Introduction
Know your problem
3. Typical questions
Know your data
Visualize your data – numerical variables
Work with your chart
Check yourself

Instruction

To really understand our problem (and therefore select the proper chart) we have to ask the right questions. As we've seen, the type of questions we ask often tell us the problem we are trying to solve.

Typical relationship questions are:

  • Are variables independent of each other (i.e. the value of one variable doesn't determine the value of the second variable)? Or is there a correlation between them – a relationship?
  • If there is a relationship, what kind is it? For example, do the first variable's values decrease as the second variable's values increase?
  • Are there some unusually frequent or unusually rare combinations of values of both variables? What are these combinations?

If at least one of these questions is posed during the data analysis phase, you want to focus on the relationship between variables.

The above questions refer strictly to the problem of relationship – they ask for trends in data. But we can also ask about specific variables. This will tell us more about what type of chart will best represent the data. Specific questions about variables include:

  • What are the rarest values for each variable? What are the most frequent?
  • What are the minimal and maximal values of each variable?
  • What is the range of values for each variable?

Later, we will verify whether our chart type can answer these questions.