Parallel Sets is a multivariate visualization technique used for showing relationships between dimensions and categories. This technique is good for exploring small sets of ordinal and nominal data.
Parallel Sets are always better than a pie chart!
Each horizontal bar represents all the categories across one dimension. Each category bar is sized based on the percent of data in that category relative to the dimension. Starting with the first dimension, each category is connected to each category of the dimension below. These connections continue across all dimensions making this visualization useful for exploring "paths" of data like an icicle plot. The dimensions and categories can be rearranged to further explore the relationships between them.