Histogram ggplot2

[Wasteva]

 Can someone help me to include corresponding percentages of histograms sit on the individual blocks. Please I don't mean percentages on the y-axis.
Thanks.

[Brandon Hurr]

Can you please provide an example dataset and your current working code to produce said histogram?

See the links below about providing a reproducible example if you're stuck. dput() is your friend and so is Legion™. 

On Wed, Oct 24, 2012 at 7:13 PM, Wasteva <steve...@gmail.com> wrote:

 Can someone help me to include corresponding percentages of histograms sit on the individual blocks. Please I don't mean percentages on the y-axis.
Thanks.

--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
 
To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

[Brandon Hurr]

On Wed, Oct 24, 2012 at 7:30 PM, Steve Ampah <steve...@gmail.com> wrote:

x<- c(rnorm(30),rnorm(30))
y<-rep(c("A","B","C"),20)

data<-data.frame(x,y)

 

ggplot(data,aes( x )+geom_histogram(binwidth=1)+facet_grid(y~.,margins=T)

 

I want a histogram of x across y so that with each A the total is 100% , that is if within A there are 4 blocks I expect percentages like 10%, 30% , 40%, 20%, and something similar for B and C.

You'll have to bin things outside of ggplot and then compute the percentages for each bin. 

From: https://stat.ethz.ch/pipermail/r-help/2008-January/151594.html

If you want the SAME bins as used by hist(), you can use:

hist(Vec, breaks = ..., plot = FALSE)$counts

Another alternative is to use cut() to bin a vector and return a factor, 
upon which you can then use table() to get the counts. For example:

   table(cut(Vec, breaks = ...))

See ?hist and ?cut for more information.

HTH,

Marc Schwartz 

[Ista Zahn]

Hi,

On Wed, Oct 24, 2012 at 2:13 PM, Wasteva <steve...@gmail.com> wrote:
> Can someone help me to include corresponding percentages of histograms sit
> on the individual blocks. Please I don't mean percentages on the y-axis.

There are really two issues here: 1) getting percentages on the
y-axis, and 2) getting the labels on top of the bars.

For the first one:

ggplot(mtcars, aes(x=mpg, y=..count../sum(..count..))) +
geom_histogram() +
scale_y_continuous("Percent", labels=percent)

For the second one:

ggplot(mtcars, aes(x=mpg, y=..count.., label=..count..)) +
geom_histogram() +
geom_text(stat="bin")

putting it all together we get

ggplot(mtcars, aes(x=mpg, y=..count../sum(..count..),
label=scales::percent(..count../sum(..count..)))) +
geom_histogram() +
geom_text(stat="bin") +
scale_y_continuous("Percent", labels=percent)

Hope that helps!
Ista

[Wasteva]

Bravo Isata, it worked but there are 0%0%0%0%0% intersecting the x-axis as I have written what could that be.
Thanks

[Wasteva]

Again how do I get the x-values to stay in the middle of the bins instead of staying to the left of the bins.
Thanks.

[Ista Zahn]

There are bins where the count is zero, and these are labeled as well.
Perhaps you should consider increasing the binwidth.

[Ista Zahn]

On Wed, Oct 24, 2012 at 4:19 PM, Wasteva <steve...@gmail.com> wrote:
> Again how do I get the x-values to stay in the middle of the bins instead of
> staying to the left of the bins.

I don't follow the question...