Currently, I have two data sets with different sizes. The data sets show the how many cells appear in a specific location within an area. Cause the data sets have different sizes. With using ggplot, I achieved the normalisation by aes(fill=..density..) to get the density (percentage) of each hex. For each hex, it's the number of cells in this hex divided by the total number. My question is how do I quantify the difference between these two plots? How can I get subtraction results of these two plots?
x1 <- runif(1000,-195.5,195.5)
y1 <- runif(1000,-49,49)
data1 <- data.frame(x1,y1)
x2 <- runif(2000,-195.5,195.5)
y2 <- runif(2000,-49,49)
data2 <- data.frame(x2,y2)
outline<-"grey50"
min=-225
max=225
brks <- (max-min)/4
plot_1 <- ggplot(data1,aes(x=x1,y=y1)) +
geom_hex(bins=30,color=outline,alpha=1, aes(fill=..density..)) +
scale_x_continuous(breaks= seq(min, max, brks), limits= c(min, max), labels=noDec) +
scale_y_continuous(breaks= seq(min, max, brks), limits= c(min, max), labels=noDec)
plot_1
plot_2 <- ggplot(data2,aes(x=x2,y=y2)) +
geom_hex(bins=30,color=outline,alpha=1, aes(fill=..density..)) +
scale_x_continuous(breaks= seq(min, max, brks), limits= c(min, max), labels=noDec) +
scale_y_continuous(breaks= seq(min, max, brks), limits= c(min, max), labels=noDec)
plot_2
I also tried another method. With using hexbin function from hexbin package firstly, and then apply the hdiffplot function. The problem is how can I get the quantified result (e.g.a number or whatever) instead of a plot as results?
xbnds <- range(data1$x1,data2$x2)
ybnds <- range(data1$y1,data2$y2)
hbin_1 <- hexbin(x=data1$x1,y=data1$y1,xbins=30,xbnds=xbnds, ybnds=ybnds,shape=98/391)
hbin_2 <- hexbin(x=data2$x2,y=data2$y2,xbins=30,xbnds=xbnds,ybnds=ybnds,shape=98/391)
erodebin1 <- erode.hexbin(smooth.hexbin(hbin_1))
erodebin2 <- erode.hexbin(smooth.hexbin(hbin_2))
hdiffplot(erodebin1,erodebin2)
Thanks in advance!
