80 types of charts for data visualization. Part 3: Part-to-Whole and Hierarchical
by Ivan Kilin — Data Visualization Specialist & Dieuwertje van Dijk — Data Visualization Designer @
Recently we’ve rewritten our essential article about 80 types of charts. Our main goal was to make the information structured and helpful right away. We talk about every chart starting from the look and feel and then going into practical advice on when you should and when you shouldn’t use this particular chart type. We will start publishing parts of this article here. One chart group per each article.
In the third part, we’ll talk about part-to-whole graphs. Part-to-whole graphs are used to represent how individual pieces contribute to a complete picture. They focus on composition and proportions. Stacked bar chart in majority of the cases tend to be the best option amount part-to-whole graphs.
Let’s explore part-to-whole graphs one by one. Keep on reading!
Stacked bar chart & stacked column chart
What is a stacked bar chart?
A stacked bar chart builds on a regular bar (or column) chart by slicing each bar into coloured segments. The full bar represents the total for a category, and each segment shows how much a sub‑category contributes to that total. If you scale every bar to the same length, you create a 100 percent stacked version. In this case, the segments display relative shares rather than absolute values.
When should you use a stacked bar chart?
Use a stacked bar chart to show the overall size of each category and how its sub‑categories contribute to that total. It works best when the number of segments is small and their order is consistent. This way, the reader can quickly see which category is largest, what makes it up, and how those parts compare across categories. A 100 percent stacked version is handy when you care more about proportional shares than absolute values.
When you shouldn’t use a stacked bar chart
Avoid stacked bars if there are many segments. Thin stacks are hard to label and even harder to compare. They are also a poor choice when the goal is to compare individual segment sizes across categories. This is because only the bottom segments share a common baseline. If precise segment‑to‑segment comparison matters, a grouped bar chart or small multiples will be clearer.
Stacked bars also struggle with negative values or datasets where segments appear in some categories but not others. Both make the visual harder to read.
Diverging (stacked) bar/column chart
What is a diverging bar chart?
A diverging bar chart is a variation of a regular bar chart. Unlike a standard bar chart, it has a baseline in the center, usually set at zero. Bars extend in both directions from this midpoint. This makes it easy to compare values that move in opposite directions. For example, bars on the left often show negative values or sentiment, while bars on the right show positive values or sentiment.
The design typically relies on a diverging color palette. It consists of one hue for the bars that extend left and a contrasting hue for those that extend right. This way, readers can instantly see which side of zero each value falls on.
A common variation is the diverging stacked bar chart. In this chart, each side of the baseline is split into colored segments to show how different subgroups contribute to the negative or positive totals.
When should you use a diverging bar chart?
Choose a diverging bar chart when your measure can move in two directions and you want to compare those directions side by side. It is ideal for data that mixes positive and negative values. Examples are profit versus loss, approval versus disapproval, or temperature above and below average. Survey results on Likert scales are a common example, but any dataset with a natural midpoint (zero or neutral) can benefit from this format.
When you shouldn’t use a diverging bar chart
Do not pick a diverging bar chart unless your data truly has a meaningful midpoint. The centred baseline implies a clear split into two opposite directions, often negative on the left and positive on the right. If your values are all positive, all negative, or lack a natural “zero,” this chart will mislead.
Also, limit the number of stacks. When there are many thin segments, it becomes hard to see or label them, and even harder to compare matching stacks across categories. Finally, if readers need to measure exact differences between individual segments, a grouped bar chart or small multiples will be easier to read.
Population pyramid
Alternative names: Age-sex pyramid, Age structure diagram
What is a population pyramid?
A population pyramid looks like a diverging bar chart. However, it is used only to show how a population is distributed by age and sex. Bars for one sex extend to the left of a central baseline, and bars for the other sex extend to the right. Each horizontal layer represents an age group. So the overall shape reveals whether a population is young, ageing, or balanced.
When should you use a population pyramid?
Choose a population pyramid when your data list age together with exactly two gender categories. It gives a quick snapshot of how each age band splits between the two and is standard in demographic and population studies to spot youth bulges, ageing trends, or gender imbalances.
When you shouldn’t use a population pyramid
Skip a population pyramid if your data is not limited to two sex categories. The classic left‑versus‑right layout forces everything into a male‑female split and hides people who identify outside that binary. In this case, use a chart that can show three or more groups. Examples are side‑by‑side bars or small multiples.
A population pyramid is also the wrong chart type when age bands are uneven or for comparing many regions in one frame. A grouped bar chart or small multiples will be clearer in those cases.
Icon array
Alternative name: Pictograph
What is an icon array?
An icon array is a grid of small icons arranged in rows and columns. The entire grid represents the total number of units in the dataset. Each icon stands for one unit, such as a person, a vote, or an item. Icons that belong to the same subgroup share a colour or shape, letting you see quickly how the whole divides into its parts. A ten‑by‑ten grid is common because it neatly shows one hundred units, but other grid sizes can be used just as well.
When should you use an icon array?
Use an icon array when you need a quick part‑to‑whole picture that anyone can understand instantly. Each icon equals one unit, so readers can count symbols instead of decoding percentages or tables. The format suits survey results, market shares, or demographic splits. It also works for simple yes‑or‑no data where filled icons show “yes” and the remainder show “no.” Originating with the Isotype movement, icon arrays remain a clear way to turn numbers into pictures.
When you shouldn’t use an icon array
Avoid icon arrays when counts are very large or uneven, for example 13 487. A grid that big is cluttered, and rounding can mislead. This type of chart also struggles with small differences. The smallest change you can make, adding or removing a single icon, often represents a sizable chunk of the data. Moreover, with more than a few categories, the chart becomes cluttered. If there are more than 5 categories, we recommend adding a “rest” category.
Visit our icon array resource page for pro tips on designing clear and effective icon arrays.
Waffle chart
Alternative names: Square pie chart, Square area chart, Gridplot
What is a waffle chart?
A waffle chart looks like an icon array but replaces icons with a grid of 100 squares cells. Each cell equals one percent of the total. By colouring groups of cells, you can show how much of the whole is complete, how far you are from a goal, or how several categories contribute to one hundred percent.
When should you use a waffle chart?
Choose a waffle chart when you need a quick, part‑to‑whole picture that even non‑technical readers can understand. It works well for progress indicators and survey results, and any other data where you want to turn percentages into a neat, countable grid. Because its layout is square, some audiences find it easier to read than a pie chart.
When you shouldn’t use a waffle chart
Skip a waffle chart if your numbers are not close to whole percentages, since you must round to the nearest cell, and that can distort small values. It also becomes cluttered when you have more than three or four categories. In this case, the coloured blocks start to blend, and labels get hard to place. For precise comparison between multiple groups, a bar or stacked bar chart will be clearer.
Pie chart
Alternative names: Pie graph, Pizza chart, Circle chart
What is a pie chart?
A pie chart is a circular graphic for showing how a total breaks into parts. The full circle represents the whole, while each slice represents a category’s share of that whole. The slices must add up to a meaningful total, which is often 100 percent. If your data are absolute numbers such as dollars or units, all categories together should equal a sensible total.
When should you use a pie chart?
Use a pie chart when you have only a few categories and you want readers to see their relative sizes instantly. It is especially effective when one slice is much larger or smaller than the others, because the difference stands out immediately.
When you shouldn’t use a pie chart
Avoid a pie chart when you need precise comparisons. Human vision judges angles and curved areas poorly, so small differences between slices are hard to read. The chart is also limited if you have many categories, because narrow wedges become crowded and labels overlap. For detailed or multi‑category comparisons, a bar chart, stacked bar, or another linear chart will communicate the data more clearly.
Read the deep dive pie chart article to see our arguments for using pie charts. And if you want to create a really good pie chart yourself, don’t miss out on the pie chart resource page full of pro design tips.
Donut chart
Alternative names: Doughnut chart
What is a donut chart?
A donut chart is a pie chart with a circular hole in the centre. The ring still shows how a total splits into slices, but the empty middle frees up space for extra information.
When should you use a donut chart?
Choose a donut chart when you want the simple part‑to‑whole message of a pie chart plus room in the centre for a title, key metric, or icon. The ring shape also shifts the read from comparing wedge areas to comparing arc lengths, which many viewers find a little easier. Use it for small sets of categories where one or two slices dominate and a quick visual impression is enough.
When you shouldn’t use a donut chart
Skip a donut chart if you need precise comparisons or have many categories. Thin wedges are hard to judge and to label, so a stacked bar chart or grouped bar chart will show detailed differences more clearly.
For your convenience, we also created a donut chart resource page with valuable design tips for your next donut chart.
Semicircle donut chart
Alternative name: Half moon chart, Half donut chart, Semi-circle doughnut chart
What is a semicircle donut chart?
A semicircle donut chart is a donut chart sliced in half. All the pieces still add up to 100 percent, but they are arranged over only half of the circle. This forms a half‑moon shape with a straight edge at the bottom.
When should you use a semicircle donut chart?
Use a semicircle donut when you want the part‑to‑whole story of a donut chart, yet need extra blank space for labels, numbers, or a pointer. The flat edge also makes it a handy base for simple gauge visuals that show progress toward a target.
When you shouldn’t use a semicircle donut chart
Avoid this chart for precise comparisons or when you have many categories. Narrow wedges become hard to read, and judging arc lengths in a half‑circle is even tougher than in a full circle. In those cases, a stacked or grouped bar chart will communicate the data more clearly.
Explore more pie and donut chart ideas on our inspiration page.
Marimekko chart
Alternative names: Mekko chart, Mosaic chart, Mosaic plot
What is a Marimekko chart?
Think of a Marimekko chart as stacked bars that can vary in width. The width of each bar shows one measure. Each bar is then divided into coloured stacks whose heights show a second measure. All the bars fit together without gaps, so you can see both how big each bar is compared with the total and how each bar’s pieces break down inside it.
When should you use a Marimekko chart?
Use a Marimekko chart to compare part‑to‑whole relationships across two dimensions at once. It is popular in marketing and sales analysis. It can show, for example, how revenue is split by region (column width) and by product mix inside each region (segment height). Any dataset that needs to display two nested proportions in a single view can benefit from this format.
When you shouldn’t use a Marimekko chart
Skip a Marimekko chart if there are many columns or many small segments. Narrow columns and thin slices are hard to label and even harder to compare. The chart is not ideal for precise reading. So if you need exact comparisons, a grouped bar chart, a stacked bar chart, or a pair of simpler bar charts will be clearer.
Treemap
What is a treemap?
A treemap is a rectangular diagram that shows how a total breaks down into parts. Each top‑level category gets its own rectangle, and you can split these into smaller rectangles for sub‑categories. The area of every rectangle is sized in proportion to the value it represents. This lets you instantly see which parts are large and which are small.
When should you use a treemap?
Choose a treemap when you have many categories arranged in a hierarchy and need to display all of them in one compact view. The chart makes good use of space, so it can show a large dataset on a single screen while still revealing relationships within and between groups. This makes treemaps popular for file‑system usage, market share by brand and product line, budget breakdowns, and similar tasks.
When you shouldn’t use a treemap
Avoid a treemap if it will contain too many tiny rectangles. An overfilled grid becomes hard to read and overwhelms the viewer. Treemaps also struggle with precise comparison because people judge rectangular areas less accurately than bar lengths. When exact values or fine differences matter, a bar chart, stacked bar chart, or separate small multiples will communicate the data more clearly.
For more background, a deep‑dive article and a detailed treemap resource page are available if you would like extra tips before building your own.
Circular treemap
Alternative name: Circular packing, Circle packing
What is a circular treemap?
A circular treemap uses circles instead of rectangles to show a hierarchy. Each top‑level category is drawn as a large circle, and these can be divided into smaller circles for sub‑categories. The area of every circle matches the value it represents, so larger values appear as larger circles within the overall packing.
When should you use a circular treemap?
Choose a circular treemap when you want to display a hierarchy in a way that feels organic and visually engaging. Although it is less space‑efficient than a rectangular treemap, a well‑designed circle packing can be eye‑catching and easy for viewers to explore. It works best for overviews where the exact proportions are less critical and visual appeal is a priority.
When you shouldn’t use a circular treemap
Avoid a circular treemap when precise comparison matters more than showing the hierarchy. People judge circle areas less accurately than rectangle sizes or bar lengths, so fine differences between values can be hard to spot. If space is tight or you need exact comparisons, a rectangular treemap or a bar chart will serve better.
Convex treemap
Alternative names: Voronoi treemap, Polygonal partition
What is a convex treemap?
A convex treemap visualises hierarchical data in the same way a regular treemap does. However, it replaces rectangles with convex polygons. These are shapes with no inward dents, such as hexagons or irregular blobs whose edges always bulge outward. A simple way to tell if a polygon is convex is the following. Draw a straight line between any two points inside the shape, and that line will stay entirely inside the shape.
Because these polygons can be warped and packed into almost any outline, the entire chart can fit inside shapes such as circles, triangles, or custom silhouettes. Each polygon’s area still represents the value of a category, and polygons can be further subdivided to show sub‑categories. So the hierarchy remains intact while the overall layout looks more organic and space‑efficient than a strict rectangular grid.
When should you use a convex treemap?
Use a convex treemap when the chart must fit a non‑rectangular shape or avoid skinny tiles in deep hierarchies. Its flexible polygons suit infographics with custom outlines, dashboards with tight spaces, and visuals that benefit from an organic look.
When you shouldn’t use a convex treemap
Avoid a convex treemap when exact area comparison matters. Irregular polygons are harder to judge than rectangles, and the view turns cluttered when many small categories appear. This layout also takes more computation to generate. If a rectangular grid will do, a standard treemap or bar chart is clearer.
For an example of a well‑designed convex treemap, see our treemap deep‑dive article.
Dendrogram
Alternative name: Phylogenetic tree
What is a dendrogram?
A dendrogram is a tree‑shaped diagram that shows how items cluster into a hierarchy. It begins with each item as an individual leaf; branches then merge step by step until all items join a single trunk. The height where two branches join represents the distance between those items. Clusters near the bottom reflect the raw data most directly and are the most reliable. As you move up the tree the merges rely on earlier averages, so the relationships become progressively less precise.
When should you use a dendrogram?
Use a dendrogram when you need to present the results of hierarchical clustering. Biologists rely on it to map evolutionary links among species, while other fields use the same tree layout for gene clustering, customer segmentation, or grouping similar documents. The branching structure lets readers see which clusters lie close together and where the major splits occur.
When you shouldn’t use a dendrogram
Avoid a dendrogram if the hierarchy is shallow or the dataset is huge. Too few levels give a tree that could be shown more clearly as simple groups. On the contrary, hundreds of leaves produce a tangle of lines that is hard to read. If your main goal is to compare cluster sizes or values, use a bar chart or heatmap instead.
Venn diagram
Alternative name: Set diagram, Logic diagram
What is a Venn diagram?
A Venn diagram uses overlapping circles to show every possible intersection among two or more sets, even when some intersections contain no items. Shared regions reveal the elements that the sets have in common. Non‑overlapping regions hold elements unique to each set. Because the layout includes empty overlaps as well as real ones, it presents the full logical picture of how the sets could relate.
When should you use a Venn diagram?
Choose a Venn diagram when you want a familiar, easy‑to‑read picture of how two or three sets relate, including places where they could overlap but do not. This makes it useful in lessons, presentations, and simple reports where highlighting both real and potential intersections helps explain the concept.
When you shouldn’t use a Venn diagram
Skip a Venn diagram if you have more than three sets or if you care only about the overlaps that actually exist. Extra circles or empty intersections quickly make the graphic cluttered. In those situations, an Euler diagram, which shows only real overlap, will be clearer.
Euler diagram
What is an Euler diagram?
An Euler diagram (pronounced “OY‑ler”) looks a lot like a Venn diagram, but it shows only the overlaps that actually exist in the data. In a Venn diagra, every possible intersection appears, even if no items fall in those intersections. An Euler diagram omits any empty intersections. So each region of the drawing corresponds to a real‑world set or overlap and nothing more.
When should you use an Euler diagram?
Use an Euler diagram to show only the overlaps that truly exist. It works well for two or three sets that share some items yet stay distinct elsewhere. This lets readers focus on real connections without empty intersections.
When you shouldn’t use an Euler diagram
Do not use an Euler diagram when you must display every possible intersection, even ones that are empty. It is also limited when you have many sets or very complex overlaps, as the visual gets messy. In those situations, choose a matrix chart instead.
If you’re curious to understand it better, we recommend this article that explains the difference between the Venn and Euler diagram.
Circular gauge
Alternative names: Angular gauge, Radial gauge chart
What is a circular gauge?
A circular gauge shows a single value on a round or half‑round scale. A needle (or pointer) sweeps across the scale to mark the current reading, much like the dial on a speedometer or an analogue clock. Coloured bands on the arc can signal safe, warning, or critical zones.
When should you use a circular gauge?
Use a circular gauge when you need a quick, instant check of one key metric in a dashboard. It works well for KPIs such as system load, sales progress, or temperature. Here, the viewer only needs to judge whether the value sits in a good, caution, or danger range.
When you shouldn’t use a circular gauge
Skip a circular gauge when you need to display several metrics side by side or when readers must read exact values. Multiple gauges quickly fill a dashboard and make comparisons awkward. Moreover, curved scales are harder to read precisely. In these situations, a bullet chart is a better choice. This chart shows exact numbers and lets you compare several values against reference ranges in a compact space.
Sunburst chart
Alternative names: Multi-level pie chart, Multilayer pie chart, Sunburst graph, Ring chart, Radial treemap
What is a sunburst chart?
A sunburst chart has many names, but whatever you call it, it’s a spectacular type of graph. It shows a hierarchical dataset through a series of concentric outward rings. Each of those rings corresponds to a different hierarchy level. The inner circle looks like a donut chart, but each outer ring can be sliced up depending on its relationship to the inner (parent) circle.
When should you use a sunburst chart?
Sunburst charts are often a good alternative to treemaps, but if you do opt for this type of chart, keep in mind that its radial layout takes more space than a rectangular shape of a treemap.
When you shouldn’t use a sunburst chart
Avoid a sunburst chart for very deep hierarchies or when there are many branches. Outer slices become thin, labels are hard to place, and comparisons across branches are difficult. If space is limited or you need easier comparisons, a treemap will work better.
Funnel chart
Alternative name: Triangle chart
What is a funnel chart?
A funnel chart stacks several horizontal bars on top of one another, widest at the base and narrowing toward the tip. Each band represents a level in a hierarchy, such as awareness, interest, and purchase in a marketing funnel.
When should you use a funnel chart?
Use these charts when you need a quick visual of how numbers shrink or grow as they move through ordered stages. They are common in sales and marketing dashboards to show how many prospects reach each step and what revenue is tied to those steps. They also work anywhere a clear top‑to‑bottom (or entry‑to‑exit) flow helps tell the story. They are widely used in infographics and business presentations and dashboards.
When you shouldn’t use a funnel chart
Avoid these charts when you need precise comparisons between stages. Because the bars taper, they do not share a common baseline, and people find it hard to judge their exact sizes. A standard bar or column chart shows differences more clearly. Pyramid and funnel charts also become unreadable if you include too many stages because narrow slices are hard to read and label.
That’s it for part-to-whole and hierarchical charts. In the next part we’ll explore data-over-time charts. Take care and see you soon!
You can also check all other parts here:
Part 1: Comparison charts
Part 2: Correlation charts
Part 4: Data-over-time charts
Part 5: Distribution charts
Part 6: Geospatial and other charts
