They show the same data using different representations.

They show different data subsets with different representations.

They can address the problem of change blindness when applied to time-series data.

They can be effective for comparing data from multiple time steps.