Small multiples (facet grid) in Matplotlib - Step-by-Step Execution

Concept Flow - Small multiples (facet grid)

Load data

↓

Choose variable to facet by

↓

Create grid layout

↓

Plot each subset in its grid cell

↓

Display all plots together

We split data by a variable, create a grid, and plot each subset in its own small plot, showing all plots together.

Execution Sample

Matplotlib

import matplotlib.pyplot as plt
import seaborn as sns

sns.set()
data = sns.load_dataset('tips')
g = sns.FacetGrid(data, col='time')
g.map(plt.hist, 'total_bill')
plt.show()

This code loads a dataset, creates a facet grid by 'time', plots histograms of 'total_bill' for each time, and shows the plots.

Execution Table

Step	Action	Evaluation	Result
1	Load dataset 'tips'	Data loaded with columns including 'time' and 'total_bill'	DataFrame with 244 rows
2	Create FacetGrid with col='time'	Unique 'time' values: ['Lunch', 'Dinner']	Grid with 2 columns for Lunch and Dinner
3	Map histogram of 'total_bill' to each facet	For Lunch: histogram bins calculated For Dinner: histogram bins calculated	Two histograms plotted separately
4	Show plot	All facets displayed side by side	Window with 2 histograms labeled Lunch and Dinner

💡 All subsets plotted; visualization complete

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	Final
data	None	DataFrame loaded	Same	Same	Same
g	None	None	FacetGrid object with 2 facets	Same	Same

Key Moments - 2 Insights

Why do we see two separate plots instead of one combined plot?

Because the FacetGrid splits data by the 'time' variable into Lunch and Dinner subsets, then plots each subset in its own grid cell as shown in execution_table step 2 and 3.

What does the map function do in the FacetGrid?

It applies the plotting function (here plt.hist) to each subset of data in the grid cells, creating separate histograms as shown in execution_table step 3.

Visual Quiz - 3 Questions

Look at the execution table, how many facets does the grid have after step 2?

A 1

B 2

C 3

D 4

💡 Hint

Check execution_table row for step 2 showing unique 'time' values and grid columns

At which step are the histograms for each subset created?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Look at execution_table row describing mapping histograms to facets

If we changed col='time' to col='day', what would change in the execution table?

A Number of facets in step 2 would change to unique days

B Histograms would plot 'time' instead of 'total_bill'

C Data would be loaded differently in step 1

D No change at all

💡 Hint

FacetGrid splits data by the column specified in step 2

Concept Snapshot

Small multiples (facet grid) split data by a variable.
Create a grid layout with one plot per subset.
Use FacetGrid in seaborn to manage this.
Map plotting functions to each facet.
Shows multiple small plots side by side for easy comparison.

Full Transcript

Small multiples or facet grids help us see data split by a category. We load data, pick a variable to split by, then create a grid with one plot per category. Each plot shows data only for that category. In the example, we load the 'tips' dataset, split by 'time' (Lunch or Dinner), and plot histograms of total bills for each. The FacetGrid creates two plots side by side. This helps compare distributions easily. The map function applies the histogram to each subset. Finally, we show all plots together. This method is great to compare groups visually.

Practice

1. What is the main purpose of using small multiples (facet grid) in matplotlib?

easy

A. To display multiple related charts side by side for easy comparison

B. To create a single large plot with multiple lines

C. To animate a plot over time

D. To change the color scheme of a plot

Solution

1. Step 1: Understand the concept of small multiples

   Small multiples are multiple small charts arranged in a grid to compare different groups or categories easily.

2. Step 2: Identify the purpose in matplotlib

   Matplotlib uses small multiples to show many charts side by side, making it easier to compare data visually.

3. Final Answer:

   To display multiple related charts side by side for easy comparison -> Option A

4. Quick Check:

   Small multiples = multiple charts side by side [OK]

Hint: Small multiples = many small charts for comparison [OK]

Common Mistakes:

• Confusing small multiples with animations
• Thinking it changes colors only
• Assuming it creates one big plot

2. Which of the following is the correct way to create a 2x2 grid of subplots in matplotlib?

easy

A. fig, axes = plt.subplots(4)

B. fig, axes = plt.subplots(1, 4)

C. fig, axes = plt.subplots(2, 2)

D. fig, axes = plt.subplots(2)

Solution

1. Step 1: Recall plt.subplots() syntax

   plt.subplots(rows, columns) creates a grid of subplots with given rows and columns.

2. Step 2: Match the grid size

   To create a 2x2 grid, use plt.subplots(2, 2).

3. Final Answer:

   fig, axes = plt.subplots(2, 2) -> Option C

4. Quick Check:

   plt.subplots(2, 2) = 2 rows and 2 columns [OK]

Hint: Use plt.subplots(rows, columns) for grid size [OK]

Common Mistakes:

• Using single number for grid shape
• Confusing rows and columns
• Missing one dimension in arguments

3. What will be the output of this code snippet?

import matplotlib.pyplot as plt
fig, axes = plt.subplots(1, 3)
data = [1, 2, 3]
for i, ax in enumerate(axes):
    ax.plot([data[i]] * 3)
plt.show()

medium

A. Three line plots each with three points of the same value 1, 2, and 3 respectively

B. A single plot with three lines of values 1, 2, and 3

C. Error because axes is not iterable

D. Three scatter plots with points 1, 2, and 3

Solution

1. Step 1: Understand plt.subplots(1, 3)

   This creates 1 row and 3 columns of subplots, so axes is an array of 3 axes objects.

2. Step 2: Loop plots each subplot

   For each axis, it plots a line with three points all equal to data[i] (1, then 2, then 3).

3. Final Answer:

   Three line plots each with three points of the same value 1, 2, and 3 respectively -> Option A

4. Quick Check:

   Loop over axes plots lines with repeated values [OK]

Hint: Loop over axes to plot each subplot separately [OK]

Common Mistakes:

• Thinking axes is a single plot
• Assuming error due to axes type
• Confusing line plot with scatter plot

4. Identify the error in this code for creating a 2x2 grid of plots:

fig, axes = plt.subplots(2, 2)
for i in range(4):
    axes[i].plot([1, 2, 3])
plt.show()

medium

A. The plot data list is invalid

B. plt.subplots(2, 2) creates only 2 plots, not 4

C. plt.show() is missing

D. axes is a 2D array, so axes[i] indexing causes an error

Solution

1. Step 1: Understand axes shape from plt.subplots(2, 2)

   axes is a 2x2 numpy array of Axes objects, not a flat list.

2. Step 2: Identify indexing error

   axes[i] tries to index a 2D array with one index, causing an error. Correct is axes[row, col] or flatten axes first.

3. Final Answer:

   axes is a 2D array, so axes[i] indexing causes an error -> Option D

4. Quick Check:

   2D axes need two indices or flatten before looping [OK]

Hint: 2D axes need two indices or flatten before indexing [OK]

Common Mistakes:

• Assuming axes is 1D array
• Ignoring error from wrong indexing
• Thinking plt.show() is missing

5. You have a dataset with sales data for 3 regions. How would you create a 1-row, 3-column grid of plots showing sales trends for each region separately using matplotlib?

hard

A. Use plt.plot() three times without subplots

B. Use plt.subplots(1, 3), loop over axes, and plot each region's data on each subplot

C. Use plt.subplots(3, 1) and plot all regions on each subplot

D. Use plt.subplots(1, 1) and plot all regions on the same plot

Solution

1. Step 1: Create a 1x3 grid for three regions

   Use plt.subplots(1, 3) to create one row and three columns of plots.

2. Step 2: Loop over axes and plot each region's data

   Loop through each axis and plot the sales data for each region separately to compare trends side by side.

3. Final Answer:

   Use plt.subplots(1, 3), loop over axes, and plot each region's data on each subplot -> Option B

4. Quick Check:

   One row, three columns, loop to plot each region [OK]

Hint: Create grid then loop axes to plot groups separately [OK]

Common Mistakes:

• Plotting all data on one plot
• Using wrong grid shape
• Not looping over axes