[R] problem with hist()

[hadley wickham]

On 6/14/07, Mario Dejung <forum at dejung.net> wrote:
> Hey everybody,
> I try to make a graph with two different plots.
>
>
> First I make a boxplot of my data. It is a collection off correlation
> values of different pictures. For example:
>
> 0.23445 pica
> 0.34456 pica
> 0.45663 pica
> 0.98822 picb
> 0.12223 picc
> 0.34443 picc
> etc.
>
> Ok, I make this boxplot and I get for every picture the boxes. After this
> I want to know, how many correlations per picture exist.
> So I make a new vector y <- as.numeric(data$picture)
>
> So I get for my example something like this:
>
> y
> [1] 1 1 1 1 1 1 1 1 1 1
> [11] 1 1 1 1 1 1 1 1 2 2
> ...
> [16881] 59 59 59 60 60 60 60 60 60 60
>
> After this I make something like this
>
> boxplot(cor ~ pic)
> par(new = TRUE)
> hist(y, nclass = 60)
>
> But there is my problem. I have 60 pictures, so I get 60 different
> boxplots, and I want the hist behind the boxes. But it makes only 59
> histbars.
>
> What can I do? I tried also
> hist(y, 1:60) # same effect
> and
> hist(y, 1:61)
> this give me 60 places, but only 59 bars. the last bar is 0.
>
> I hope anyone can help me.

What does the y axis represent?  It will be counts for the histogram,
and correlations for the boxplots.  These aren't comparable, so you're
probably better off making two separate graphics.

Hadley