Making Charts Readable

A chart is readable when the viewer can answer the intended question without guessing what the marks mean.

Readable does not mean fancy. It means clear.

The snippets in this chapter polish the familiar subject-total bar chart:

import pandas as pd
import matplotlib.pyplot as plt

records = [
    {"subject": "math", "minutes": 45, "date": "2026-05-10"},
    {"subject": "history", "minutes": 30, "date": "2026-05-11"},
    {"subject": "programming", "minutes": 60, "date": "2026-05-11"},
    {"subject": "math", "minutes": 25, "date": "2026-05-12"},
    {"subject": "programming", "minutes": 90, "date": "2026-05-13"},
    {"subject": "writing", "minutes": 45, "date": "2026-05-14"},
]

df = pd.DataFrame(records)
subject_totals = df.groupby("subject")["minutes"].sum().sort_values(ascending=False)

The basics are:

fig, ax = plt.subplots(figsize=(7, 4))
ax.bar(subject_totals.index, subject_totals.values)
ax.set_title("Total study minutes by subject")
ax.set_xlabel("Subject")
ax.set_ylabel("Minutes")
ax.tick_params(axis="x", rotation=30)
fig.tight_layout()

The figsize argument controls the figure size in inches. The fig.tight_layout() call asks matplotlib to adjust spacing so labels are less likely to be cut off.

Tick Labels

If category labels are long, rotate them:

ax.tick_params(axis="x", rotation=30)

Use rotation carefully. A small rotation can help. A chart where every label is vertical can become hard to read. If labels are too long, consider whether shorter display names would be clearer.

Figure Size

A chart with three subjects does not need a huge figure. A chart with ten subjects may need more width.

fig, ax = plt.subplots(figsize=(8, 4))

Figure size should serve the data. If the chart feels cramped, increase the size a little. If it feels empty, reduce it.

Color

Color can help, but it can also distract.

For a simple subject-total bar chart, one color is often enough:

ax.bar(subject_totals.index, subject_totals.values, color="steelblue")

Using a different bright color for every bar may look lively, but it usually does not add meaning. Reserve color differences for actual differences in the data, such as highlighting one selected subject or separating two meaningful groups.

Also remember that not everyone sees color the same way. A chart should not depend only on red versus green to communicate an important difference. Labels, position, and shape matter too.

Legends

Use a legend when there is more than one plotted series and the reader needs to know which is which:

ax.legend()

Do not add a legend when the labels are already obvious. A single blue line labeled by the title and axes usually does not need a legend that says “minutes”.

Avoiding Clutter

Clutter is anything that competes with the data without helping the reader understand it.

Common clutter:

• too many categories in one chart,
• labels that overlap,
• colors with no meaning,
• legends that repeat the title,
• titles that are vague,
• and chart types that add decoration without adding clarity.

Here is the same subject-total data with several of these problems at once — a vague title, no axis labels, colors that mean nothing, a redundant legend, unsorted bars, and tick labels squeezed until they are cut off:

Every number in it is correct. Compare it with the chart at the top of this chapter, which shows the same data and answers the same question without making the reader work.

Write the comparison down while it is fresh: critique the cluttered chart in your own words.

A useful chart checklist is:

1. one clear question,
2. one prepared summary,
3. a title,
4. labeled axes,
5. readable tick labels,
6. and no extra decoration.

That is enough. The point is not to impress the viewer. The point is to help the viewer read the data.

Now apply the checklist in a new context: the next exercise asks you to make a readable chart of workshop registrations.

df = pd.DataFrame([
    {"workshop": "Python Basics", "registrations": 18, "date": "2026-06-01"},
    {"workshop": "Data Stories", "registrations": 25, "date": "2026-06-01"},
    {"workshop": "Python Basics", "registrations": 12, "date": "2026-06-02"},
    {"workshop": "Web Design", "registrations": 10, "date": "2026-06-02"},
    {"workshop": "Web Design", "registrations": 8, "date": "2026-06-03"},
])

Check Your Understanding

1. Which three text elements help readers understand most charts?
2. When does rotating tick labels help, and when does it start to hurt?
3. When should bars get different colors?
4. What is clutter, and why does it matter?