I tried to use the Kernel Density plot method from Hayfield and Racine (2008)
np package for my own data, but somehow ended up with different type of plots and I have no idea what the difference is between my data and the example data provided by the package.
I used the Italy GDP example that is provided in the np package and described in Racine's primer. The example itself worked well and I ended up with a nice 3D plot.
When I used it with my own data I just got two 2D plots instead of one 3D plot. Why is that? Do I miss something content wise (that's what I think and it's the reason being why I guess this is rather a CV than a SO question) or is it a syntax problem?
# get a bandwith object bw2 <- npcdensbw(formula=mydf$values~ordered(mydf$datefield), tol=.1, ftol=.1) str(mydf) # returns data.frame': 780 obs. of 2 variables: $ datefield: Ord.factor w/ 104 levels "1984-04-01"<"1984-07-01"<..: 1 1 1 1 1 1 1 1 1 1 ... $ values : num 50.7 58.5 56.1 55.5 62.7 ... tables(mydf$datefield) # shows that I have 13 entries per quarter. quartz() plot(bw2) npplot(bw2) # same 2D plot for both :(
all of these properties are exactly in line with the example data. Still I get only 2D plots instead of a 3D plot. I do have three dimensions: quarters,values and conditional density. What's wrong?