Multivariate Analysis and MDS Scratchpad

mds.r

library(MASS)
library(car)

makeProfilePlot <- function(list,words) {
  # find out the minimum and maximum values of the variables:
  mymin <- 1e+20
  mymax <- 1e-20
  for (i in 1:length(list)) {
    vectori <- mylist[[i]]
    mini <- min(vectori)
    maxi <- max(vectori)
    if (mini < mymin) { mymin <- mini }
    if (maxi > mymax) { mymax <- maxi }
  }
  # plot the variables
  for (i in 1:length(list)) {
    vectori <- list[[i]]
    wordi <- words[i]
    if (i == 1) { 
      plot(vectori, col="green", type="l",ylim=c(mymin,mymax)) 
    } else { 
      points(vectori,col="green", type="l")                                   
    }
    lastxval <- length(vectori)
    lastyval <- vectori[length(vectori)]
    #poor colour choices, I know!
    text((lastxval-10),(lastyval),wordi,col="red",cex=0.6)
  }
}

#update for your working directory
setwd("~/development/js/susan/")
data <- read.csv('data.csv')

#sample profile plot to show variance of
names <- c("Very","Really","So","Like","Sort of", "Kind of")
sample_comp_list<-list(data$very, data$really, data$so, data$like, data$sortof, data$kindof)
makeProfilePlot(sample_comp_list, names)

#remove categorical data, extract sample variables
numData <- data[11:length(data)]
scatterplotMatrix(numData)

#non metric mds, not including word count
data[11:length(data)]
d <- dist(data[11:length(data)])
fit <- isoMDS(d, k=2)
fit
x <- fit$points[,1]
y <- fit$points[,2]
plot(x, y, xlab="x", ylab="y", main="Nonmetric MDS", type="n")
text(x, y, labels = data$character, cex=.7)

#metric mds with no reliance on external libs
(mds <- cmdscale(d))
plot(mds, type = 'n')
text(mds[, 1], mds[, 2], data$character)

#TODO: visualisation of categorical variables (age, show type etc)