User Tools

Site Tools


en:mix:logo

Logo (of heart) on the top of this website

This is just simple horseshoe artefact, which arises from linear ordination analysis (PCA in this case) when applied on heterogeneous data. I used simulated community dataset, structured by one artificial “environmental” gradient, and applied principal component analysis (PCA) on it. The R code which I used to generate this figure is below:

simul1.spe <- read.delim ('http://www.davidzeleny.net/anadat-r/data-download/simul1-spe.txt', row.names = 1)
#png (width = 128, height = 98, bg = NA)
par (mar = c (2,2,1,1))
ordiplot (rda (log1p (simul1.spe)), display = 'si', axes = F, type = 'n')
points (rda (log1p(simul1.spe)), bg= 'white', pch = 3, col = 'red', cex = .7)
title (xlab = 'PCA1', line = .5, ylab = 'PCA2', cex.lab = 0.8)
#dev.off ()

(you may remove # to get *.png file with the exact size of logo and without white background)

en/mix/logo.txt · Last modified: 2017/03/21 03:10 by David Zelený