R help - Jun 2014 - Extracting data from density plots

I have a dataframe
​df ​
with 3 columns. Details
​of df are ​
as follows
> summary(df)
      Date                        TestVar            type
 Min.   :2002-05-10 00:00:00   Min.   :-3.8531   Bottom: 313
 1st Qu.:2005-05-09 12:00:00   1st Qu.:-0.7773   Other :2501
 Median :2008-05-07 00:00:00   Median : 0.2482   Top   : 313
 Mean   :2008-05-07 00:00:00   Mean   : 0.1980
 3rd Qu.:2011-05-05 12:00:00   3rd Qu.: 1.2250
 Max.   :2014-05-05 00:00:00   Max.   : 3.6633

> str(df)
'data.frame':   3127 obs. of  3 variables:
 $ Date   : POSIXlt, format: "2002-05-10" "2002-05-13"
"2002-05-14" ...
 $ TestVar: num  1.34 2.02 1.39 1.54 2.45 ...
 $ type   : Factor w/ 3 levels "Bottom","Other",..: 3 3 3 3
3 3 3 3 2 2 ...

I have plotted the following density plot for column 2 i.e. TestVar and then
color coded according to factor variable "type"

<http://r.789695.n4.nabble.com/file/n4691999/Rplot01.jpeg>
> ggplot(df)+  geom_density(aes(x=TestVar, color=type))


I now want to filter out data samples such that the density
​curve
of "Top" is
higher than "Bottom". Basically I need TestVar values for which blue
line is
higher than red line.

How can this be achieved? Is there a way i can extract density values out of
the plot?

	[[alternative HTML version deleted]]

I don't know if you can get the information from the plot, but you can
certainly get it from the density() function directly.  For example,

# fake data
TestVar <- rnorm(60)
type <- rep(c("Bottom", "Other", "Top"), 20)

# density of TestVar for each type, estimated at 100 points along range of
TestVar
d <- tapply(TestVar, type, density, from=min(TestVar), to=max(TestVar),
n=100)
results <- data.frame(TestVar=d$Top$x, sapply(d, "[[", 2))

# select only those records where the Top density is greater than the
Bottom density
results[results$Top > results$Bottom, ]

Jean



On Thu, Jun 12, 2014 at 10:25 PM, Parimal Lokhande <
parimallokhande@gmail.com> wrote:
> I have a dataframe
> ​df ​
> with 3 columns. Details
> ​of df are ​
> as follows
>
> > summary(df)
>       Date                        TestVar            type
>  Min.   :2002-05-10 00:00:00   Min.   :-3.8531   Bottom: 313
>  1st Qu.:2005-05-09 12:00:00   1st Qu.:-0.7773   Other :2501
>  Median :2008-05-07 00:00:00   Median : 0.2482   Top   : 313
>  Mean   :2008-05-07 00:00:00   Mean   : 0.1980
>  3rd Qu.:2011-05-05 12:00:00   3rd Qu.: 1.2250
>  Max.   :2014-05-05 00:00:00   Max.   : 3.6633
>
>
> > str(df)
> 'data.frame':   3127 obs. of  3 variables:
>  $ Date   : POSIXlt, format: "2002-05-10" "2002-05-13"
"2002-05-14" ...
>  $ TestVar: num  1.34 2.02 1.39 1.54 2.45 ...
>  $ type   : Factor w/ 3 levels "Bottom","Other",..: 3 3
3 3 3 3 3 3 2 2 ...
>
> I have plotted the following density plot for column 2 i.e. TestVar and
> then
> color coded according to factor variable "type"
>
> <http://r.789695.n4.nabble.com/file/n4691999/Rplot01.jpeg>
>
> > ggplot(df)+  geom_density(aes(x=TestVar, color=type))
>
>
>
> I now want to filter out data samples such that the density
> ​curve
> of "Top" is
> higher than "Bottom". Basically I need TestVar values for which
blue line
> is
> higher than red line.
>
> How can this be achieved? Is there a way i can extract density values out
> of
> the plot?
>
>         [[alternative HTML version deleted]]
>
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
>
	[[alternative HTML version deleted]]