These data measure protein consumption in twenty-five European countries for nine food groups. It is possible to use multivariate methods to determine whether there are groupings of countries and whether meat consumption is related to that of other foods.
A complete linkage cluster analysis of all nine variables shows that the diets of the countries can be divided into several groups according to their geographic locations. In the plot below, each group has been denoted by a different color. From top to bottom in the plot, the groups are: Eastern Europe (blue) containing East Germany, Czechoslovakia, Poland, USSR, and Hungary; Scandinavia (green) containing Sweden, Denmark, Norway, and Finland; Western Europe (red) containing UK, France, West Germany, Belgium, Ireland, the Netherlands, Austria, and Switzerland; Iberian (purple) containing Spain and Portugal; Mediterranean (orange) containing Italy and Greece; and the Balkans (yellow) containing Yugoslavia, Romania, Bulgaria, and Albania.
Clustering the observations using single linkage gives different results that are not so easily interpreted.
A principal components analysis of the nine variables reveals four large dimensions of variability. The first principal component may be interpreted roughly as a measure of total meat consumption. The second, third, and fourth principal components may be interpreted as red meat, white meat, and fish consumption respectively.