R code solutions - mdepasca/miniature-adventure GitHub Wiki

Kernel density plot

visualise the density of redshift measures in a sample using R. The simplest way is to use the density function, provided by R-base

Read-in the data

df <- read.csv2('products/SIMGEN_PUBLIC_DES_REDSHIFTS.csv', dec='.')

plot(density(df$redshift))

A bandwidth for the kernel (in this case the default gaussian) can be provided by specifying the bw keyword.

Comparative kernel density plot

To compare redshift distribution in training and testing the sm package can be used (it has to be installed). To disinguish the two groups, the train_flag value will be used.

install.packages('sm')
library(sm)

sm.density.compare(df$redshift, df$train_flag)

train.factor <- factor(df$train_flag, levels=c(1,0), 
                labels=c("Test", "Train"))  # to be use in legend
colfill <- c(2:(1+length(levels(train.factor))))
legend(locator(1), levels(train.factor), fill=colfill)