Choosing Between Bars, Pies, and Tables

Choosing a Display for the Question

A categorical data set can often be shown in more than one reasonable way. A bar chart, pie chart, and table may all summarize the same responses, but they make different tasks easier. A bar chart supports comparisons between categories; a pie chart emphasizes how categories make up a whole; and a table gives readers direct access to recorded values. The best choice depends less on which display is most familiar and more on what readers are being asked to see.

Before choosing, identify the data structure and the purpose of the display. Is there one categorical variable or are categories being compared across groups? Are there only a few categories, or so many that labels and slices become difficult to follow? Should a reader see broad differences, or retrieve an exact count or percentage? These questions lead to a practical choice rather than a rule that one display is always best.

Definition: A bar chart displays categories as separate bars whose lengths represent counts or proportions. A pie chart divides a circle into sectors whose sizes represent categories’ shares of a whole. A table arranges category labels and their counts, proportions, or percentages in rows and columns for direct lookup.

What Each Display Makes Easy

A bar chart is usually a strong choice when the main task is comparing category sizes. Readers can compare bar lengths along a common scale, and category labels can be listed clearly. It works with counts when those counts are the relevant comparison, or with proportions when samples of different sizes need to be compared. With many categories, a bar chart can still be useful if the labels remain readable and the ordering helps the reader find categories.

A pie chart is most useful when the question is about the composition of one whole: what shares of the total belong to each category? It is generally clearest with a small number of categories, particularly when the purpose is to show the broad division of the whole rather than to compare close values. Sectors must represent parts of a single total, so the categories should be mutually exclusive and collectively account for that total.

A table is preferable when the reader needs exact values, when category names are long, or when there are many categories to list. A table can also present multiple variables or groups without forcing a reader to estimate values from a visual scale. Its trade-off is that patterns and relative sizes may be less immediately visible than in a chart. A chart and a table can be used together when readers need both a quick visual summary and precise values.

Reader’s main task	Often suitable	Reason
Compare category sizes or rank categories	Bar chart	Bar lengths can be compared against a common scale.
See how a few categories divide one total	Pie chart	Each sector shows a category’s share of the whole.
Look up exact counts or percentages	Table	Values are printed directly rather than estimated.
Compare patterns across groups and retain details	Table, sometimes with a bar chart	A table preserves group-specific values; a chart can emphasize the pattern.

The number of categories matters, but it is not a fixed numerical cutoff. A pie chart with many similarly sized sectors can be hard to compare, while a bar chart with many categories can become crowded if labels are long. If category-by-category lookup matters, a table may be clearer. If the audience needs to notice the largest or smallest categories quickly, a bar chart may remain effective, especially when categories are ordered in a meaningful way.

Key principle: Choose the display that makes the intended comparison easiest without hiding information the reader needs. The question, number of categories, and need for exact values should be considered together.

Why Bar Charts Help With Comparisons

In a bar chart, each category’s value is represented by a length measured on the same scale. Comparing two values therefore amounts to comparing two lengths with a shared reference. A pie chart instead represents each value as a sector of a circle. Its angles are proportional to category shares, but estimating and comparing angles is less direct than comparing lengths, particularly for categories of similar size. This is why a bar chart is generally preferable when the central task is deciding which category is larger or by how much.

The relationship between a category’s proportion and its angle also clarifies the purpose of a pie chart. If a category represents proportion \(p\) of a whole circle, its angle is \(360p\) degrees. Because the proportions of all categories in the whole sum to 1, their sector angles sum to a complete circle:

\sum_{i=1}^{k} 360p_i =360\left(\sum_{i=1}^{k}p_i\right) =360(1) =360^\circ.

Thus, a pie chart is not just a collection of unrelated category sizes: its sectors jointly represent the whole. That property makes it suitable for composition, but does not make it the best choice for every comparison.

Worked Example: Showing One Survey Question

Suppose a hypothetical campus survey asks 80 students which of four library study areas they use most often. The counts are: quiet floor, 32; group rooms, 24; café area, 16; and outdoor tables, 8. Choose a display for a short report that should show the overall pattern and answer which area was most common.

First check the total: \(32+24+16+8=80\), so each response is represented in one of the four categories. The category proportions are:

\hat{p}_{\text{quiet}}=\frac{32}{80}=0.40,\quad \hat{p}_{\text{group}}=\frac{24}{80}=0.30,\quad \hat{p}_{\text{café}}=\frac{16}{80}=0.20,\quad \hat{p}_{\text{outdoor}}=\frac{8}{80}=0.10.

These proportions sum to \(0.40+0.30+0.20+0.10=1.00\), or 100%. A pie chart is a reasonable choice if the report’s emphasis is how the 80 responses divide among the four areas. A bar chart is the stronger choice for the specific comparison of which area was most common: the quiet floor has 32 responses, compared with 24 for group rooms, 16 for the café area, and 8 for outdoor tables. A table would be useful if readers also need to retrieve each exact count. For the stated goal of showing the pattern and identifying the largest category, choose a bar chart; a small table can accompany it if exact counts matter.

When Exact Values or Group Comparisons Matter

A table often becomes more valuable as the number of categories or dimensions increases. With two categorical variables, a two-way table can show counts for combinations of categories, such as response by study year. If the comparison is about how groups differ in their response distributions, a bar chart may help make the pattern visible; if the audience needs exact cell values, the table is the direct source. When comparing groups of unequal sizes, use proportions that match the comparison rather than letting larger groups appear different simply because they contain more observations.

Worked Example: Many Categories and Exact Counts

A hypothetical help desk records the type of issue in 60 support requests: password access, 18; software installation, 14; account setup, 10; printing, 7; network access, 5; email, 4; equipment, 2. Suppose the report must help staff find the exact number of requests in each category, and several issue types have similar counts.

The counts total \(18+14+10+7+5+4+2=60\). For instance, the proportion of requests about printing is:

\hat{p}_{\text{printing}}=\frac{7}{60}=0.1167\approx 11.7\%.

A table is the best primary display because exact counts are the stated need and it can list all seven categories without requiring readers to estimate bar lengths. A bar chart would be a useful addition if staff also want to spot the most common issue quickly; the categories could be ordered from largest to smallest. A pie chart is less suitable here: seven sectors would make comparisons among the smaller categories difficult, and the exact counts would still need to be printed or looked up. The choice follows the task, not a claim that seven categories always rules out a pie chart.

For multiple groups, distinguish the question “How many observations are in each category?” from “What proportion of each group is in each category?” The first may call for counts; the second calls for within-group proportions. A chart of counts can make a larger group look larger across all categories simply because more people were observed. A table can report both group totals and within-group percentages, while a bar chart can emphasize the percentage pattern. Exact values and visual comparison can therefore call for complementary displays.

Worked Example: Comparing Response Patterns Across Groups

In a hypothetical survey, 40 first-year students and 60 final-year students each choose one preferred study location: library, home, or another place. First-year counts are 20, 12, and 8; final-year counts are 24, 24, and 12, respectively. The question is whether the response pattern differs by year, rather than which group supplied more responses.

The totals are \(20+12+8=40\) and \(24+24+12=60\). Calculate proportions within each year so the group sizes are accounted for:

\begin{aligned} \text{First-year: }& \frac{20}{40}=0.50,\quad \frac{12}{40}=0.30,\quad \frac{8}{40}=0.20;\\ \text{Final-year: }& \frac{24}{60}=0.40,\quad \frac{24}{60}=0.40,\quad \frac{12}{60}=0.20. \end{aligned}

The proportions within each group sum to 1.00. For a quick visual comparison of response patterns, a 100% stacked bar chart is suitable because each group is represented as a whole and its segments show within-group proportions. A two-way table with the counts and percentages is better if readers need exact values. A single pie chart is not appropriate for comparing these two groups: each pie would show one group separately, making direct comparisons between corresponding categories less convenient. For this question, use a 100% stacked bar chart, with a table alongside it if exact counts are important.

A Practical Selection Process

Use the following process to make a defensible choice. It does not impose a rigid threshold for how many categories a display can contain; it prompts you to match the display to the actual reading task.

Identify the structure.
Decide whether the display concerns one categorical variable or a comparison involving groups or a second categorical variable.

Name the reader’s task.
Is the reader comparing category sizes, seeing parts of one whole, comparing group patterns, or looking up values?

Check category count and labels.
Consider whether a pie chart would have too many sectors to compare or a bar chart too many labels to read. Long labels and many categories often favor a table.

Choose the measurement to display.
Use counts when totals are the relevant quantities; use proportions or percentages when comparing shares, especially for groups of different sizes.

Decide whether one display is enough.
If readers need both a clear visual pattern and precise values, pair a chart with a table rather than expecting one display to serve both purposes equally well.

Common Mistakes

Choosing a pie chart just because values are percentages. Percentages can be shown in bars or tables too. A pie chart is appropriate when the categories represent shares of one whole and composition is the focus.
Using a table when the main task is visual comparison. A table provides exact values, but readers may have to scan and compare entries individually. A bar chart can make relative sizes and rankings easier to notice.
Comparing raw counts from groups of different sizes. The wrong reasoning is that a group with larger counts must have a stronger preference for a category. The correction is to compare within-group proportions when the question concerns response patterns, and to state that these are proportions of each group.
Assuming there is a universal maximum number of categories for each display. A fixed cutoff ignores label length, category sizes, and the task. Judge whether the resulting display lets readers make the intended comparison clearly.

Key takeaway: Use bars primarily for comparing category values, pies for showing a few categories as parts of one whole, and tables for exact lookup and detailed category listings. When the task requires both pattern and precision, a chart and a table can work together.

Check Your Understanding

For each situation, decide which display or combination of displays would best support the stated task, and explain why.

A poll records the preferred season of 120 people. The report’s main purpose is to show how responses divide across four seasons. Which display is a reasonable choice?
A store has 15 product categories and needs a report listing each category’s exact sales count. Which display should be primary?
Two departments have different numbers of employees. A survey asks employees to choose one of three benefits, and the goal is to compare preference patterns. Should the comparison use counts or within-department proportions?
A chart must make it easy to identify the largest of six categories whose values are close. Which display is generally more suitable: a bar chart or a pie chart? Explain.
Give one reason to present a table alongside a chart rather than relying on the chart alone.