How to overlay point density (ggplot2) on OSM tile in R?

Question

I have question regarding my R script: How do I overlay a point_density using google map or osm tile as background in R? 1) This works fine, but is not what I want:

library(ggmap)
map <- get_map("Vilnius", zoom = 14, source = "osm", color = "bw")
mapPoints <- ggmap(map)
p<-ggplot(geotag@data, aes(Y, X))
p + stat_density2d(aes(fill= ..level..), alpha =0.5, geom="polygon") + geom_point(aes(Y, X), colour="red", data = geotag@data, alpha = .5)

2) if I instead use a osm tile as background the point density is distorted (unfortunately I cannot show illustrate an image). It seems that the point density function does not recognize the point distribution and creates a symmetric overlay:

library(ggmap)
map <- get_map("Vilnius", zoom = 14, source = "osm", color = "bw")
mapPoints <- ggmap(map)
mapPoints + stat_density2d(aes(fill= ..level..), alpha =0.5, geom="polygon") + geom_point(aes(Y, X), colour="red", data = geotag@data, alpha = .5)

Here is some sample data:

OID_    Y   X
0   25.29315500000  54.68288700000
0   25.29375600000  54.68260200000
0   25.28416600000  54.68472200000
0   25.29776900000  54.68051900000
0   25.29549400000  54.68064800000
0   25.25535200000  54.67742600000
0   25.29541800000  54.68429700000
0   25.29751100000  54.68445600000
0   25.29541800000  54.68429700000
0   25.29541800000  54.68429700000
0   25.29751100000  54.68445600000
0   25.29751100000  54.68445600000
0   25.29751100000  54.68445600000
0   25.29751100000  54.68445600000
0   25.28865900000  54.68074300000
0   25.28943200000  54.67457100000
0   25.29133200000  54.68690000000
0   25.29176000000  54.68873800000
0   25.28049400000  54.67952800000

what I am doing wrong?

Thanks for any help

Answer 1

Is this what you're looking for? Calling your sample df:

library(ggmap)
map <- get_map("Vilnius", zoom = 14, source = "osm", color = "bw")
mapPoints <- ggmap(map)
mapPoints + 
  stat_density2d(data=df,aes(Y,X, fill= ..level..), alpha =0.5, geom="polygon") + 
  geom_point(data=df, aes(Y, X), colour="red", alpha = .5)

The problem was that, in the first snippet, the default dataset is set to geotag@data in the call to ggplot(...), whereas in the second snippet the default dataset is defined internally in the call to ggmap(...). In that case you have to explicitly define the (local) dataset, and the X and Y mappings, in the call to stat_density(...).