This entry is part of the Chart Index, the reference library for the Chart Design Field Guide.

The lollipop chart is a bar chart on a diet. The bar is replaced with a thin line; the value is marked by a small circle at the end. Same data, same encoding (length and position), but a fraction of the ink. For visualisations that would otherwise be a wall of bars, the lollipop offers density without the visual heaviness.

It is a perfect example of Tufte's data-ink principle in action: remove ink that does not represent the data, and the data remains intact. The lollipop says here is the value with the smallest possible mark — the dot — and uses the stick only to anchor the dot to the category axis.

What it is

A lollipop chart maps a categorical variable to position along one axis and a quantitative variable to a position along the other. Each data point is rendered as a circle (the lollipop head) at the value position, joined to the baseline by a thin line (the stick). It is structurally identical to a bar chart with the bar collapsed to a hairline.

The hero chart above ranks fourteen Sydney suburbs by the median rent for a two-bedroom unit. Read the heads, not the sticks: they form a descending curve that flattens through the middle of the list, and the eye picks out the two breaks — between Chatswood and Parramatta, and again between Parramatta and Liverpool — where the rent steps down by more than the usual increment. Fourteen filled bars would have shown the same numbers as a solid wedge of blue; here the page stays light and the shape of the ranking is visible. Randomise the values and the heads slide along their sticks, and the sticks follow.

When to use it

Lollipops are the right choice when:

  • You have many categories to display (10–30 typically) and a regular bar chart would be visually heavy.
  • The reader's question is "which is biggest?" or "rank these" — same as a bar chart.
  • The values are close together and a thick bar would obscure small differences.
  • You want to reduce visual weight without sacrificing the length encoding.
  • The chart is dense or editorial — lollipops feel lighter and more analytical than bars.

When not to use it

  • Few categories. Three or four lollipops on a page look thin and decorative. A standard bar chart is more substantial and reads better.
  • Stacked or grouped composition. Lollipops do not stack cleanly. For composition, use stacked bars or dot plots with multiple series.
  • Very small differences. While lollipops show small differences better than thick bars, when differences are within the size of the dot, the chart loses precision. Use a dot plot or magnify the relevant region.
  • Time series. Lollipops imply discrete categories. For continuous time, use a line chart with markers.

Design principles

Use it where bars would be a wall

The saving is easiest to see side by side. The two figures below hold the same fourteen suburbs.

Fourteen suburbs as bars
Every value is a solid block, and fourteen of them together read as one slab of blue that happens to have a ragged right edge. The ranking is there, but the eye has to work along that edge to find it, and the page is heavy.
The same fourteen suburbs as lollipops
Identical encoding, a small fraction of the ink. The heads trace the ranking as a curve rather than an edge, the gaps between neighbouring suburbs are easier to judge, and there is room to breathe between rows.

Make the stick thin and quiet

The line should be 1–2 pixels. Any thicker and the chart becomes a thin bar chart. The data-ink saving is the whole point. The line is a visual guide; the eye should read past it.

Make the dot deliberate

The dot is the data. Use an 8–12 pixel circle with enough weight to stand apart from the stick — a solid fill, or a ring with a heavy stroke as in this entry — so that it is unambiguously the centre of attention on each row. A faint outline that fades into the line loses the encoding.

Sort by value

Lollipops live or die by sorting. Random or alphabetical ordering produces a forest of sticks with no rhythm. Sorted by value, the chart becomes a ranking — the eye reads from longest to shortest in one sweep.

Alphabetical order
The same rents listed A to Z. Each head has to be found on its own; the reader can look up a suburb but cannot see the shape of the market, and the second-highest rent is halfway down the list.
Sorted by value
The same ten suburbs ranked. The heads now form a single descending curve, and the two-tier structure of the list — five suburbs above 900 dollars, five well below — is visible before a single label is read.

Direct-label the dot

Place the value to the right of the dot. The reader's eye lands on the dot and reads the number without leaving. Few's principle of eliminating visual lookup applies — the chart should not need an axis if every value is labelled.

Values printed beside the heads
Each head carries its number. The reader gets the rank from position and the exact rent from the label without a trip to the axis, which could now be dropped altogether.

Highlight the lead

If the chart has a hero — the largest value, the one of interest — use the accent colour for its dot and let the rest fade to neutral. The hero becomes the focal point; the others are reference.

One suburb in colour, the rest in grey
Parramatta is the suburb the story is about, so it alone takes the house blue. The nine grey lollipops still give its rent a context — a little above the outer suburbs, well below the harbourside ones — without competing for attention.

Stagger labels if necessary

When categories have long names, the label column on the left can dominate. Wrap labels to two lines, abbreviate, or use a horizontal layout where the labels sit above the dots. The category column should not visually weigh more than the data column.

Avoid pairing dots and bars

A lollipop is one form; a bar is another. Mixing them on the same axes (some rows as bars, some as lollipops) breaks the visual contract. Pick one.

Anatomy

The composition of a lollipop chart
An anatomical guide

A lollipop chart has three structural elements: the dots, the sticks, and the categorical axis. The data-ink calculation favours the form whenever the bar version would feel visually heavy.

  • Bar chart — the parent form. Lollipops are bars with the rectangle collapsed.
  • Dot plot — drops the stick entirely. Even less ink; relies on a clear category axis.
  • Cleveland dot plot — two dots per category, joined by a line, encoding two values for comparison.
  • Range plot / dumbbell — two dots and a line, encoding from–to rather than a single value.
  • Forest plot — dot with horizontal error bars, used in meta-analysis.

Reading list

  • Tufte, E. (2001). The Visual Display of Quantitative Information. The data-ink principle.
  • Cleveland, W. (1993). Visualizing Data. The dot plot and its perceptual case.
  • Few, S. (2012). Show Me the Numbers. On reducing visual weight in dense displays.