Exercises: Customization Mastery

These exercises are hands-on. Run them. Modify them. The difference between a matplotlib user who has "read about" customization and one who has "practiced" it is large, and it only closes with repetition.

Part A: Conceptual (6 problems)

A.1 ★☆☆ | Recall

State the four ways matplotlib accepts color specifications. Give an example of each.

A.2 ★☆☆ | Recall

What is rcParams in matplotlib? Give two examples of settings that are commonly overridden.

A.3 ★★☆ | Understand

Explain the difference between setting rcParams globally (mpl.rcParams[...] = ...) and using a context manager (with plt.rc_context({...}):). When would you prefer each?

A.4 ★★☆ | Understand

The chapter's threshold concept says that "professional matplotlib is systematic, not ad hoc." Explain what this means and describe the two main tools for building a style system.

A.5 ★★☆ | Analyze

The chapter argues that the same rcParams can produce three different "house styles" for different contexts (academic, news, slide deck). Identify at least three rcParams settings that would differ across the three contexts, and explain why.

A.6 ★★★ | Evaluate

The chapter includes a "publication-ready checklist" in Section 12.12. In your own words, explain why the checklist is organized around Chapters 4, 6, 7, 8, and 12. What does each chapter contribute to the "is this chart ready?" question?

Part B: Applied Customization (11 problems)

B.1 ★★☆ | Apply

Take the "ugly climate plot" from Chapter 10 Section 10.9 and apply the Chapter 6 declutter procedure through matplotlib method calls. Specifically: remove the top and right spines, lighten the remaining spines to gray, remove any heavy grid lines, and add light horizontal gridlines only. Save the result.

fig, ax = plt.subplots(figsize=(12, 4))
ax.plot(climate["year"], climate["anomaly"], color="#d62728", linewidth=1.5)

# Declutter
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
ax.spines["left"].set_color("gray")
ax.spines["bottom"].set_color("gray")
ax.spines["left"].set_linewidth(0.8)
ax.spines["bottom"].set_linewidth(0.8)

# Light horizontal gridlines
ax.grid(True, axis="y", linestyle="-", linewidth=0.5, color="#cccccc", alpha=0.6, zorder=0)
ax.set_axisbelow(True)

ax.set_title("Temperature Anomaly (Decluttered)")
ax.set_ylabel("Temperature Anomaly (°C)")
fig.savefig("climate_decluttered.png", dpi=300, bbox_inches="tight")

B.2 ★★☆ | Apply

Write a descriptive title for your decluttered climate chart and then rewrite it as an action title (per Chapter 7). Apply the action title with ax.set_title(..., loc="left", fontweight="semibold", fontsize=14, pad=12).

B.3 ★★☆ | Apply

Add annotations to the decluttered climate chart calling out the 2016 and 2023 record years. Use ax.annotate() with both xy (the target point) and xytext (the text position) and include an arrow.

ax.annotate(
    "2023: +1.18°C record",
    xy=(2023, 1.18),
    xytext=(2005, 1.4),
    fontsize=10,
    color="#8b0000",
    arrowprops=dict(arrowstyle="->", color="gray", lw=0.6),
)

ax.annotate(
    "2016: +1.01°C",
    xy=(2016, 1.01),
    xytext=(2000, 1.22),
    fontsize=10,
    color="#8b0000",
    arrowprops=dict(arrowstyle="->", color="gray", lw=0.6),
)

B.4 ★★☆ | Apply

Add an on-image source attribution to your climate chart using fig.text(...). Place it at the bottom left of the figure in small gray italic text.

fig.text(
    0.125, 0.02,
    "Source: NASA Goddard Institute for Space Studies. Baseline: 1951-1980.",
    fontsize=8,
    color="gray",
    style="italic",
)

B.5 ★★☆ | Apply

Write a Python function apply_clean_style(ax) that takes a matplotlib Axes and applies the declutter principles: removes top/right spines, lightens the remaining spines, adds light horizontal gridlines, and sets tick parameters. The function should return the Axes.

def apply_clean_style(ax):
    ax.spines["top"].set_visible(False)
    ax.spines["right"].set_visible(False)
    ax.spines["left"].set_color("gray")
    ax.spines["bottom"].set_color("gray")
    ax.spines["left"].set_linewidth(0.8)
    ax.spines["bottom"].set_linewidth(0.8)
    ax.tick_params(
        axis="both", which="major",
        labelsize=10, colors="gray",
        length=4, width=0.5, direction="out",
    )
    ax.grid(True, axis="y", linestyle="-", linewidth=0.5, color="#cccccc", alpha=0.6, zorder=0)
    ax.set_axisbelow(True)
    return ax

B.6 ★★☆ | Apply

Create a custom matplotlib style sheet as a .mplstyle file with at least 10 specific rcParams settings. Apply it with plt.style.use("/path/to/your/style.mplstyle") and verify that a default chart inherits the settings.

# my_style.mplstyle
font.family: sans-serif
font.sans-serif: Inter, Helvetica, Arial
font.size: 11
figure.figsize: 10, 6
figure.dpi: 100
savefig.dpi: 300
savefig.bbox: tight
axes.spines.top: False
axes.spines.right: False
axes.titlesize: 14
axes.titleweight: semibold
axes.labelsize: 11
axes.grid: True
axes.grid.axis: y
axes.axisbelow: True
xtick.labelsize: 9
ytick.labelsize: 9
grid.color: cccccc
grid.linewidth: 0.5
grid.alpha: 0.6
pdf.fonttype: 42
ps.fonttype: 42

B.7 ★★☆ | Apply

Format a y-axis that shows values in millions of dollars with thousand-separators and a dollar sign prefix (e.g., $1,500`, `$15,000). Use matplotlib.ticker.StrMethodFormatter or FuncFormatter.

from matplotlib.ticker import FuncFormatter

def dollars(x, pos):
    return f"${x:,.0f}"

ax.yaxis.set_major_formatter(FuncFormatter(dollars))

from matplotlib.ticker import StrMethodFormatter
ax.yaxis.set_major_formatter(StrMethodFormatter("${x:,.0f}"))

B.8 ★★☆ | Apply

Create a bar chart with five categories, and use ax.set_ylim(0, max_value * 1.1) to explicitly enforce Chapter 4's zero-baseline rule. Verify that the bars sit on a visible zero baseline.

categories = ["A", "B", "C", "D", "E"]
values = [120, 85, 150, 95, 200]

fig, ax = plt.subplots(figsize=(8, 5))
ax.bar(categories, values, color="steelblue")
ax.set_ylim(0, max(values) * 1.1)
ax.axhline(0, color="gray", linewidth=0.8)
ax.set_title("Bar Chart with Explicit Zero Baseline")
ax.set_ylabel("Value")

B.9 ★★★ | Create

Write a reusable function style_publication(ax, title=None, subtitle=None, xlabel=None, ylabel=None, source=None) that applies full publication styling: the declutter from B.5, plus a left-aligned action title, optional subtitle, axis labels, and on-image source attribution.

def style_publication(ax, title=None, subtitle=None, xlabel=None, ylabel=None, source=None):
    apply_clean_style(ax)  # from B.5

    if title:
        ax.set_title(title, fontsize=14, loc="left", fontweight="semibold", pad=12)
    if subtitle:
        fig = ax.get_figure()
        fig.text(0.125, 0.91, subtitle, fontsize=11, color="gray")
    if xlabel:
        ax.set_xlabel(xlabel, fontsize=11)
    if ylabel:
        ax.set_ylabel(ylabel, fontsize=11)
    if source:
        fig = ax.get_figure()
        fig.text(0.125, 0.02, source, fontsize=8, color="gray", style="italic")

    return ax

fig, ax = plt.subplots(figsize=(12, 4))
ax.plot(years, anomaly, color="#d62728")
style_publication(
    ax,
    title="Global Temperatures Have Risen 1.2°C Since 1900",
    subtitle="NASA GISS surface temperature analysis",
    ylabel="Temperature Anomaly (°C)",
    source="Source: NASA GISS",
)

B.10 ★★★ | Create

Combine everything you have built in B.1-B.9 into a single script that loads the climate data, creates a decluttered chart, applies the publication style, adds annotations, and saves the result at dpi=300. This is the final polished version of the climate chart from Section 12.13.

B.11 ★★★ | Create

Save your polished climate chart in three different formats: PNG (dpi=300), SVG (for editing), and PDF (with pdf.fonttype=42 set). Open each in an appropriate viewer and verify that the fonts render correctly.

import matplotlib as mpl
mpl.rcParams["pdf.fonttype"] = 42
mpl.rcParams["ps.fonttype"] = 42

# ... chart code ...

fig.savefig("climate.png", dpi=300, bbox_inches="tight")
fig.savefig("climate.svg", bbox_inches="tight")
fig.savefig("climate.pdf", bbox_inches="tight")

Part C: Synthesis and Style Systems (5 problems)

C.1 ★★★ | Evaluate

Take a chart you have previously made (any chart, from any project) and apply the publication-ready checklist from Section 12.12. For each item on the checklist, note whether your chart passes or fails. For any failures, describe the fix.

C.2 ★★★ | Create

Design your own personal matplotlib style system. Write a brief document (under 500 words) describing: (a) the house style you want to apply by default (fonts, colors, spine treatment, etc.), (b) the matplotlib implementation (as a .mplstyle file, rcParams block, or style function), and (c) the contexts in which you would deviate from the default (e.g., slide decks, academic papers, web graphics).

C.3 ★★★ | Create

Take the climate chart and produce three versions: a print journal version (single-column, serif, black-and-white), a news graphics version (wide, colored fill, action title), and a slide deck version (dark theme, large fonts). Use context managers and rc_context to isolate the styles.

C.4 ★★★ | Analyze

The chapter says matplotlib customization is "systematic, not ad hoc." Look at matplotlib code from a real-world data analysis project you have done (or find one on GitHub). Identify places where the code would benefit from systematizing: repeated customization blocks that could be factored into a function, ad-hoc styling choices that could be moved to a style sheet, etc. Propose a refactor.

C.5 ★★★ | Evaluate

The chapter's threshold concept says that professional matplotlib is systematic. Defend or critique this claim. Are there contexts where ad-hoc customization is actually better than a systematic style system? Where is the line between "appropriate customization" and "inconsistent code smell"?

These exercises are hands-on. The payoff for doing them is large: once you have built a personal style function and a custom .mplstyle file, every subsequent chart you produce inherits your discipline automatically. Do at least five Part B exercises before moving on to Chapter 13.