Pie charts and **donut (pie) charts** are confusing. At most, a donut can show deeper relationships by using **concentric** Donut charts:  However, in either case, a different display, including **stacked bar charts**, are probably more appropriate. Another alternative to pie charts are **waffle charts** (aka. square area charts), especially if you are showing a progression:  As another example, the waffle chart is great for showing progression towards a target (e.g., a Key Performance Indicator):  **Stacked bar charts** should avoided, unless there is a clear advantage to use them over **grouped bar charts**: By stacking the bars, you can compare the total size of each group and understand each group's part-to-whole relationship, at the cost of making it harder to compare the individual components of each group. That is to say, stacked column charts can be the right pick when that group's size is more important:  A **tree map** makes sense when you want to group several areas together by color. You should consider if a (grouped or stacked) bar chart isn't the better option for your data. The advantage is that you can get both the grouping (comparison) and stacking (part-to-whole) visualization packed into a single chart:  A more complex visualization of hierarchical relations *and* a part-to-whole version of the concentric Donut chart is the **Sunburst diagram** that can visualize more faceted, tree-like data, much like a tree map:  *Heat maps* come in many variations; Often, they are used as a standard grid to [display a trend](Display%20a%20trend.md) or to [compare variables](Compare%20variables.md). However, you can create a **Clustermap** to group similar covariates together and thereby visualize hierarchical relationships:  For example, such heat maps can be used to group correlated genes or stock tickers: