StatisticsSciPy 1.18 · pandas 3.0 · statsmodels 0.15 · Python 3.12 or 3.13
Dashboard
0%
1
Curious builder0 XP earned · 300 to level 2
0 daysFinish a lesson to begin
Badge collection0 of 6 unlocked
57 small wins to finish your pathNext lesson →

Bar charts and pie charts

A bar chart is a chart of categorical data that draws one separate bar per category, with the bar's height equal to the category's frequency; a pie chart shows the same counts as slices of a circle sized by their share.

Last updated: 07 Oct, 2026 · SciPy 1.18

The Frequency distribution lesson counted nine flowers: Rose 3, Lily 4, Sunflower 2. A table of counts is exact but slow to read. The same numbers drawn as bars, or as slices, show at a glance which category is the largest.

Bar graphs from the frequency table · from the Complete Statistics for Data Science in 6 Hours video · 36:51 to 38:09

Drawing a bar chart from the flower table

The video draws the chart from the frequency table. The x axis lists the categories, Rose, Lily and Sunflower; the y axis is the frequency, 1, 2, 3 and up. Each category gets one bar as tall as its count: Rose to 3 in red, Lily to 4 in blue, Sunflower to 2 in green. It is still descriptive statistics: a summary of the data in hand.

Two things make it a bar chart and not a histogram. The bars stand apart, with gaps, because the categories are separate. And the order along the axis carries no meaning for nominal data like flowers, so the bars are often sorted from tallest to shortest; for ordinal data, such as T-shirt sizes, keep the natural order. The y axis starts at zero, because the reader compares bar lengths.

A bar chart also suits a discrete variable with a few values, such as the number of people at a table: one bar per value. A numeric variable with many values needs a histogram.

Drawing a pie chart

A pie chart draws the relative frequencies as slices of one circle. Each slice's angle is its share of 360 degrees, so the slices always fill the circle exactly:

Slice angles for the flowers

A pie chart answers one question well: what share of the whole is each part? It fits a few categories that make up one total. With many categories, or close values, the slices are hard to compare, and a bar chart is the better choice.

Plotting the flowers in matplotlib

The frequency table as lists

python
import matplotlib.pyplot as plt

flowers = ["Rose", "Lily", "Sunflower"]
counts = [3, 4, 2]                            # the frequency table
colors = ["tab:red", "tab:blue", "tab:green"] # the board's pen colours

Counting raw values with seaborn

When the data is the raw list of flowers rather than the counts, seaborn counts and draws in one call:

python
import seaborn as sns

sns.countplot(x=["Rose", "Lily", "Sunflower", "Rose", "Lily",
                 "Sunflower", "Rose", "Lily", "Lily"])

Drawing the bar chart and the pie chart

ExampleFrom the video, run on matplotlib 3.11.2
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 4))
ax1.bar(flowers, counts, color=colors)
ax1.set_title("Bar chart of the flowers")
ax1.set_xlabel("Flower")
ax1.set_ylabel("Frequency")
_, _, pct = ax2.pie(counts, labels=flowers, colors=colors, autopct="%1.1f%%", startangle=90)
for t in pct:
    t.set_color("white")                      # readable percentages on the slices
ax2.set_title("Pie chart of the flowers")
plt.show()
print("slice angles:", [360 * c / sum(counts) for c in counts])
A bar chart with Rose at 3 in red, Lily at 4 in blue and Sunflower at 2 in green, next to a pie chart of the same flowers with slices of 33.3%, 44.4% and 22.2%.

Reading the two charts

  • The bars stand at 3, 4 and 2, the frequencies, with Lily the tallest.
  • The slices are 33.3%, 44.4% and 22.2%, the relative frequencies 3/9, 4/9 and 2/9.
  • The angles are 120°, 160° and 80°, which add up to 360°.
  • Both charts show the same table: the bar chart makes the counts easy to compare, the pie chart makes the share of the whole easy to see.

Bar chart vs pie chart

Bar chartPie chart
ShowsCounts (or shares) per categoryShares of one whole
Compares byBar length, which eyes judge wellAngle and area, which eyes judge less well
Number of categoriesManyA few, about five or fewer
NeedsA y axis starting at zeroParts that add up to one total
In matplotlibplt.bar(categories, counts)plt.pie(counts, labels=...)

Where you use bar and pie charts

  • Survey results: answers per option as bars, sorted by count.
  • Exploring categorical columns: sns.countplot on each column of a dataset, such as the day of the week in tips.
  • Market share or a budget split: a pie chart of a few parts that make up 100%.
Watch out. A bar chart whose y axis starts above zero exaggerates differences: with an axis from 1.5 to 4, Lily's bar looks five times Sunflower's instead of twice. Keep the axis at zero (matplotlib's default for bar), and only use a pie chart when the parts add up to one whole.
Try it yourself
  • Draw the notes' colours: flowers = ["Green", "Red", "Yellow"] and counts = [3, 3, 2]. What are the slice angles now?
  • Swap ax1.bar for ax1.barh to draw horizontal bars, which suit long category names.
  • Add ax1.set_ylim(1.5, 4.5) and look at how much bigger Lily's bar seems.
PreviousFrequency distribution

Slow is fine. Stopping is the only problem.