R-project Graphic plot

Question

I'm trying to plot from the data Vocabulary.txt education in relation to vocabulary.

This is the code I used

plot(jitter(education)~jitter(vocabulary),pch=23,xlim=c(0,30),ylim=c(0,30))

my graphic looks like this enter image description here

It doesn't look right maybe someone can explain to me what I did wrong and further on what exactly the command jitter does?

I cant see anything from this data I wonder If theres a better way to plot it.I read ?jitter but I dont understand it. — J.Goodcat, Jun 13 '16 at 17:58
What is it you want to see? There isn't enough here to provide an answer. — gung - Reinstate Monica, Jun 13 '16 at 18:00
The exercise ist just to make a good graphic which could be published and I just wanted to know if I ploted the data right because it looked wrong to me — J.Goodcat, Jun 13 '16 at 18:06
You have to either add some more context so that we can understand why it "looks wrong" to you; or delete the question since the discussion looks counter-productive. — YakovL, Jun 13 '16 at 18:26
see `?jitter` and examples. it tries hard to add some noise (usually just enough with default parameters but it can be tuned). it helps to see points on the very same coordinates by jittering them. — Vincent Bonhomme, Jun 13 '16 at 20:18

Vincent Bonhomme · Accepted Answer · 2018-03-19T12:28:20.917

I think the two outputs below are ~pub-ready.

The first uses base R and jitter, used to add some noise to data so that points with the very same coordinates appear on different positions. That's a nice approach in such case (providing you mention the jittering as data are slightly modified). If you have many points, you can combine this approach with some transparency.

First of all, we make the example reproducible:

df <- read.table("http://socserv.socsci.mcmaster.ca/jfox/Books/Applied-Regression-3E/datasets/Vocabulary.txt", header=TRUE)
plot(jitter(education)~jitter(vocabulary), df, pch=20, col="#00000011",
     xlim=range(vocabulary), ylim=range(education),
     xlab="vocabulary", ylab="education")

enter image description here

But fundamentally, you may be trying to plot a contingency table, so the second, using ggplot2:

library(ggplot2)
# creates a contingency table
tab.df <- as.data.frame(with(df, table(education, vocabulary)))
ggplot(tab.df) + aes(x=vocabulary, y=education, fill=Freq, label=Freq) + 
# colored tiles and labels (0s are omitted)
geom_tile() + geom_text(data=subset(tab.df, subset = Freq != 0), size=2) +
# cosmectics
scale_fill_gradient(low="white", high="red") + theme_linedraw()

enter image description here

Plotting percentages (both for tiles and labels) may be a better choice but your question was vague about your objectives. And if you want the first plot, but ala ggplot2 you still can work around:

ggplot(df) + aes(x=education, y=vocabulary) + geom_jitter(alpha=0.05)

enter image description here

Your phrase that starts with "But fundamentally" ends with [backtick]gpplot2 (I can't put 1 backtick into a comment so that it is displayed as backtick) which seems to be something mistyped (presumably "2" should be another backtick, shouldn't it? or just another backtick is missing? since I'm not familiar with `r`, I can't really say) — YakovL, Mar 19 '18 at 12:22

R-project Graphic plot

1 Answers1