = [str(x) for x in x$ix]. but this is R and I've forgotten most of what I used to know.
I found that you can do as.character(x$ix) in order to get the strings,
> as.character(x$ix)
[1] "23" "3" "18" "30" "13" "8" "4" "14" "11" "17" "12" "29" "9" "15" "19" "16"
[17] "7" "1" "20" "2" "6" "28" "21" "10" "5" "22" "24" "26" "25" "27"
and I can use this as the input to pch. But only the first character is used (and according to the docs, that's normal).
I know there's a way to do this; I did it in college. But I can't for the life of me remember how I did it.
Chart without labels:
Chart with labels, but incorrect:
This should work:
x = 1:4
y = x
plot(x, y, ann=F, axis=F, col="blue", pch=16)
text(x, y, labels=c("1st", "2nd", "3rd", "4th"), col="red", pos=c(3,4,4,1), offset=0.6)
Just convert your non-data vector (the one containing the labels) to a character vector: labels = as.character(label_vector)
and then substitute this for the third argument in line 4 above.
The 'Text' function is fairly versatile because of the various arguments you can pass in--e.g., (as in the example above) you can set the text to a different color than your data points using "col"; You can also specify the position (relative to the data point annotated by a given text label) for each text label separately. That's often useful to keep text labels from, for instance, overlapping one of the axes, which is what happened the first time i ran this example without setting 'pos'. So by setting 'pos' (as c(3, 4, 4, 1)) i set the position of the text labels as "above", "right", "right", and "below"--moving the first data point up so it doesn't hit the bottom x-axis, and moving the fourth one down so it doesn't hit the top x-axis. Additionally, using 'offset' (which has a default value of 0.5) you can specify the magnitude of the position adjustment.
Here is a way to do this using the ggplot2
package:
library(ggplot2)
x <- rnorm(10)
y <- rnorm(10)
labs <- 1:10
ggplot()+geom_text(aes(x=x,y=y,label=labs))
Here's a very basic solution using lattice
:
dat <- data.frame(x = rnorm(10), y = rnorm(10),
labs = as.character(1:10))
xyplot(y~x,data = dat,panel = function(x,labs){
panel.text(x,y,labels = labs)},
labs = dat$labs)
精彩评论