AI Glossary / Data science and analytics / Beginner
Pie chart
Definition
A pie chart is a circular chart divided into slices that show how categories contribute to one whole. Each slice's angle, and therefore its area in a flat circular chart, is proportional to its category's share of the total.
What it means
A pie chart answers a part-to-whole question, such as how a budget is divided among a few departments or how completed requests are distributed across exclusive outcome categories. The complete circle represents the defined total.
Its meaning depends on that denominator. A chart of selected products is not automatically a chart of all sales, and a chart of respondents who answered a question is not automatically a chart of everyone invited to the survey.
Pie charts can make a large share, such as roughly half or three quarters, easy to recognize. They are less effective for precise comparisons between similar slices or for displaying many categories. Choosing the chart should follow the question and data, rather than a preference for a colorful circle.
How it works
Define a meaningful whole and categories that divide it without overlap. Add the category values, divide each value by the total and multiply by 100 to obtain its percentage. Multiplying the proportion by 360 gives the slice angle in degrees.
Use nonnegative values with a positive total. Negative contributions or overlapping responses do not fit the ordinary slice-of-one-whole interpretation. If the data omit categories, either include a clearly defined remainder or label the restricted total explicitly.
A flat two-dimensional presentation preserves the intended geometry. Perspective and three-dimensional effects can distort perceived proportions. Direct labels with category names and values reduce the need to match colors to a separate legend.
Check rounding and provide an accessible textual description or data table. Colors should support identification without being the only way readers can understand the categories. For precise comparisons, a bar chart may be easier to read.
Three practical examples
- A project manager shows how a fixed event budget is divided among venue, catering and other costs, using the complete budget as the denominator.
- A support team displays the shares of closed cases assigned to mutually exclusive final outcomes.
- A website editor shows the device distribution of recorded visits, stating the period, measurement source and handling of unknown devices.
Think of it this way
Think of cutting a single cake into portions. A slice's size describes its share of that cake, and all portions together make the whole.
The analogy has limits. The underlying categories may represent counts or money rather than physical portions, and two equal-sized chart circles can represent very different totals. Slice shares alone do not reveal the absolute quantity behind each chart.
Worked example
Suppose a fictional website records 200 visits in a stated period: 120 from mobile devices, 60 from desktops and 20 from tablets. Assume every visit belongs to exactly one of these categories.
Mobile accounts for 120/200 = 60%, so its angle is 0.60 × 360 = 216 degrees. Desktop accounts for 30%, or 108 degrees. Tablet accounts for 10%, or 36 degrees. The shares total 100% and the angles total 360 degrees.
Now suppose a different report records 80 people using a chatbot and 70 using a search engine among 100 respondents, with multiple answers allowed. These figures total 150 because people can appear in both groups. Turning them into two slices of “respondents” would misrepresent the data.
A bar chart showing 80% and 70%, labeled as multiple-response percentages, would preserve that meaning. Alternatively, a pie would require mutually exclusive combinations such as chatbot only, search only, both and neither.
Illustrative case study
Illustrative fictional case: Harbour Publishing asks an AI assistant to visualize why readers visit its site. The survey permits several reasons per person. The assistant produces a pie chart and rescales the selected answers to sum to 100%.
An editor notices that the chart now describes the share of recorded selections, not the share of readers choosing each reason. The team replaces it with labeled bars and states that respondents could select multiple reasons.
If the team later chooses to analyze selections as a whole, it will name that denominator explicitly. The fictional lesson is that a technically drawable chart can still answer the wrong question. No measured audience result is claimed.
Common misconceptions and mistakes
Do not use a pie merely because values are expressed as percentages. Percentages from different denominators or overlapping categories do not necessarily form one whole.
Avoid three-dimensional effects and unexplained “other” slices. Do not compare equal-looking pies as if their totals were equal. A slice growing from 20% to 30% can still represent fewer units if the overall total falls. Finally, rounding differences should be explained rather than silently altering values to force a cosmetically perfect total.
Business use
Use a pie when a small number of categories make up a clearly defined whole and the main message concerns shares. Include the total, period and source so the chart can stand on its own.
When the business question concerns small differences, rankings or changes over time, consider a chart better suited to those comparisons. Review AI-generated charts for denominator errors, overlap and invented categories before using them in a presentation. A brief sentence stating the intended takeaway is a useful check that the visual communicates a real analytical point.
People also ask
Can a pie chart show counts rather than percentages?
Yes. Counts can determine slice sizes when they divide one meaningful total. Labels may show counts, percentages or both.
Must the displayed rounded percentages add to exactly 100%?
Not always. Rounding can produce a small discrepancy, which should be noted when relevant; the underlying unrounded shares should still describe the whole.
Is a doughnut chart different in meaning?
It uses the same part-to-whole idea with a central opening. The change in appearance does not remove the need for a valid denominator and clear categories.
Check your understanding
Three survey options receive 60%, 50% and 30% because respondents can select several options. Should these become three slices labeled as shares of respondents?
Answer: No. The categories overlap and do not partition the respondents. Use a presentation that preserves the multiple-response meaning or define mutually exclusive categories.
Related terms
- Data visualization
- Proportion
- Bar chart
- Categorical variable
- denominator
- doughnut chart
- percentage
- accessibility
Learn next
Read proportion and bar chart. For one proposed pie, write a sentence defining the whole and verify that every record belongs to exactly one category. Compare a pie and a bar version, choosing the one that best supports the intended comparison.
More in Data science and analytics
- Persistence module - A persistence module is an indexed family of algebraic objects together with maps that describe how information passes from one index to later indices.
- Persistent Betti number - A persistent Betti number counts the independent homology classes present at an earlier filtration parameter that remain nonzero and independent at a later...
- Persistent homology group - A persistent homology group is the image of the homology map from an earlier stage of a filtration to a later stage.
- Pivot table - A pivot table is a summary table that groups source records by selected categories and calculates measures such as sums, counts or averages.
- Post hoc analysis - Post hoc analysis is analysis undertaken after results have been examined, often to explore an unexpected pattern or investigate a broader finding in more...
- Predictive analytics - Predictive analytics uses data and statistical or machine-learning methods to estimate future or otherwise unknown outcomes.
Sources
Office for National Statistics, appropriate use and limitations of pie charts: https://digitalblog.ons.gov.uk/2017/01/26/the-humble-pie-chart/
Datawrapper, pie-chart design considerations: https://www.datawrapper.de/academy/what-to-consider-when-creating-a-pie-chart
Office for National Statistics, accessible colour use in charts: https://service-manual.ons.gov.uk/data-visualisation/colours/using-colours-in-charts
The visit counts and publishing case are fictional teaching examples.
Previous: PID controller Next: Pilot job