Seaborn figure-level vs axes-level in Matplotlib - Visual Side-by-Side Comparison

Concept Flow - Seaborn figure-level vs axes-level

Start: Choose plot type

↓

Is plot figure-level?

Yes→Create new Figure + Axes internally

↓

Return Figure object

↓

Is plot axes-level?

Yes→Use existing Axes object

↓

Return Axes object

↓

Error: Unknown plot type

Decide if the plot creates a new figure (figure-level) or uses an existing axes (axes-level). Figure-level plots manage the whole figure, axes-level plots draw on given axes.

Execution Sample

Matplotlib

import seaborn as sns
import matplotlib.pyplot as plt

# Figure-level plot
sns.catplot(data=tips, x='day', y='total_bill')

# Axes-level plot
fig, ax = plt.subplots()
sns.boxplot(data=tips, x='day', y='total_bill', ax=ax)

Shows how figure-level plot creates its own figure, axes-level plot uses existing axes.

Execution Table

Step	Action	Plot Type	Figure Created	Axes Used	Return Object
1	Call sns.catplot()	Figure-level	Yes (new Figure)	New Axes created internally	FacetGrid (figure-level object)
2	Plot draws internally	Figure-level	Yes	Internal Axes	FacetGrid object manages figure and axes
3	Call plt.subplots()	Setup	Yes (new Figure)	New Axes returned	Figure and Axes objects
4	Call sns.boxplot() with ax=ax	Axes-level	No (uses existing Figure)	Existing Axes passed in	Axes object
5	Plot draws on passed Axes	Axes-level	No	Existing Axes	Axes object
6	End of plotting	-	-	-	-

💡 Execution stops after plots are drawn and objects returned.

Variable Tracker

Variable	Start	After sns.catplot()	After plt.subplots()	After sns.boxplot()
fig	None	Created internally by catplot	Created by plt.subplots()	Same as plt.subplots() figure
ax	None	Created internally by catplot	Created by plt.subplots()	Passed to sns.boxplot()
plot_obj	None	FacetGrid object returned	None	Axes object returned

Key Moments - 2 Insights

Why does sns.catplot() create a new figure while sns.boxplot() can use an existing axes?

sns.catplot() is a figure-level function that manages the whole figure and creates it internally (see execution_table step 1). sns.boxplot() is axes-level and draws on an existing axes passed via the ax parameter (see step 4).

What object type is returned by figure-level vs axes-level plots?

Figure-level plots like sns.catplot() return a FacetGrid object that manages figure and axes (step 2). Axes-level plots like sns.boxplot() return the Axes object they draw on (step 5).

Visual Quiz - 3 Questions

Look at the execution table, what is created internally by sns.catplot() at step 1?

A Only a new Figure

B A new Figure and Axes

C Only a new Axes

D No new objects

💡 Hint

Check the 'Figure Created' and 'Axes Used' columns at step 1 in the execution_table.

At which step does sns.boxplot() use an existing Axes object?

A Step 4

B Step 2

C Step 1

D Step 3

💡 Hint

Look at the 'Axes Used' column for sns.boxplot() in the execution_table.

If you do not pass ax to sns.boxplot(), what will happen?

A It will create a new figure and axes internally

B It will raise an error

C It will plot on the last active axes

D It will return a FacetGrid object

💡 Hint

Axes-level plots use existing axes or the current active axes if none is passed.

Concept Snapshot

Seaborn plots are either figure-level or axes-level.
Figure-level plots (e.g., catplot) create and manage their own figure and axes.
Axes-level plots (e.g., boxplot) draw on existing matplotlib axes passed via ax.
Figure-level returns a FacetGrid object; axes-level returns the Axes object.
Use figure-level for complex multi-plot layouts; axes-level for simple single plots.

Full Transcript

This visual execution trace shows the difference between Seaborn figure-level and axes-level plots. Figure-level plots like catplot create a new figure and axes internally and return a FacetGrid object that manages them. Axes-level plots like boxplot require an existing matplotlib axes object passed via the ax parameter and return that axes after drawing. The execution table traces calls and object creation step-by-step. Variable tracking shows how figure and axes variables change. Key moments clarify why figure-level plots create new figures and what objects are returned. The quiz tests understanding of which objects are created and used at each step. The concept snapshot summarizes the main differences and usage tips.

Practice

1. Which of the following best describes a figure-level function in Seaborn?

easy

A. It creates a complete plot including figure and axes automatically.

B. It only modifies existing axes without creating a new figure.

C. It is used to customize axis labels after plotting.

D. It requires manual creation of figure and axes before plotting.

Solution

1. Step 1: Understand figure-level function role

   Figure-level functions in Seaborn create the entire plot including figure and axes automatically.

2. Step 2: Compare with axes-level functions

   Axes-level functions only draw on existing axes and do not create the figure.

3. Final Answer:

   It creates a complete plot including figure and axes automatically. -> Option A

4. Quick Check:

   Figure-level = creates full plot [OK]

Hint: Figure-level functions create whole plots; axes-level modify existing axes [OK]

Common Mistakes:

• Confusing figure-level with axes-level functions
• Thinking figure-level functions require manual axes creation
• Assuming axes-level functions create figures automatically

2. Which of the following is the correct way to use an axes-level function in Seaborn on existing axes?

easy

A. sns.scatterplot(data=df, x='age', y='height')

B. sns.scatterplot(ax=ax, data=df, x='age', y='height')

C. sns.relplot(data=df, x='age', y='height')

D. sns.relplot(ax=ax, data=df, x='age', y='height')

Solution

1. Step 1: Identify axes-level function usage

   Axes-level functions like scatterplot can accept an ax parameter to plot on existing axes.

2. Step 2: Check options for correct syntax

   sns.scatterplot(ax=ax, data=df, x='age', y='height') correctly passes ax=ax to scatterplot, an axes-level function.

3. Final Answer:

   sns.scatterplot(ax=ax, data=df, x='age', y='height') -> Option B

4. Quick Check:

   Axes-level functions use ax= parameter [OK]

Hint: Axes-level functions accept ax= parameter to plot on existing axes [OK]

Common Mistakes:

• Using figure-level functions with ax= parameter (not supported)
• Confusing relplot (figure-level) with scatterplot (axes-level)
• Omitting ax= when plotting on existing axes

3. What will the following code output?

import seaborn as sns
import matplotlib.pyplot as plt

df = sns.load_dataset('tips')
fig, ax = plt.subplots()
sns.scatterplot(data=df, x='total_bill', y='tip', ax=ax)
plt.title('Scatterplot on existing axes')
plt.show()

medium

A. A scatterplot of total_bill vs tip on a new figure with default title

B. An empty plot with no points

C. A scatterplot of total_bill vs tip on the existing axes with custom title

D. An error because scatterplot cannot accept ax parameter

Solution

1. Step 1: Analyze code creating figure and axes

   The code creates a figure and axes with plt.subplots() and stores axes in ax.

2. Step 2: Understand scatterplot with ax parameter

   scatterplot is axes-level and plots on the given ax. Title is set on the figure.

3. Final Answer:

   A scatterplot of total_bill vs tip on the existing axes with custom title -> Option C

4. Quick Check:

   Axes-level plot on existing axes = scatterplot with ax= [OK]

Hint: Axes-level plots use ax= to draw on existing axes [OK]

Common Mistakes:

• Expecting scatterplot to create new figure automatically
• Thinking ax= causes error with scatterplot
• Assuming title applies only to figure-level plots

4. Identify the error in this code snippet:

import seaborn as sns
import matplotlib.pyplot as plt

df = sns.load_dataset('tips')
fig, ax = plt.subplots()
sns.relplot(data=df, x='total_bill', y='tip', ax=ax)
plt.show()

medium

A. plt.show() must be called before relplot

B. plt.subplots() is missing required arguments

C. DataFrame 'df' is not loaded correctly

D. relplot does not accept ax parameter; it creates its own figure

Solution

1. Step 1: Check relplot function parameters

   relplot is a figure-level function and does not accept an ax parameter.

2. Step 2: Understand error cause

   Passing ax=ax to relplot causes an error because it manages figure creation internally.

3. Final Answer:

   relplot does not accept ax parameter; it creates its own figure -> Option D

4. Quick Check:

   Figure-level functions ignore ax= and raise error if given [OK]

Hint: Figure-level functions like relplot do NOT accept ax= [OK]

Common Mistakes:

• Passing ax= to figure-level functions
• Confusing relplot with scatterplot usage
• Assuming plt.subplots() is incorrect here

5. You want to create a figure with two different plots side by side: a histogram and a scatterplot. Which approach correctly uses Seaborn's figure-level and axes-level functions together?

hard

A. Use sns.histplot() on one axes and sns.scatterplot() on another axes created by plt.subplots().

B. Use sns.relplot() twice, each creating its own figure, then combine figures manually.

C. Use sns.histplot() with ax= parameter and sns.relplot() without ax= on the same axes.

D. Use sns.relplot() with ax= parameter for both plots on shared axes.

Solution

1. Step 1: Understand figure-level vs axes-level functions

   relplot is figure-level and creates its own figure; histplot and scatterplot are axes-level and can plot on existing axes.

2. Step 2: Plan side-by-side plots

   Creating subplots with plt.subplots() and plotting axes-level functions on each axes allows side-by-side plots in one figure.

3. Step 3: Evaluate options

   Use sns.histplot() on one axes and sns.scatterplot() on another axes created by plt.subplots(). correctly uses axes-level functions on subplots. Options A, B, and D misuse figure-level functions or ax= parameter.

4. Final Answer:

   Use sns.histplot() on one axes and sns.scatterplot() on another axes created by plt.subplots(). -> Option A

5. Quick Check:

   Axes-level functions + subplots = combined figure [OK]

Hint: Use axes-level functions with plt.subplots() for combined plots [OK]

Common Mistakes:

• Trying to use relplot with ax= parameter
• Using multiple figure-level functions expecting one figure
• Mixing figure-level and axes-level functions on same axes