Exercises: seaborn Philosophy

These exercises require a Python environment with seaborn installed (pip install seaborn). All exercises assume import seaborn as sns and use seaborn's built-in datasets.

Part A: Conceptual (6 problems)

A.1 ★☆☆ | Recall

Name the three seaborn function families and describe the type of question each answers.

A.2 ★☆☆ | Recall

Explain the difference between figure-level and axes-level seaborn functions. Give one example of each.

A.3 ★★☆ | Understand

What does "tidy data" mean in the context of seaborn, and why does seaborn prefer it?

A.4 ★★☆ | Understand

Explain what hue, style, and size parameters do in seaborn. Give an example using all three.

A.5 ★★☆ | Analyze

The chapter's threshold concept is the shift from imperative matplotlib to declarative seaborn. Explain this shift in your own words and give a specific example where it matters.

A.6 ★★★ | Evaluate

The chapter says that seaborn is not always better than matplotlib. Under what circumstances would you prefer matplotlib?

Part B: Applied (10 problems)

B.1 ★☆☆ | Apply

Install seaborn (if not already installed), import it as sns, and load the tips dataset. Print the first 5 rows.

import seaborn as sns
tips = sns.load_dataset("tips")
print(tips.head())

B.2 ★☆☆ | Apply

Create a simple scatter plot of tip vs. total_bill from the tips dataset using sns.scatterplot.

import seaborn as sns
import matplotlib.pyplot as plt

tips = sns.load_dataset("tips")
sns.scatterplot(data=tips, x="total_bill", y="tip")
plt.show()

B.3 ★★☆ | Apply

Extend B.2 to color points by smoker (yes/no) using the hue parameter. Add a title and save as PNG.

fig, ax = plt.subplots(figsize=(8, 6))
sns.scatterplot(data=tips, x="total_bill", y="tip", hue="smoker", ax=ax)
ax.set_title("Tip vs. Total Bill, by Smoker Status")
fig.savefig("tips_by_smoker.png", dpi=300, bbox_inches="tight")

B.4 ★★☆ | Apply

Use sns.relplot (figure-level) to create a scatter plot of tip vs. total_bill, faceted by day (one panel per day of the week).

g = sns.relplot(
    data=tips,
    x="total_bill",
    y="tip",
    col="day",
    kind="scatter",
    height=3,
    aspect=1,
)

B.5 ★★☆ | Apply

Apply a seaborn theme with sns.set_theme(style="whitegrid", context="notebook") and reproduce the chart from B.3. Compare the appearance to the un-themed version.

sns.set_theme(style="whitegrid", context="notebook")

fig, ax = plt.subplots(figsize=(8, 6))
sns.scatterplot(data=tips, x="total_bill", y="tip", hue="smoker", ax=ax)
ax.set_title("Tip vs. Total Bill, by Smoker Status")

B.6 ★★☆ | Apply

Create a wide-form DataFrame with columns [year, temperature, co2, sea_level] using synthetic data. Convert it to tidy form using pd.melt and verify the result has three columns: year, variable, value.

import pandas as pd
import numpy as np

np.random.seed(42)
years = np.arange(1880, 2025)
wide = pd.DataFrame({
    "year": years,
    "temperature": -0.3 + (years - 1880) * 0.01,
    "co2": 290 + (years - 1880) * 0.9,
    "sea_level": (years - 1880) * 1.5,
})

tidy = wide.melt(id_vars="year", var_name="variable", value_name="value")
print(tidy.head(10))
print(tidy.columns.tolist())  # ['year', 'variable', 'value']

B.7 ★★★ | Apply

Use the tidy DataFrame from B.6 to create a three-panel line chart with sns.relplot, one panel per variable, using col_wrap=1 for vertical stacking and facet_kws={"sharey": False}.

g = sns.relplot(
    data=tidy,
    x="year",
    y="value",
    col="variable",
    kind="line",
    col_wrap=1,
    height=3,
    aspect=3.5,
    facet_kws={"sharey": False},
)

B.8 ★★☆ | Apply

Access the underlying matplotlib Axes for B.7's FacetGrid and customize each panel's title.

variable_names = {"temperature": "Temperature (°C)", "co2": "CO2 (ppm)", "sea_level": "Sea Level (mm)"}
for ax, var in zip(g.axes.flat, tidy["variable"].unique()):
    ax.set_title(variable_names[var], fontsize=10, loc="left")

B.9 ★★☆ | Apply

Load the penguins dataset with sns.load_dataset("penguins"). Create a histogram of body_mass_g split by species using sns.histplot with hue and multiple="stack".

penguins = sns.load_dataset("penguins")

fig, ax = plt.subplots(figsize=(8, 5))
sns.histplot(data=penguins, x="body_mass_g", hue="species", multiple="stack", ax=ax)
ax.set_title("Penguin Body Mass by Species")

B.10 ★★★ | Create

Using the penguins dataset, create a faceted chart with sns.relplot that shows bill_length_mm vs. bill_depth_mm, colored by species, with a separate panel for each island.

g = sns.relplot(
    data=penguins,
    x="bill_length_mm",
    y="bill_depth_mm",
    hue="species",
    col="island",
    kind="scatter",
    height=4,
    aspect=1,
)

Part C: Synthesis (4 problems)

C.1 ★★★ | Analyze

Compare the matplotlib and seaborn versions of a three-panel climate chart (Section 16.10). Count the lines of code in each version. Identify three specific things seaborn handles automatically that matplotlib requires you to write explicitly.

C.2 ★★★ | Evaluate

Find a matplotlib tutorial or Stack Overflow answer that shows a multi-panel grouped chart (grouped bar chart, small multiple, or similar). Rewrite it in seaborn. Compare the two versions for readability and conciseness.

C.3 ★★★ | Create

Build a personal "seaborn starter template" file that includes the standard imports, a set_theme call with your preferred style and context, and a comment block describing when to use figure-level vs. axes-level functions. Save it for reuse.

"""seaborn starter template."""

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

sns.set_theme(style="whitegrid", context="notebook", palette="colorblind")

# Usage notes:
# - For faceted charts: use figure-level (relplot, catplot, displot)
# - For single charts integrated with matplotlib layouts: use axes-level (scatterplot, lineplot, histplot)
# - For tidy data in a DataFrame, prefer the data= parameter with column names
# - For customization beyond seaborn's parameters, access .fig and .axes after the plot

C.4 ★★★ | Evaluate

The chapter argues that seaborn is a "higher-level interface to the same system" (matplotlib). Does this mean you should always learn matplotlib first, or can you learn seaborn without understanding matplotlib?

These exercises introduce the seaborn API. Do at least five Part B exercises before moving on to Chapter 17, which uses the same API for distributional visualization.