[R] ggplot2 multiple legends modification (resent witout attachments)

[David Winsemius]

On Oct 2, 2013, at 11:12 AM, rolf.kemper at renesas.com wrote:

> 
> I resent this as I recognized that attachments are not allowed.

Attachments are allowed, but only certain types . To attach data it should be MIME-text and the usual way to get attachments recognized as such jointly by your mail-client and the R mail-server is to name the file with a '.txt' extension. Extensions such as '.csv', '.dat', '.fil', '.doc' will not be forwarded to R-help subscribers. I'm not sure if '.rtf' files pass muster, so I'll just try in this reply:

-------------- next part --------------


I wonder if the googlegroups discussion forum for ggplot2 would allow attachmets of whatever type yours was?

https://groups.google.com/forum/?fromgroups#!forum/ggplot2



> Dear ALL,
> 
> please find attached a small test file. (code is copied in this mail too)
> Basically (as shown) I would like to combine several results in one graph.
> I could succeed to plot the data correctly, but the legend generation does not fit.
> For the dfStartPoints and the dfEndpoints I just want to be able to select the color AND (if possible the shape of each one)
> 
> The legend I'm getting could be modified for its title , but I was not able to to change the label value itself.
> 
> My final wish is to get 3 Legends (the one for dfXY is fully OK)
> Legend one (Titled Cell/Area...) is fully OK as it is
> Legend two should have Title "Start Point"s"  color "red" AND the count of data points as value
> Legend three should have Tile ?nd Points color "green"AND the count of data points as value.
> 
> I think I can easily get the count , but I have no idea how to modify the related  legends in code
> 
> There are many examples of this (e.g. http://docs.ggplot2.org/current/guide_legend.html )  , but  nothing fits really.
> 
> Hope someone can help on this
> 
> Rolf
> 
> ######################## code as attached ###############################
> require(sqldf)
> options(gsubfn.engine = "R")
> require(ggplot2)
> 
> rm(list=ls(all=TRUE))

Personally I consider it impolite to place that as uncommented code. I have utility functions that get blown away by that code. It would be more considerate to place it as a comment

# ------- To be uncommented to be more reproducible
# rm(list=ls(all=TRUE))


> maxCells = max(primitiveDensityMap$c)
> 
> #craete a 100 * 100 grid with random values(cellCount)
> dfXY <- data.frame(x=rep(0:99, each=100) , y=rep(0:99, each=1)  , cellCount=sample(0:99, 10000 , replace=T) )
> 
> dfStartPoints <- data.frame(x=sample(10:20, 3 , replace=T) , y=sample(30:40, 3 , replace=T)  )
> dfEndPoints <- data.frame(x=sample(70:80, 3 , replace=T) , y=sample(90:100, 3 , replace=T)  )
> 
> 
> 
> nrow(dfXY)
> #get max and min
> maxCells = max(dfXY$cellCount)
> 
> ggplot(dfXY, aes(x=x,y=y)) +
> 	geom_tile(aes(fill= cellCount), as= 1) +
> 	labs(title="Primitive Cell Density", x="mm", y="mm") +
> 	scale_fill_continuous(  space= "Lab" , low="white", high="black" , breaks=seq(0,maxCells ,by=round(maxCells/10) ), guide = guide_legend( title =
> "Cells/Area\n(100um^2)" ) ) +
> 	coord_fixed() +
> 	geom_point(data=dfStartPoints  , aes( colour="red"  ) ) +
> 	labs(color="Start Points" ) +
> 	geom_point(data=dfEndPoints  , aes( colour="green"  ) ) +
> 	labs(color="End Points" )
> 
> #######################################################################

David Winsemius
Alameda, CA, USA