GoldenViz examples: 25 rules

Each example is a complete, standalone Matplotlib snippet. The good-chart code highlights the lines that were added or corrected.

Installation check output

This cell reproduces the installation smoke test. The first output is the Matplotlib chart; the second output is the GoldenViz report embedded as HTML.

import matplotlib.pyplot as plt
from IPython.display import HTML, display
from pathlib import Path
import GoldenViz as gv

fig, ax = plt.subplots(figsize=(6, 3.2))
ax.plot([2021, 2022, 2023], [10, 13, 16])
ax.set_title("Revenue trend by year")
ax.set_xlabel("Year")
ax.set_ylabel("Revenue (M EUR)")

# The chart output appears first.
display(fig)

# GoldenViz analyzes the figure; the HTML report below is embedded in this notebook.
report = gv.analyze(fig)
report_html = Path("_static/goldenviz_installation_check_report.html")
if not report_html.exists():
    report_html = Path("docs/_static/goldenviz_installation_check_report.html")
display(HTML(report_html.read_text(encoding="utf-8")))
plt.close(fig)

GoldenViz report Automatic visual QA for the 25 Golden Rules.

PASS 25 WARNING 0 FAIL 0

View warnings and failures

Axis: Revenue trend by year

Needs attention (0)

No warnings or failures for this axis.

Passing checks (25)

Rule	Status	Assessment
Clear title	PASS	Title detected: 'Revenue trend by year'.
Axis labels	PASS	Both x-axis and y-axis labels are present.
Units and scale clarity	PASS	Numeric labels include detectable units or scale where applicable.
Legend clarity	PASS	No obvious legend clarity issue detected.
Annotation context	PASS	No obvious annotation context issue detected. Revenue trend by year
Uncertainty cues	PASS	No uncertainty cue issue detected. Revenue trend by year
Readable labels and ticks	PASS	No obvious readability issue detected.
Color accessibility	PASS	No obvious color accessibility issue detected.
Direct labeling	PASS	No direct-labeling opportunity detected.
Avoid chartjunk	PASS	No obvious chartjunk issue detected. Revenue trend by year
Too many categories	PASS	Category count appears manageable. Revenue trend by year
Sort categorical bars	PASS	No categorical bar sorting issue detected. Revenue trend by year
Scatter overplotting	PASS	No scatter overplotting issue detected. Revenue trend by year
Decimal precision	PASS	Tick precision appears readable. Revenue trend by year
Date axis formatting	PASS	No obvious date-axis formatting issue detected. Revenue trend by year
Visual economy	PASS	Chart appears visually economical. Revenue trend by year
Appropriate scale	PASS	No obvious scale issue detected.
Chart type	PASS	Detected chart type: line.
Color map quality	PASS	No problematic color map detected.
Avoid dual axes	PASS	No dual-axis layout detected. Revenue trend by year
Area baseline	PASS	No area baseline issue detected. Revenue trend by year
Aspect ratio sanity	PASS	Figure aspect ratio appears reasonable. Revenue trend by year
Histogram bin quality	PASS	No histogram bin issue detected. Revenue trend by year
Category color consistency	PASS	No category color issue detected. Revenue trend by year
Diverging zero reference	PASS	No diverging zero-reference issue detected. Revenue trend by year

Completeness rules

Rules 1-6 check whether the chart contains enough information to be understood on its own.

Rule 1 · Clear title

A title should tell the reader what the chart is about before they inspect the marks.

Bad chart code

import matplotlib.pyplot as plt
 
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun"]
revenue = [18, 20, 21, 23, 25, 28]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(months, revenue, marker="o")
ax.set_xlabel("Month")
ax.set_ylabel("Revenue (k EUR)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun"]
revenue = [18, 20, 21, 23, 25, 28]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(months, revenue, marker="o")
ax.set_title("Monthly revenue increased from Jan to Jun 2026")
ax.set_xlabel("Month")
ax.set_ylabel("Revenue (k EUR)")
plt.show()

Rule 2 · Axis labels

Axis labels and units make the measurement unambiguous.

Bad chart code

import matplotlib.pyplot as plt
 
ad_spend = [8, 12, 16, 20, 24, 30, 34, 38]
signups = [110, 140, 175, 205, 245, 310, 335, 390]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.scatter(ad_spend, signups)
ax.set_title("Campaign performance")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
ad_spend = [8, 12, 16, 20, 24, 30, 34, 38]
signups = [110, 140, 175, 205, 245, 310, 335, 390]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.scatter(ad_spend, signups)
ax.set_title("Campaign performance")
ax.set_xlabel("Ad spend (k EUR)")
ax.set_ylabel("New signups")
plt.show()

Rule 3 · Units and scale clarity

Large values should be scaled and labeled so the reader knows what the numbers mean.

Bad chart code

import matplotlib.pyplot as plt
 
years = [2022, 2023, 2024, 2025]
revenue = [1250000, 1480000, 1730000, 1910000]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(years, revenue, marker="o")
ax.set_title("Revenue trend")
ax.set_ylabel("Revenue")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
years = [2022, 2023, 2024, 2025]
revenue = [1250000, 1480000, 1730000, 1910000]
revenue_millions = [value / 1_000_000 for value in revenue]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(years, revenue_millions, marker="o")
ax.set_title("Revenue trend")
ax.set_xlabel("Year")
ax.set_ylabel("Revenue (million EUR)")
plt.show()

Rule 4 · Legend clarity

Multiple series need clear labels so readers can identify each line.

Bad chart code

import matplotlib.pyplot as plt
 
years = [2021, 2022, 2023, 2024, 2025]
series = {
    "Basic": [12, 14, 16, 18, 20],
    "Pro": [9, 13, 17, 22, 28],
    "Enterprise": [6, 8, 12, 17, 25],
}
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
for values in series.values():
    ax.plot(years, values, marker="o")
ax.set_title("Subscriptions by plan")
ax.set_ylabel("Subscriptions (k)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
years = [2021, 2022, 2023, 2024, 2025]
series = {
    "Basic": [12, 14, 16, 18, 20],
    "Pro": [9, 13, 17, 22, 28],
    "Enterprise": [6, 8, 12, 17, 25],
}
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
for label, values in series.items():
    ax.plot(years, values, marker="o", label=label)
ax.legend(loc="center left", bbox_to_anchor=(1.02, 0.5))
ax.set_title("Subscriptions by plan")
ax.set_ylabel("Subscriptions (k)")
plt.show()

Rule 5 · Annotation context

Important events or takeaways should be visible when they explain the pattern.

Bad chart code

import matplotlib.pyplot as plt
 
weeks = [1, 2, 3, 4, 5, 6, 7, 8]
adoption = [12, 15, 18, 23, 34, 39, 43, 46]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(weeks, adoption, marker="o")
ax.set_title("Feature adoption")
ax.set_xlabel("Week")
ax.set_ylabel("Adoption (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
weeks = [1, 2, 3, 4, 5, 6, 7, 8]
adoption = [12, 15, 18, 23, 34, 39, 43, 46]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(weeks, adoption, marker="o")
ax.annotate("Onboarding email launched", xy=(5, 34), xytext=(3.2, 42), arrowprops={"arrowstyle": "->"})
ax.set_title("Feature adoption increased after onboarding email")
ax.set_xlabel("Week")
ax.set_ylabel("Adoption (%)")
plt.show()

Rule 6 · Uncertainty cues

Estimates should show uncertainty when the uncertainty matters.

Bad chart code

import matplotlib.pyplot as plt
 
groups = ["A", "B", "C", "D"]
mean = [52, 57, 61, 55]
error = [4, 7, 3, 6]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(groups, mean)
ax.set_title("Average test result")
ax.set_ylabel("Score")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
groups = ["A", "B", "C", "D"]
mean = [52, 57, 61, 55]
error = [4, 7, 3, 6]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(groups, mean, yerr=error, capsize=6)
ax.set_title("Average test result with uncertainty")
ax.set_ylabel("Score")
plt.show()

Readability rules

Rules 7-16 check whether the chart can be read, scanned, and compared without unnecessary effort.

Rule 7 · Readable labels

Long labels should not collide or force the reader to decode a crowded axis.

Bad chart code

import matplotlib.pyplot as plt
 
segments = [
    "Returning enterprise customers",
    "New small business customers",
    "One-time promotional buyers",
    "Students using education plan",
    "Trial users awaiting onboarding",
]
counts = [180, 145, 96, 125, 72]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(segments, counts)
ax.set_title("Customers by segment")
plt.show()

Good chart code

import matplotlib.pyplot as plt
import textwrap
 
segments = [
    "Returning enterprise customers",
    "New small business customers",
    "One-time promotional buyers",
    "Students using education plan",
    "Trial users awaiting onboarding",
]
counts = [180, 145, 96, 125, 72]
labels = [textwrap.fill(s, 22) for s in segments]
pairs = sorted(zip(counts, labels))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.barh([label for _, label in pairs], [count for count, _ in pairs])
ax.set_title("Customers by segment")
ax.set_xlabel("Customers")
plt.show()

Rule 8 · Color accessibility

Charts should remain readable for people with color-vision deficiencies.

Bad chart code

import matplotlib.pyplot as plt
 
segments = ["Home", "Food", "Auto", "Health", "Shopping"]
spend = [42, 26, 18, 12, 8]
colors = ["#d7191c", "#fdae61", "#ffffbf", "#a6d96a", "#1a9641"]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(segments, spend, color=colors)
ax.set_title("Household spend by category")
ax.set_ylabel("Spend (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
segments = ["Home", "Food", "Auto", "Health", "Shopping"]
spend = [42, 26, 18, 12, 8]
colors = ["#0072B2", "#E69F00", "#56B4E9", "#009E73", "#CC79A7"]
hatches = ["", "//", "\\", "..", "xx"]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
bars = ax.bar(segments, spend, color=colors)
for bar, hatch in zip(bars, hatches):
    bar.set_hatch(hatch)
ax.set_title("Household spend by category")
ax.set_ylabel("Spend (%)")
plt.show()

Rule 9 · Direct labeling

When lines are easy to label directly, readers should not have to bounce between the plot and legend.

Bad chart code

import matplotlib.pyplot as plt
 
years = [2021, 2022, 2023, 2024, 2025]
series = {
    "Basic": [12, 14, 16, 18, 20],
    "Pro": [9, 13, 17, 22, 28],
    "Enterprise": [6, 8, 12, 17, 25],
    "Education": [4, 7, 10, 13, 19],
}
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
for label, values in series.items():
    ax.plot(years, values, marker="o", label=label)
ax.legend()
ax.set_title("Subscriptions by plan")
ax.set_ylabel("Subscriptions (k)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
years = [2021, 2022, 2023, 2024, 2025]
series = {
    "Basic": [12, 14, 16, 18, 20],
    "Pro": [9, 13, 17, 22, 28],
    "Enterprise": [6, 8, 12, 17, 25],
    "Education": [4, 7, 10, 13, 19],
}
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
for label, values in series.items():
    ax.plot(years, values, marker="o")
    ax.text(years[-1] + 0.05, values[-1], label, va="center")
ax.set_xlim(2021, 2025.9)
ax.set_title("Subscriptions by plan")
ax.set_ylabel("Subscriptions (k)")
plt.show()

Rule 10 · Avoid chartjunk

Decoration should not compete with the data.

Bad chart code

import matplotlib.pyplot as plt
 
regions = ["North", "South", "East", "West"]
profit = [18, 12, 15, 21]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.set_facecolor("#f3d9a5")
ax.bar(regions, profit, edgecolor="black", linewidth=2)
ax.grid(True, axis="both")
ax.set_title("!!! PROFIT !!!")
ax.set_ylabel("Profit (k EUR)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
regions = ["North", "South", "East", "West"]
profit = [18, 12, 15, 21]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(regions, profit)
ax.grid(True, axis="y")
ax.set_title("Profit by region")
ax.set_ylabel("Profit (k EUR)")
plt.show()

Rule 11 · Too many categories

Too many categories make comparison slow and crowded.

Bad chart code

import matplotlib.pyplot as plt
 
categories = [f"C{i}" for i in range(1, 19)]
values = [34, 28, 26, 22, 20, 18, 16, 12, 11, 10, 9, 8, 7, 6, 5, 5, 4, 3]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(categories, values)
ax.set_title("Requests by category")
ax.set_ylabel("Requests")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
categories = [f"C{i}" for i in range(1, 19)]
values = [34, 28, 26, 22, 20, 18, 16, 12, 11, 10, 9, 8, 7, 6, 5, 5, 4, 3]
top_labels = categories[:7] + ["Other"]
top_values = values[:7] + [sum(values[7:])]
pairs = sorted(zip(top_values, top_labels))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.barh([label for _, label in pairs], [value for value, _ in pairs])
ax.set_title("Requests by category")
ax.set_xlabel("Requests")
plt.show()

Rule 12 · Sort categorical bars

Sorted bars make ranking and comparison faster.

Bad chart code

import matplotlib.pyplot as plt
 
teams = ["Ops", "Sales", "Support", "Product", "Finance"]
hours = [42, 18, 31, 24, 36]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.barh(teams, hours)
ax.set_title("Average resolution time")
ax.set_xlabel("Hours")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
teams = ["Ops", "Sales", "Support", "Product", "Finance"]
hours = [42, 18, 31, 24, 36]
pairs = sorted(zip(hours, teams))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.barh([team for _, team in pairs], [hour for hour, _ in pairs])
ax.set_title("Average resolution time")
ax.set_xlabel("Hours")
plt.show()

Rule 13 · Scatter overplotting

Repeated or dense points should reveal density instead of hiding it.

Bad chart code

import matplotlib.pyplot as plt
import numpy as np
 
rng = np.random.default_rng(42)
centers = np.array([[2, 2], [3, 3], [4, 2.5], [4.5, 4]])
points = np.repeat(centers, [35, 45, 25, 30], axis=0)
points = points + rng.normal(0, 0.04, size=(135, 2))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.scatter(points[:, 0], points[:, 1], s=28)
ax.set_xlabel("Value score")
ax.set_ylabel("Satisfaction score")
plt.show()

Good chart code

import matplotlib.pyplot as plt
import numpy as np
 
centers = np.array([[2, 2], [3, 3], [4, 2.5], [4.5, 4]])
counts = np.array([35, 45, 25, 30])
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.scatter(centers[:, 0], centers[:, 1], s=counts * 18, alpha=0.55, edgecolor="black")
ax.set_xlabel("Value score")
ax.set_ylabel("Satisfaction score")
ax.set_title("Point size shows repeated observations")
plt.show()

Rule 14 · Decimal precision

Labels should avoid unnecessary decimals that add noise without adding meaning.

Bad chart code

import matplotlib.pyplot as plt
 
products = ["A", "B", "C", "D"]
conversion = [12.3421, 11.9874, 13.2251, 12.7718]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
bars = ax.bar(products, conversion)
ax.bar_label(bars, fmt="%.4f%%")
ax.set_ylim(0, 15)
ax.set_ylabel("Conversion (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
products = ["A", "B", "C", "D"]
conversion = [12.3421, 11.9874, 13.2251, 12.7718]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
bars = ax.bar(products, conversion)
ax.bar_label(bars, fmt="%.1f%%")
ax.set_ylim(0, 15)
ax.set_ylabel("Conversion (%)")
plt.show()

Rule 15 · Date axis formatting

Date ticks should be formatted at a readable interval.

Bad chart code

import matplotlib.pyplot as plt
import numpy as np
 
dates = [np.datetime64("2026-01-01") + np.timedelta64(i * 14, "D") for i in range(10)]
visits = [120, 135, 128, 150, 162, 158, 170, 181, 190, 205]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(dates, visits, marker="o")
ax.set_title("Website visits")
ax.set_ylabel("Visits (k)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import numpy as np
 
dates = [np.datetime64("2026-01-01") + np.timedelta64(i * 14, "D") for i in range(10)]
visits = [120, 135, 128, 150, 162, 158, 170, 181, 190, 205]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(dates, visits, marker="o")
ax.set_title("Website visits")
ax.set_ylabel("Visits (k)")
ax.xaxis.set_major_locator(mdates.MonthLocator())
ax.xaxis.set_major_formatter(mdates.DateFormatter("%b %Y"))
ax.tick_params(axis="x", rotation=30)
plt.show()

Rule 16 · Visual economy

Use enough visual encoding to explain the data, but avoid redundant styling.

Bad chart code

import matplotlib.pyplot as plt
 
products = ["A", "B", "C", "D", "E", "F"]
values = [18, 25, 22, 31, 27, 20]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(products, values, hatch="//", edgecolor="black")
ax.plot(products, values, marker="D", color="red")
ax.grid(True, axis="both")
ax.set_title("Product sales with redundant styling")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
products = ["A", "B", "C", "D", "E", "F"]
values = [18, 25, 22, 31, 27, 20]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(products, values)
ax.grid(True, axis="y")
ax.set_title("Product sales")
ax.set_ylabel("Sales (k EUR)")
plt.show()

Integrity rules

Rules 17-25 check whether the chart avoids misleading scale, encoding, or comparison choices.

Rule 17 · Appropriate scale

For bars, a truncated baseline can exaggerate small differences.

Bad chart code

import matplotlib.pyplot as plt
 
branches = ["North", "Central", "South"]
satisfaction = [82, 86, 88]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(branches, satisfaction)
ax.set_ylim(80, 90)
ax.set_title("Customer satisfaction by branch")
ax.set_ylabel("Satisfied customers (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
branches = ["North", "Central", "South"]
satisfaction = [82, 86, 88]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(branches, satisfaction)
ax.set_ylim(0, 100)
ax.set_title("Customer satisfaction by branch")
ax.set_ylabel("Satisfied customers (%)")
plt.show()

Rule 18 · Suitable chart type

Categorical comparisons are easier to read as bars than as a line that implies sequence.

Bad chart code

import matplotlib.pyplot as plt
 
products = ["Shoes", "Bags", "Watches", "Perfume", "Jewelry"]
sales = [54, 31, 46, 28, 62]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(products, sales, marker="o")
ax.set_title("Sales by product category")
ax.set_ylabel("Sales (k EUR)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
products = ["Shoes", "Bags", "Watches", "Perfume", "Jewelry"]
sales = [54, 31, 46, 28, 62]
pairs = sorted(zip(sales, products))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.barh([p for _, p in pairs], [s for s, _ in pairs])
ax.set_title("Sales by product category")
ax.set_xlabel("Sales (k EUR)")
plt.show()

Rule 19 · Color map quality

Continuous color should use a perceptual scale and a labeled color bar.

Bad chart code

import matplotlib.pyplot as plt
import numpy as np
 
heat = np.outer(np.linspace(0, 1, 12), np.linspace(0.2, 1, 12))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
image = ax.imshow(heat, cmap="rainbow")
ax.set_title("Demand intensity")
fig.colorbar(image, ax=ax)
plt.show()

Good chart code

import matplotlib.pyplot as plt
import numpy as np
 
heat = np.outer(np.linspace(0, 1, 12), np.linspace(0.2, 1, 12))
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
image = ax.imshow(heat, cmap="viridis")
ax.set_title("Demand intensity")
colorbar = fig.colorbar(image, ax=ax)
colorbar.set_label("Orders per store")
plt.show()

Rule 20 · Avoid dual axes

Dual axes can imply relationships that come from scale choices rather than data.

Bad chart code

import matplotlib.pyplot as plt
 
months = [1, 2, 3, 4, 5, 6]
revenue = [1.2, 1.5, 1.8, 2.1, 2.4, 2.7]
churn = [9, 8, 7, 6, 5, 4]
 
fig, ax1 = plt.subplots(figsize=(6.4, 3.8))
ax2 = ax1.twinx()
ax1.plot(months, revenue, marker="o", label="Revenue")
ax2.plot(months, churn, marker="o", color="red", label="Churn")
ax1.set_ylabel("Revenue (M EUR)")
ax2.set_ylabel("Churn (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
months = [1, 2, 3, 4, 5, 6]
revenue = [1.2, 1.5, 1.8, 2.1, 2.4, 2.7]
churn = [9, 8, 7, 6, 5, 4]
 
fig, axes = plt.subplots(2, 1, figsize=(6.4, 4.2), sharex=True)
axes[0].plot(months, revenue, marker="o")
axes[0].set_ylabel("Revenue (M EUR)")
axes[1].plot(months, churn, marker="o", color="red")
axes[1].set_ylabel("Churn (%)")
axes[1].set_xlabel("Month")
plt.show()

Rule 21 · Area baseline

Filled areas should use an honest baseline because area size carries meaning.

Bad chart code

import matplotlib.pyplot as plt
 
years = [2020, 2021, 2022, 2023, 2024, 2025]
share = [62, 64, 66, 67, 69, 70]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.fill_between(years, share, 60)
ax.plot(years, share, marker="o")
ax.set_ylim(60, 72)
ax.set_ylabel("Share (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
years = [2020, 2021, 2022, 2023, 2024, 2025]
share = [62, 64, 66, 67, 69, 70]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.fill_between(years, share, 0)
ax.plot(years, share, marker="o")
ax.set_ylim(0, 100)
ax.set_ylabel("Share (%)")
plt.show()

Rule 22 · Aspect ratio sanity

Extreme aspect ratios can flatten or exaggerate trends.

Bad chart code

import matplotlib.pyplot as plt
 
quarters = [1, 2, 3, 4, 5, 6, 7, 8]
index = [12, 14, 15, 16, 18, 19, 21, 22]
 
fig, ax = plt.subplots(figsize=(7.8, 2.2))
ax.plot(quarters, index, marker="o")
ax.set_title("Compressed aspect ratio hides the trend")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
quarters = [1, 2, 3, 4, 5, 6, 7, 8]
index = [12, 14, 15, 16, 18, 19, 21, 22]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(quarters, index, marker="o")
ax.set_title("Balanced aspect ratio shows the trend clearly")
plt.show()

Rule 23 · Histogram bin quality

A histogram needs enough bins to reveal the distribution shape without creating noise.

Bad chart code

import matplotlib.pyplot as plt
import numpy as np
 
rng = np.random.default_rng(42)
scores = rng.normal(72, 9, 420)
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.hist(scores, bins=3)
ax.set_title("Exam scores")
ax.set_xlabel("Score")
ax.set_ylabel("Students")
plt.show()

Good chart code

import matplotlib.pyplot as plt
import numpy as np
 
rng = np.random.default_rng(42)
scores = rng.normal(72, 9, 420)
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.hist(scores, bins=18)
ax.set_title("Exam scores")
ax.set_xlabel("Score")
ax.set_ylabel("Students")
plt.show()

Rule 24 · Category color consistency

The same category should keep the same color throughout a chart.

Bad chart code

import matplotlib.pyplot as plt
 
quarters = ["Q1", "Q2", "Q3"]
desktop = [42, 45, 48]
mobile = [31, 34, 38]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(quarters, desktop, marker="o", color="blue", label="Desktop")
ax.plot(quarters, mobile, marker="o", color="orange", label="Mobile")
ax.scatter(["Q2"], [45], color="orange", s=90)
ax.scatter(["Q2"], [34], color="blue", s=90)
ax.legend()
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
quarters = ["Q1", "Q2", "Q3"]
desktop = [42, 45, 48]
mobile = [31, 34, 38]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.plot(quarters, desktop, marker="o", color="blue", label="Desktop")
ax.plot(quarters, mobile, marker="o", color="orange", label="Mobile")
ax.legend()
ax.set_title("Traffic by device")
ax.set_ylabel("Sessions (k)")
plt.show()

Rule 25 · Diverging zero reference

Positive and negative values need a clear zero reference.

Bad chart code

import matplotlib.pyplot as plt
 
departments = ["Ops", "Sales", "Support", "Product"]
delta = [-8, 12, -5, 9]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(departments, delta)
ax.set_title("Change vs target")
ax.set_ylabel("Change (%)")
plt.show()

Good chart code

import matplotlib.pyplot as plt
 
departments = ["Ops", "Sales", "Support", "Product"]
delta = [-8, 12, -5, 9]
colors = ["#b42318" if value < 0 else "#2f855a" for value in delta]
 
fig, ax = plt.subplots(figsize=(6.4, 3.8))
ax.bar(departments, delta, color=colors)
ax.axhline(0, color="black", linewidth=1.2)
ax.set_title("Change vs target")
ax.set_ylabel("Change (%)")
plt.show()