I have a .ppp object of trees in a forest. It has two marks: ash disease and species codes. The ash disease obviously only affects the ash trees; the column reads "NA" for the rest of the species. Nevertheless, I want to extract the four nearest neighbours of the ash trees, whatever species, and be able to then plot against the extent of the disease. In other words, to find a correlation between ash disease and different neighbour species.
Having extracted the nearest neighbours, how do I add that data frame (or column by column) back to the original data frame for later analysis?
library(spatstat)
x <- c(1, 3, 2, 4, 5, 6, 3, 2, 6, 7)
y <- c(3, 4, 5, 2, 2, 6, 7, 8, 4, 3)
ashdis <- c(0.3, 0.95, 0.05, rep(NA, 7))
species <- factor(c("ash", "ash", "ash", "oak", "oak", "beech", "maple", "hickory", "hickory", "beech"))
df <- data.frame(x, y, ashdis, species)
m <- data.frame(ashdis, species)
#convert to .ppp
X <- ppp(df$x, df$y, owin(c(0, 10), c(0, 10)))
marks(X) <- m
plot(X)
#then extract the nearest neighbours of each point in X:
nn <- nnwhich(X, k=1:4)
nn
#trying to add which.1 to data frame as a new column fails:
nn$nn1 <- nn$which.1
#I get this message: "Error in nn$which.1 : $ operator is invalid for atomic vectors"
#alternative way of extracting nearest 4 neighbours also produces error: "Sorry, not implemented when the marks are a data frame"
marktable(X, 4, )
plot(X, pch = 16, cols = "blue", size = 1)
plot(X)
#I should add that the above methods DO work when there is a single mark. But then I may lose the connection to the other variables?
#I also would be grateful for ideas on how to meaningfully combine the nearest neighbour columns into a group which can represent a species combo ~ ash disease - but that is more than a technical point
CodePudding user response:
Can't you just convert nn
to a data frame and cbind
to the original?
cbind(df, as.data.frame(nn))
#> x y ashdis species which.1 which.2 which.3 which.4
#> 1 1 3 0.30 ash 2 3 4 5
#> 2 3 4 0.95 ash 3 1 4 5
#> 3 2 5 0.05 ash 2 1 7 8
#> 4 4 2 NA oak 5 2 9 1
#> 5 5 2 NA oak 4 10 9 2
#> 6 6 6 NA beech 9 10 7 2
#> 7 3 7 NA maple 8 3 2 6
#> 8 2 8 NA hickory 7 3 2 6
#> 9 6 4 NA hickory 10 6 5 4
#> 10 7 3 NA beech 9 5 4 6