Investigating the relationship between two attributes
The best way to investigate the relationships between attributes visually is to do it in pairs. The tools we use for investigating the relationship between a pair of attributes depends on the type of attributes. In what follows, we will cover these tools based on the following pairs: numerical-numerical, categorical-categorical, and categorical-numerical.
Visualizing the relationship between two numerical attributes
The best tool for portraying the relationship between two numerical attributes is the scatter plot. In the following example, we will use a tool called scatter matrix that creates a matrix of scatterplots for a dataset with numerical attributes.
Example of using scatterplots to investigate relationships between numerical attributes
In this example, we will use a new dataset, Universities_imputed_reduced.csv. This dataset's definition of data objects is Universities in the USA, and these data objects...
