Readability

The main outputs provided by the MCA() function of the FactoMineR package is a graphical representation of both individuals and categories in a two-dimensional graph thqt allows an easy interpretation of the data.
When a large number of poeple are surveyed, too many points are to be represented (either individuals or categories) which prevents the user from easy decision-making.

To make the interpretation of such graphs easier, the EnQuireR package propose two points of view via the functions ENlisib() and ENdensity().

ENlisib()

ENlisib() proceeds in two steps.
The first step consists in selecting the individuals and categories with a good quality of representation. That is to say individuals whose coordinates are upper than the threshold (which is equal to the mean plus one standard deviation) and categories that are significantly linked to the components provided by the MCA.
The second step consists in selecting randomly a given number or percentage of what remains.
Individuals and categories having been selected during these two steps are plotted.

data(tea)
res.enmca=ENMCA(tea[,1:18])
ENlisib(res.enmca$MCA,nbvar=0.05,nbind=50,axes=c(1,2)) #tea: the data set used whith the columns (i.e. the categorical variables) on which to perform the MCA
#nbvar: number or percentage for categories
#nbind: number or percentage for individuals
#axes: selected axes

Click to view

Click to view

ENdensity()

ENdensity() uses the distribution of the coordinates of the individuals to draw density curves to visualize the shape of the scatter plot.

data(tea)
res.enmca=ENMCA(tea[,1:18])
ENdensity(res.enmca$MCA,nlevels=20)

Click to view

data(tea)
res.enmca=ENMCA(tea[,1:18])
cluster=res.enmca$data[,ncol(res.enmca$data)]
ENdensity(res.enmca$MCA,cluster,nlevels=10) #tea: the data set used whith the columns (i.e. the categorical variables) on which to perform the MCA
#nlevels: number of contour levels desired

Click to view