problem-4.6

problem-4.6  A table shows the relationship between manufacturer and vitamin type by shelf location:
## load and attach data set
> library(MASS); attach(UScereal)
> table(mfr, vitamins, shelf)
, , shelf = 1

, , shelf = 3

   vitamins
mfr 100% enriched none
  G 3    6        0
  K 2    8        0
  N 0    1        0
  P 0    6        0
  Q 0    1        1
  R 0    1        0
    
The most obvious relationship is that 100% fortified cereal appears only on the highest shelf. This leaves space for the kid-desirable, sugar-laden cereals on the coveted second shelf.
The bubble plot is produced with the commands
> plot(calories ~ sugars, data=UScereal, cex=2*sqrt(fat))
    
There appears to be a linear relationship between sugars and calories, as expected. The larger bubbles appear to be above the smaller ones, indicating that more fat implies more calories as well.
A pairs plot (pairs(UScereal)) shows many relationships. For example, the fibre, shelf plot shows some relationship. A better plot than the scatterplot for showing a factor and a numeric variable is the boxplot:
> boxplot(fibre ~ shelf, data=UScereal)
    
The high-fiber cereals live on the top shelf for presumably the same reason that the fortified cereals do.