Defining new variables
When given a data set to analyse, think about whether its variables are the most useful ones to analyse. Sometimes a simple transformation provides a variable whose values are more meaningful or highlight a different aspect of the data.
For example, although the raw GDP of countries indicates their relative importance, their GDP per capita gives better comparison of the relative wealths of individuals in these countries:
GDP per capita = | Country's total GDP |
![]() |
|
Population |
As another example, although it is interesting to compare the calorie intake of countries, it is also appropriate to look at their percentage change over different periods:
%Increase = 100 × | 1998 calories − 1993 calories |
![]() |
|
1993 calories |