Overview
A pie chart is a circular chart divided into slices, where each slice shows one category’s share of a single meaningful whole, usually expressed as a percentage. The deciding factor for using one is your data and task: pie charts work best for simple part-to-whole composition with a few categories, and another chart is usually clearer when you need precise comparison, ranking, or trends over time.
That single rule shapes everything else in this guide. Because a full circle represents 100% of one total, a pie chart only makes sense when your categories are non-negative parts that genuinely add up to that total. The moment your goal shifts from “roughly how is this whole divided?” to “exactly which category is bigger, and by how much?” or “how did this change across quarters?”, a bar chart or line chart typically communicates the answer faster.
This article walks through what a pie chart shows, how to calculate one from raw values, when it helps, when it misleads, and which alternatives fit specific analytical tasks — with a short worked example and a review checklist you can reuse before you publish.
What a pie chart shows
A pie chart answers one practical question: how is a single total split among a handful of categories? Each wedge, or slice, represents a category, and the size of that slice — its angle and area — encodes the category’s share of the whole. As the “From Data to Viz” project puts it, a pie chart is a circle divided into sectors that each represent a proportion of the whole, most often shown as percentages.
To read one, you compare slice sizes and match each to its label or a legend. A budget pie might show that rent takes the largest wedge, groceries a medium one, and savings a thin sliver. The chart communicates the composition of the total at a glance, which is why pie charts remain one of the most recognizable chart types in reports and presentations, as Infogram notes in its overview of pie chart pros and cons.
The catch is that this reading depends entirely on a meaningful total. If the categories do not describe a complete whole — or if they overlap, double-count, or leave gaps — the slices stop meaning what readers assume they mean. Everything below builds on getting that whole right first.
Slices, percentages, and the whole
Each slice represents a category’s share of the total, and all slices together are meant to describe one complete whole. If rent is 40% of a monthly budget, its slice fills 40% of the circle; the remaining slices fill the other 60%. The implicit promise of the format is that the categories are mutually exclusive and collectively exhaustive — nothing is counted twice, and nothing important is missing.
This is why proportional integrity matters so much. Eval Academy’s guide to common pie chart misuses points out that non-proportional data can exceed 100%, which breaks the entire premise of a pie. When “share of a whole” values sum to more than 100% — for example, survey responses where people could pick multiple options — a pie chart quietly misstates the data, because a circle cannot hold more than one full total.
Pie chart, pie graph, and circle graph
“Pie chart,” “pie graph,” and “circle graph” are generally used to mean the same thing: a circle divided into proportional slices. The terminology varies by region, textbook, and tool, but the underlying idea is identical, so you can treat these labels as interchangeable in most contexts.
Related circular charts are a different matter. A donut chart, polar area diagram, rose diagram, sunburst, or waffle chart may look pie-adjacent but often encode data differently — through area, radius, hierarchy, or a grid of units rather than a simple slice of one circle. It is worth keeping the plain pie chart distinct from these variants, which the final sections of this guide define individually.
How to calculate a pie chart from raw values
Turning raw numbers into a pie chart is a short, repeatable process: total the values, convert each value to a percentage, and convert each percentage to a slice angle in degrees. Working through it once removes the mystery and helps you catch bad data before it becomes a bad chart.
Consider a small, realistic dataset — the sources of traffic to a research dataset page over one month:
- Search: 620 visits
- Direct: 310 visits
- Social: 155 visits
- Referral: 93 visits
- Email: 62 visits
These five categories are non-negative, mutually exclusive, and describe one complete whole (all visits for the month), so a pie chart is a defensible choice. The three steps below turn these values into slices.
Step 1: Add the total
Start by summing every value to get the denominator, because a slice percentage means nothing until you know the total it is a share of. Here, 620 + 310 + 155 + 93 + 62 = 1,240 visits. That 1,240 is the whole circle — 100% — and every slice will be measured against it.
Confirm the total is genuinely meaningful before continuing. If the categories did not cover all traffic, or if some visits were counted in two channels at once, the denominator would be wrong and every downstream percentage would inherit that error. A valid total is the single most important precondition for a trustworthy pie chart.
Step 2: Convert each value to a percentage
The formula for each slice is straightforward: percentage = (category value ÷ total) × 100. Apply it to each row against the total of 1,240:
- Search: 620 ÷ 1,240 × 100 = 50%
- Direct: 310 ÷ 1,240 × 100 = 25%
- Social: 155 ÷ 1,240 × 100 ≈ 12.5%
- Referral: 93 ÷ 1,240 × 100 = 7.5%
- Email: 62 ÷ 1,240 × 100 = 5%
These percentages are the values most readers actually want, and here they sum cleanly to 100%. When you round to whole numbers, keep the raw figures nearby so you can verify the total, which matters for the rounding issue covered shortly.
Step 3: Convert each percentage to degrees
A full circle is 360 degrees, so each slice’s central angle equals its share of the whole multiplied by 360. You can work from the fraction directly — angle = (category value ÷ total) × 360 — or from the percentage: multiply the percentage (as a decimal) by 360. Search at 50% becomes 0.50 × 360 = 180 degrees, a clean half-circle; Direct at 25% becomes 90 degrees; Social at 12.5% becomes 45 degrees; Referral at 7.5% becomes 27 degrees; and Email at 5% becomes 18 degrees.
Add those angles and they total 360 degrees, which confirms the math is internally consistent. Most spreadsheet and charting tools compute these angles for you, but knowing the conversion lets you sanity-check a chart when a slice looks wrong for its label — a mismatch that, as the sections below explain, is a genuine reporting risk.
Why pie chart percentages sometimes total 99% or 101%
Rounded pie chart percentages sometimes add up to 99% or 101% instead of exactly 100%, and this is almost always a rounding artifact rather than a data error. When you round each slice to a whole number, small fractions get pushed up or down independently, and the rounded parts do not always reassemble into a perfect 100.
The fix is transparency, not false precision. You can show one decimal place where it matters, note that figures are rounded, or adjust the largest slice’s label to absorb the rounding difference so the visible numbers sum to 100. What you should avoid is silently forcing values or hiding the discrepancy — if percentages genuinely exceed 100% by a wide margin, that is a sign the data is non-proportional and does not belong in a pie chart at all, as Eval Academy warns about data that can exceed 100%.
When to use a pie chart
A pie chart is a good choice when you are showing simple, static composition and readers only need the approximate share of each part. The practical criteria are: the values form one meaningful whole, they are non-negative, there are only a few categories, the differences between slices are large enough to see, and you do not need precise ranking or a trend over time.
Pie charts are habitually read faster by non-technical audiences because the format is so familiar, which can make them pragmatically useful in a briefing even when they are not perceptually optimal. That familiarity is a real advantage — but it only pays off when the underlying data actually suits the format. When the composition is stable and the story is “this one total splits roughly like so,” a pie communicates it in a single glance.
The criteria below and the good use cases that follow help you confirm the fit before you commit to the format.
- A single meaningful whole (all parts sum to one total)
- Non-negative values only
- A small number of categories with visibly different sizes
- No need for exact comparison or ranking
- No time dimension or multi-group comparison
Good use cases
Pie charts fit well when categories are non-negative parts of one whole and you want to communicate share rather than exact value. Common examples include budget allocation (how spending divides across a fixed total), survey response share for a single-choice question, traffic source mix like the worked example above, market share among a few competitors, or the category composition of a research dataset.
The traffic-source pie works precisely because Search at 50% and Email at 5% are visibly different and the five channels cover all visits. A portfolio’s asset allocation is another strong fit: stocks, bonds, and cash sum to 100% of the portfolio, and readers usually care about the rough split, not third-decimal precision. In each case the whole is unambiguous and the number of slices stays small.
Two-slice and dominant-slice pie charts
A two-slice pie can be effective for genuinely binary compositions — compliant versus non-compliant, yes versus no — where readers only need to see which side dominates and by roughly how much. Similarly, a chart with one overwhelmingly large slice can make a single point clearly: one vendor holds most of the market, or one failure category accounts for most incidents.
The caveat is that these simple pies drift easily from analysis toward rhetoric. A near-100% slice emphasizes dominance emphatically, which is fine for advocacy or storytelling but can overstate a point in an analytical report. If exact values matter, even a two-slice split is often clearer as a labeled bar, so use the dominant-slice pie deliberately, not by default.
When not to use a pie chart
Avoid a pie chart when your task needs precision, ranking, trends, or comparison across groups — or when the data simply does not form one clean whole. The core weakness is perceptual: as the “From Data to Viz” project bluntly notes, humans are pretty bad at reading angles, which makes it hard to order slices or judge small differences accurately. FP&A practitioners make the same point, observing that it is hard to read numbers accurately from a pie chart because the brain struggles to estimate 2D areas.
This is not a claim that pie charts are always wrong. Ataccama’s practitioner critique estimates that in about 90% of cases you can use bar charts instead of pies, while still acknowledging pies are easier to judge when slices land on round 90-degree angles like 25%, 50%, and 75%. The honest framing is that pie charts have a narrow zone of good fit, and several common situations fall outside it.
The subsections below cover the failure modes worth checking before you publish.
Too many or too similar slices
Pie charts break down when they hold too many categories or when several slices are close in size. With many wedges, the chart becomes a cluttered ring of colors that forces readers into the legend, and Eval Academy identifies presenting too many slices as a leading misuse that quickly makes pies unreadable. A widely repeated rule of thumb is to avoid going much beyond a handful of slices, since beyond that viewers cannot reliably distinguish them.
Similar-sized slices are just as problematic, because judging which of two near-equal wedges is larger is exactly the angle-comparison task humans do poorly. When you hit either problem, the fix is usually to group small or minor categories into a single “Other” slice — while disclosing what it contains — or to switch to a bar chart, where the same categories line up against one axis and rank themselves.
Negative values, zero values, and missing data
Negative values cannot be shown in a pie chart, because a slice represents a part of a whole and a whole has no negative portions. If your data includes negatives — profit and loss, net change, over/under budget — use a bar chart or another format that has an axis capable of going below zero.
Zero values and missing data create quieter problems. A zero-value category produces an invisible slice that can silently vanish from the chart, so if that category matters, surface it with a label, annotation, or callout rather than letting it disappear. Missing or unknown values should never be folded silently into other slices or dropped without a note; disclose them, because hiding them distorts the denominator and makes the whole look more complete than it is.
Time trends and side-by-side comparisons
Pie charts are weak for anything involving change over time. Infogram notes that pie charts are ineffective for complex data, trends over time, or relationships between variables, and a line or stacked bar chart shows a trend far more directly than a sequence of pies ever could.
Placing pies side by side to compare segments or periods is similarly unreliable. Small composition changes between two circles are nearly impossible to judge by eye, so a shift from 22% to 26% can pass unnoticed. When you need to compare compositions across groups or time, a stacked or grouped bar chart aligns the categories on a shared axis and makes the differences visible — which leads directly into the decision matrix below.
Pie chart decision matrix: when to use a pie, bar, stacked bar, waffle, treemap, or line chart
Choosing a chart is really about choosing the analytical task you want to make easy. A pie chart optimizes for approximate share of a single whole; other chart types optimize for ranking, exact comparison, composition across groups, trends, or hierarchy. The matrix below maps common tasks to the chart that usually serves them best, along with when to avoid each.
Use it as a starting point rather than a strict rule — context, audience familiarity, and interactivity all shift the tradeoffs — but it captures the defaults most visualization guidance converges on.
| Chart type | Best for | When to avoid |
|---|---|---|
| Pie chart | Approximate share of one whole, few categories, static composition | Precise comparison, ranking, trends, negative values, many slices |
| Bar chart | Precise comparison and ranking of categories on one axis | Emphasizing that parts sum to a single whole |
| Stacked bar chart | Composition across several groups or a few time periods | Reading exact values of middle segments across many stacks |
| Waffle chart | Communicating an approximate percentage in an accessible grid | Precise values or many categories at once |
| Treemap / sunburst | Hierarchical or nested part-to-whole data with many items | Simple flat composition where a pie or bar is clearer |
| Line chart | Trends and change over time | Static one-moment composition of a whole |
| Dot plot | Precise comparison across many categories in compact space | Emphasizing part-to-whole relationships |
Best chart by analytical task
Reading the matrix as a workflow: reach for a pie chart only when the task is “show the rough split of one whole among a few parts.” For precise comparison or ranking, a bar chart or dot plot wins because both categories sit against a shared axis, and Ataccama argues bar charts are easier to read because they rely on a single axis rather than area.
For the remaining tasks, match the encoding to the question. Use a stacked bar chart to compare composition across groups or a few periods, a line chart for trends over time, a waffle chart when you want an accessible, easy-to-count approximation of a percentage, and a treemap or sunburst when the data is hierarchical or has too many items for a pie. The goal is not to ban pies but to reserve them for the one job they do reasonably well.
Best practices for clear pie charts
If a pie chart is the right call, a few design choices separate a clear one from a misleading one. The priorities are direct labels, sensible slice ordering, restrained color, no distorting effects, and an unambiguous denominator. These matter more than they might seem: an audit-oriented review in The CPA Journal found that among 1,584 charts across 162 documents, 192 (12%) had a fatal proportionality flaw, and that pie charts regularly showed slices that did not match their displayed values.
Order slices in a logical sequence — often largest to smallest, starting from the top (12 o’clock) position — so readers can scan them predictably. Keep the number of categories small, label each slice with its value or percentage, and make sure the chart’s title or caption states exactly what the whole represents. The subsections below expand the three choices that most often go wrong.
Label slices directly when possible
Label each slice directly on or beside the wedge rather than forcing readers to bounce between a legend and the chart. Direct labels let people read the value and category in one place, which speeds comprehension and makes it easy to verify that a slice’s size matches its stated number.
There is a caveat worth noting from the University of Melbourne’s Statistical Consulting Centre, which lists needing to directly label every slice or add a percentage to every slice as a sign the chart is failing. In other words, if the pie only works because every wedge carries a number, the geometry is not doing its job and a bar chart may serve you better. Direct labels help a genuinely suitable pie; they do not rescue an unsuitable one.
Use color to distinguish, not decorate
Color on a pie chart should separate categories clearly, not add decoration. Use distinct, easily distinguishable hues, and avoid relying on color alone to carry meaning — pair it with direct labels so the chart still works in grayscale or for readers who cannot distinguish certain colors.
Be especially careful with red and green together. FP&A guidance flags showing green and red colors on the same page as a common mistake, both because it can imply good/bad judgments you did not intend and because it is hard for many colorblind readers to tell apart. Choose a palette that stays distinguishable across common types of color vision, and treat color as a secondary cue backing up the labels.
Avoid 3D, exploded slices, and decorative effects
Skip 3D perspective, tilted angles, exploded slices, and decorative backgrounds — they distort the very proportions the chart exists to communicate. A 3D pie tilts the circle so that slices near the front look larger than their values and slices at the back look smaller, biasing perceived proportions. The CPA Journal’s review notes that some flawed charts encode values by a shape’s diameter rather than its area, which should encode value as area — a subtle distortion that inflates apparent differences.
Exploded pies, which pull one slice away from the center, draw attention to that slice and subtly shift how its size reads relative to the rest. Used sparingly for genuine emphasis they can be acceptable, but they trade analytical accuracy for rhetorical punch. When accuracy matters, keep the pie flat, whole, and free of ornament.
Common types and related charts
Several charts are relatives of the pie, and knowing how each differs helps you pick the right one and read others correctly. Some, like the donut, are close cousins that share the pie’s strengths and limits; others, like the sunburst or waffle, encode data in a fundamentally different way. The subsections below define the ones you are most likely to meet.
The short version: donut, 3D, and exploded charts are pie variants, while polar area diagrams, rose diagrams, sunbursts, and waffle charts are related but not interchangeable. Treat the label on a chart as a hint about how to read it, not a guarantee that it behaves like a standard pie.
Donut chart
A donut chart is a pie chart with the center removed, leaving a ring of proportional slices around a hole. The hollow center can hold a total or a label, and some designers find the ring visually cleaner than a full circle, which is why donuts are common in dashboards.
Functionally, though, a donut shares almost all of a pie’s readability limits. Removing the center arguably makes angle judgments slightly harder, since the vertex where slices meet is gone. The same rules apply: keep categories few, label directly, and switch to a bar chart when you need precise comparison.
3D pie chart and exploded pie chart
A 3D pie chart renders the circle as a tilted, extruded disc, and an exploded pie chart separates one or more slices from the center. Both are variants of the standard pie, and both are usually risky for accurate interpretation.
The 3D version distorts proportions because perspective enlarges foreground slices and shrinks background ones, so the picture no longer matches the numbers. The exploded version biases attention and perceived size toward the pulled-out slice. Neither adds analytical information over a flat pie, so reserve them for cases where a single, deliberate emphasis outweighs precision — and prefer a plain pie otherwise.
Polar area diagram, rose diagram, sunburst chart, and waffle chart
These charts look pie-adjacent but encode data differently, so they are not drop-in substitutes. In a polar area diagram (and the related rose diagram, famously used by Florence Nightingale), slices often share equal angles while their radius or area carries the value — the opposite of a standard pie, where the angle carries the value.
A sunburst chart stacks concentric rings to show hierarchy, with inner rings as parent categories and outer rings as their children, which lets it handle nested part-to-whole data a flat pie cannot. A waffle chart abandons the circle entirely, filling a grid of small squares (often 100 of them) to represent percentages, which many readers find easier to count precisely than angles. Because each of these encodes proportion through a different visual channel, always confirm what the geometry means before drawing conclusions.
Accessibility and accuracy checks before publishing
Before a chart goes into a report, dashboard, or shared dataset page, it is worth a quick review for both accuracy and accessibility — especially because charts in published documents are wrong more often than people assume. The CPA Journal reports that in one sample of 50 public companies’ yearly reports, 86% had at least one broken chart, and cites a UK study finding 49% of key financial graphs distorted by at least 10%. Those numbers are a reminder that a plausible-looking chart is not automatically a correct one.
Accessibility deserves the same attention as accuracy. That means direct labels rather than color-only meaning, a palette that stays distinguishable for colorblind readers, and a short text summary of the chart’s main point so screen-reader users and skimmers both get the takeaway. If you publish datasets as interactive pages, this is where tooling helps: a platform like TablePage lets you turn a spreadsheet into a shareable page where readers get charts, AI insights, and a filterable table with no signup needed, so anyone can inspect the underlying values behind a visual rather than trusting the picture alone.
A short review checklist
Run through this checklist before publishing any pie chart or the report that contains it:
- Confirm the denominator: state exactly what the whole represents and that all parts sum to it.
- Verify values and labels: check that each slice’s size matches its stated number and that labels are correct.
- Avoid distortion: no 3D, tilt, exploded slices, or decorative effects that skew proportions.
- Make colors distinguishable: use a colorblind-safe palette and never rely on color alone.
- Include a text explanation: give the chart a clear title and a one-line summary of its main point.
- Disclose grouping and gaps: note what any “Other” slice contains and flag missing or excluded data.
Because tools like TablePage keep the filterable source table beside the chart, they make several of these checks easier — readers can confirm a slice against the raw rows, which is exactly the verification the accuracy findings above call for.
Frequently asked questions
The questions below cover the edge cases and comparisons readers most often raise once they have the basics.
Can a pie chart include negative values?
No. A pie chart shows parts of a whole, and a whole cannot contain negative portions, so a negative value has no valid slice to occupy. If your data includes negatives — such as losses, net changes, or amounts under budget — use a bar chart or another format with an axis that extends below zero, where negative bars can point downward or left and still be read accurately.
Is a pie chart better than a bar chart?
It depends on the task. A pie chart can be the better choice when you want to communicate the approximate share each part takes of a single whole and there are only a few categories. For precise comparison, ranking, or reading exact values, a bar chart is usually clearer, because it places categories on a shared axis and Ataccama estimates you could use a bar chart instead of a pie in roughly 90% of cases. Match the chart to whether your reader needs “rough split” or “exact comparison.”
How many slices should a pie chart have?
Fewer is better, and there is no universal hard limit — but readability drops quickly as slices multiply, since comparing many similar wedges is exactly the angle judgment humans do poorly. As a practical guideline, keep the count small enough that every slice is visibly distinct and directly labelable; a common rule of thumb is to stay within a handful of categories. When you have more, group minor categories into a clearly disclosed “Other” slice or switch to a bar chart, where a longer list of categories still ranks cleanly against one axis.