Histogram Outline Colour

[Nick S]

I am exploring the use of some example code I stumbled across recently
which produces back-to-back histograms.

I am having trouble figuring out how to highlight/outline the bins in
black on the top portion of the graph using a call to the colour
parameter. I can't seem to find an example of what I want to do in the
documentation. Perhaps I have just missed it.

x data can be found here: http://dl.dropbox.com/u/3734701/x.csv
y data can be found here: http://dl.dropbox.com/u/3734701/y.csv

When I try to add the following code...I get the following behavior
(http://dl.dropbox.com/u/3734701/back-to-back_2.jpg). I would like
both histograms to have black outlines for the bins.

qplot(x,geom="histogram", fill="Pre-BBXL",
binwidth=0.25,colour="black") +
geom_histogram(data=data.frame(x=y),aes(fill="Post-BBXL",
y=-..count..),binwidth= 0.25,colour="black") +
labs(x="ODs",y="count",fill="") +
xlim(0,14) +
scale_fill_brewer("Set1")

Any advise is appreciated!

[Hadley Wickham]

Hi Nick,

This is a case where it's easier to create the plot with ggplot, not qplot:

ggplot(mapping = aes(fill = "Pre-BBXL")) +
geom_histogram(aes(x), data = x, binwidth=0.25, colour="black") +
geom_histogram(aes(y), data = y, binwidth=0.25, colour="black") +
labs(x="ODs",y="count") +

xlim(0,14) +
scale_fill_brewer("Set1")

Hadley

> --
> You received this message because you are subscribed to the ggplot2 mailing list.
> Please provide a reproducible example: http://gist.github.com/270442
>
> To post: email ggp...@googlegroups.com
> To unsubscribe: email ggplot2+u...@googlegroups.com
> More options: http://groups.google.com/group/ggplot2
>

--
Assistant Professor / Dobelman Family Junior Chair
Department of Statistics / Rice University
http://had.co.nz/

[Nick S]

Thanks for the reply Hadley.

I get the following error when trying to use the code you provided:
Error: ggplot2 doesn't know how to deal with data of class numeric

Any ideas?

[Joshua Wiley]

My guess would be that your data is not in a data frame. Is this what
you want? I just filled the difference in lengths between your two
data files with NAs.

##########
library(ggplot2)
x <- read.table(file="http://dl.dropbox.com/u/3734701/x.csv", sep=",",
header=TRUE)
y <- read.table(file="http://dl.dropbox.com/u/3734701/y.csv", sep=",",
header=TRUE)
dat <- data.frame(x=c(x$x, rep(NA, (nrow(y)-nrow(x)))), y=y$x)

ggplot(data=dat, mapping = aes(fill = "Pre-BBXL")) +
geom_histogram(aes(x=x), binwidth=0.25, colour="black") +
geom_histogram(aes(x=y, y=-..count..), binwidth=0.25, colour="black") +

labs(x="ODs",y="count") +
xlim(0,14) +
scale_fill_brewer("Set1")

############

--
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/

[Nick S]

Thanks for the reply.

What I really want to do is have the top half of the histogram be
displayed in one color and the bottom half be displayed in a separate
color.

I have the data in the following format...one column with numeric
values and one column with the categorical variable I would like to
use to differentiate between the two groups (http://dl.dropbox.com/u/
3734701/data.csv)

Is it possible to do this with the data in this format? In previous
examples...I had no luck in displaying the associated category for
each side of the histogram.

[Joshua Wiley]

On Thu, Jul 1, 2010 at 10:43 AM, Nick S <n.si...@gmail.com> wrote:
> Thanks for the reply.
>
> What I really want to do is have the top half of the histogram be
> displayed in one color and the bottom half be displayed in a separate
> color.
>
> I have the data in the following format...one column with numeric
> values and one column with the categorical variable I would like to
> use to differentiate between the two groups (http://dl.dropbox.com/u/
> 3734701/data.csv)
>
> Is it possible to do this with the data in this format? In previous
> examples...I had no luck in displaying the associated category for
> each side of the histogram.

Colouring the different histograms is not an issue with that format of
data, but I am not sure how to specify ..count.. and -..count..
without setting almost everything manually. Have you considered
faceting at all? I think it easier than trying to read an upside down
histogram. Here's an example:

##################

dat <- read.table(file="http://dl.dropbox.com/u/3734701/data.csv",
header=TRUE, sep=",", stringsAsFactors=FALSE)
dat[ , "category"] <- factor(dat[ , "category"],
levels=c("Pre-BBXL","Post-BBXL"), labels=c("Pre-BBXL","Post-BBXL"))
str(dat)
names(dat)

#########
library(ggplot2)

ggplot(data=dat, aes(x=totalODdna, group=category, fill=category)) +
geom_histogram(binwidth = 0.25, colour = "black") +
facet_wrap(~category) +
xlim(0, 14) +
scale_fill_brewer("Set1")

##################

>
> On Jun 30, 6:29 pm, Joshua Wiley <jwiley.ps...@gmail.com> wrote:

<snip>