Visualizations

Last updated on 2024-12-11 | Edit this page

Estimated time: 120 minutes

Overview

Questions

How can I visualize tabular data in Python?
How can I group several plots together?

Objectives

create graphs and other visualizations using tabular data
group plots together to make comparative visualizations

Introduction

In this lesson, we will explore how to create visualizations of your data using three popular Python libraries:

Matplotlib is a foundational library for creating static visualizations in Python. It provides a wide range of charts, such as line plots, bar charts, scatter plots, histograms, and more. While it offers great flexibility, it requires more code for customization, making it best suited for basic to moderately complex visualizations.
Seaborn builds on top of Matplotlib and provides a high-level interface for creating attractive and informative statistical graphics. Seaborn comes with several built-in themes and color palettes, and it simplifies many common tasks, such as visualizing data distributions and relationships between variables. It’s especially powerful when working with Pandas DataFrames and when creating plots like boxplots, heatmaps, and pair plots.
Plotly is a modern, interactive graphing library that allows you to create beautiful and interactive web-based visualizations. It is designed for creating visualizations that allow users to zoom, hover, and interact with the chart dynamically. Plotly is particularly useful for creating dashboards, 3D plots, and other interactive visualizations that engage users in exploring the data.

These three libraries can be imported using the following aliases:

PYTHON

import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px

Basic plots with `matplotlib`

Creating a figure and axes

In matplotlib, the plt.subplots() function is a common way to create a figure (fig) and axes (ax) objects that you can work with to create and customize your plots.

What are Figures and Axes?

Figure (fig):
- The figure is the entire container for your plot. It holds everything—axes, titles, labels, and any other elements of the plot.
- You can have multiple figures in a single Python session, and each figure can hold multiple subplots or axes.
Axes (ax):
- The axes is where your actual plot will appear. It’s a region of the figure that holds the graph. Each axes object has methods for plotting, adding titles, modifying labels, and other customizations.
- An axes can represent various types of charts, like line plots, bar charts, histograms, etc. Each plot you create will be associated with one axes.

2 ways to create a matplotlib plot

In matplotlib, there are two main ways to create plots:

The state-based interface using plt.plot() and similar functions. This is simpler and often fine for quick plots.

PYTHON

# Basic line plot
plt.plot([1, 2, 3], [1, 4, 9])
plt.show()

The object-oriented approach using fig, ax = plt.subplots(), which is recommended for more control and flexibility.

PYTHON

# Create a figure and axis
fig, ax = plt.subplots()

# Plot on the axis
ax.plot([1, 2, 3], [1, 4, 9])

# Show the plot
plt.show()

When you use plt.subplots(), you get access to the figure and axes objects, allowing you to customize everything from the title, labels, grid, axis limits, and more, in a very controlled manner. This is the approach we will use in this episode.

Customizing Plots

In this example, we change the line style and color, add a title, axis labels and legend:

PYTHON

fig, ax = plt.subplots()

# Plotting data
ax.plot([1, 2, 3], [1, 4, 9], linestyle='--', color='r', label="y = x^2")

# Adding title and labels
ax.set_title('Simple Line Plot')
ax.set_xlabel('X Axis')
ax.set_ylabel('Y Axis')

# Adding a legend
ax.legend()

# Show the plot
plt.show()

Line plot

Using the education dataset we worked on in the previous episode, write the lines of code to plot the share of Mathematics in the total instruction time of students aged 6 to 13 years old in Austria.

Use the education_subset.csv file.
The X axis will show the students’ age (age column).
The Y axis will show the share of Mathematics in the total instruction time of students.
Add the axis labels and a title.

Show me the solution

PYTHON

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

# Reading the data into a DataFrame
df = pd.read_csv(r"education_subset.csv")

# Selecting Austria and Mathematics only
df = df[(df['iso3']=='AUT')&(df['subject_label']=='Mathematics')]

# Changing the type of the age column to integer
df['age'] = df['age'].str.replace('Y', '').astype(int)

# Sorting values by age, as it will appear on the graph
df.sort_values(by=['age'], ascending=True, inplace=True)

# Creating figure and axes
fig, ax = plt.subplots()

# Plotting the value for each age
ax.plot(df['age'], df['value'])

# Adding axis labels and title
ax.set_title('Austria')
ax.set_xlabel('Students age')
ax.set_ylabel('Share of Mathematics in total instruction time')

plt.show()

Types of plot

On top of the line plot that we already created using ax.plot(), matplotlib offers many other types of plots, including:

Bar plot

PYTHON

# Data
categories = ['A', 'B', 'C', 'D']
values = [3, 7, 5, 2]

# Bar plot
fig, ax = plt.subplots()
ax.bar(categories, values)

plt.show()

Horizontal bar plot

PYTHON

# Horizontal bar plot
fig, ax = plt.subplots()
ax.barh(categories, values)

plt.show()

Stacked bar plot

PYTHON

# Data
categories = ['A', 'B', 'C']
values1 = [3, 7, 5]
values2 = [2, 5, 6]

# Stacked bar plot
fig, ax = plt.subplots()
ax.bar(categories, values1, label='Category 1')
ax.bar(categories, values2, bottom=values1, label='Category 2')

ax.legend()

plt.show()

Scatter plot

PYTHON

# Data
x = [1, 2, 3, 4, 5]
y = [1, 4, 9, 16, 25]

# Scatter plot
fig, ax = plt.subplots()
ax.scatter(x, y)

plt.show()

Pie charts

PYTHON

# Data
sizes = [10, 20, 30, 40]
labels = ['A', 'B', 'C', 'D']

# Pie chart
fig, ax = plt.subplots()
ax.pie(sizes, labels=labels)

plt.show()

Multiple Subplots

Using fig, ax = plt.subplots() allows you to create multiple subplots within a single figure.

PYTHON

fig, ax = plt.subplots(2, 1)

# Plotting in the first subplot
ax[0].plot([1, 2, 3], [1, 4, 9])
ax[0].set_title('First Plot')

# Plotting in the second subplot
ax[1].bar(['A', 'B', 'C'], [3, 7, 5])
ax[1].set_title('Second Plot')

plt.show()

Saving plots to files

PYTHON

fig, ax = plt.subplots()

# Plotting
ax.plot([1, 2, 3], [1, 4, 9])

# Saving the plot
fig.savefig('plot.png')

Matplotlib supports multiple formats, including PNG, PDF, SVG, and more. Use fig.savefig('filename.format').

Multiple subplots

Using the education dataset we worked on in the previous episode, create a figure with 2 subplots (bar plots), showing the share of Mathematics in the total instruction time of students aged 6 and 13 years old in all countries (one bar per country) where data is available.

Use the education_subset.csv file.
The X axis will show the country codes (iso3 column).
The Y axis will show the share of Mathematics in the total instruction time of students (value column).
The first subplot will show values for 6 year old students, the second subplot will show values for 13 year old students
Add the axis labels and a title to each subplot.

Show me the solution

PYTHON

# Reading the data into a DataFrame
df = pd.read_csv(r"education_subset.csv")

# Selecting Mathematics
df = df[df['subject_label']=='Mathematics']

# Sorting values by country code, as it will appear on the graph
df.sort_values(by=['iso3'], ascending=True, inplace=True)

# Create 2 dataframes, one for 6 years old, one for 13 years old
df6 = df[df['age']=='Y6']
df13 = df[df['age']=='Y13']

# Creating figure and axes
fig, ax = plt.subplots(nrows=2, sharey=True, figsize=(10, 10))

# Plotting the bars in each axes
ax[0].bar(df6['iso3'], df6['value'])
ax[1].bar(df13['iso3'], df13['value'])

# Adding axis labels
ax[0].set_ylabel('Share of Mathematics\nin total instruction time')

# Rotate the country codes
for ax_i in ax: 
    ax_i.tick_params(axis='x', labelrotation = 45)
    
# Add a title to each axes
ax[0].set_title('6 years old')
ax[1].set_title('13 years old')

plt.show()

Quick and appealing plots with Seaborn

Both Seaborn and Matplotlib are popular Python libraries for data visualization, but they serve different purposes:

High-Level vs. Low-Level

Matplotlib is a low-level library, offering full control over plots but requiring more code.
Seaborn is a high-level library built on top of Matplotlib, offering easier and quicker creation of visually appealing statistical plots.

Aesthetics

Matplotlib produces basic plots by default, requiring manual styling for better visuals.
Seaborn comes with default styles, making it easier to create polished plots.

Plotting Types

Matplotlib supports a wide range of plots but requires extra work for advanced statistical plots.
Seaborn offers specialized plots for statistics (e.g., violin plots, pair plots) with minimal effort.

Integration with Pandas

Matplotlib doesn’t integrate directly with Pandas DataFrames.
Seaborn integrates smoothly with Pandas, allowing you to pass DataFrame columns directly into plotting functions.

However, Seaborn is built on top of Matplotlib, meaning that it inherits Matplotlib’s flexibility, allowing users to make more detailed customizations as needed.

Seaborn includes several built-in datasets, such as tips, iris, and flights, that we will use throughout this lesson for our examples. These datasets are great for experimenting with Seaborn’s plotting functions without needing to import external data files.

Creating Seaborn plots

With Seaborn too, the plt.subplots() function is used to create a figure (fig) and one or more axes (ax) that can be used to draw plots.

PYTHON

import seaborn as sns
import matplotlib.pyplot as plt

# Create the figure and axes
fig, ax = plt.subplots()

# Use Seaborn to create a plot on the axes
sns.set(style="whitegrid")
data = sns.load_dataset("tips")
sns.boxplot(x="day", y="total_bill", data=data, ax=ax)

# Customize with Matplotlib
ax.set_title("Total Bill by Day")
ax.set_xlabel("Day of the Week")
ax.set_ylabel("Total Bill ($)")

# Show the plot
plt.show()

In this example, we create a boxplot using Seaborn, but we specify the axes (ax) created with plt.subplots(). This allows us to use Matplotlib to customize the plot’s title, labels, and size.

The other types of graphs available in Seaborn include:

Distribution plots

PYTHON

fig, ax = plt.subplots(figsize=(8, 6))
sns.histplot(data["total_bill"], ax=ax)
ax.set_title("Histogram of Total Bill")
plt.show()

Scatter plots

PYTHON

fig, ax = plt.subplots(figsize=(8, 6))
sns.scatterplot(x="total_bill", y="tip", data=data, ax=ax)
ax.set_title("Total Bill vs Tip")
plt.show()

Seaborn’s scatterplot() function allows you to add additional features like color and size based on other variables.

Line plots

PYTHON

fig, ax = plt.subplots(figsize=(8, 6))
sns.lineplot(x="size", y="total_bill", data=data, ax=ax)
ax.set_title("Total Bill vs Size of Party")
plt.show()

Line plot

Using the education dataset we worked on in the previous episode, create a lineplot, showing the share of each subject in the total instruction time of students in Austria over time.

Use the education_subset.csv file.
The X axis will show the students’ age (age column).
The Y axis will show the share of each subject in the total instruction time of students (value column).
The color of the lines will indicate the subject (subject_label column).

Show me the solution

PYTHON

# Reading the data into a DataFrame
df = pd.read_csv(r"education_subset.csv")

# Selecting Austria
df = df[df['iso3']=='AUT']

# Changing the type of the age column to integer
df['age'] = df['age'].str.replace('Y', '').astype(int)

# Creating figure and axes
fig, ax = plt.subplots()

# Plot the lines
sns.lineplot(data=df, x='age', y='value', hue='subject_label', ax=ax)

# Adding axis labels and title
ax.set_title('Austria')
ax.set_xlabel('Students age')
ax.set_ylabel('Share of Mathematics in total instruction time')

# Change the default legend title
ax.legend(title='Subject')

plt.show()

Line plot with a loop

Now let’s do the same as before, but now we want one graph for each country where data is available.

Show me the solution

PYTHON

# Reading the data into a DataFrame
df = pd.read_csv(r"education_subset.csv")

# Changing the type of the age column to integer
df['age'] = df['age'].str.replace('Y', '').astype(int)


for iso3 in df.iso3.unique(): # For loop over the iso3 codes

    # Select the country
    df_iso = df[df['iso3']==iso3]    

    # Creating figure and axes
    fig, ax = plt.subplots()
    
    # Plot the lines
    sns.lineplot(data=df_iso, x='age', y='value', hue='subject_label', ax=ax)
    
    # Adding axis labels and title
    ax.set_title(f'{iso3}') # Using a f-string for using the iso3 variable in the string
    ax.set_xlabel('Students age')
    ax.set_ylabel('Share of Mathematics in total instruction time')
    
    # Change the default legend title
    ax.legend(title='Subject')
    
    # Saving the plot
    fig.savefig(f'{iso3}.png')

Customizing Seaborn plots

One of the benefits of using Seaborn is its simplicity and attractive default themes, but you can still customize the appearance of plots using Matplotlib functions. For example, you can set the plot’s title, labels, or customize the grid.

PYTHON

# Customizing plot appearance
fig, ax = plt.subplots(figsize=(8, 6))
sns.boxplot(x="day", y="total_bill", data=data, ax=ax)
ax.set_title("Total Bill by Day")
ax.set_xlabel("Day of the Week")
ax.set_ylabel("Total Bill ($)")
ax.grid(True)  # Add gridlines
plt.show()

Interactive plots with Plotly

Plotly is an interactive graphing library that enables the creation of sophisticated visualizations that are interactive by default. Unlike static libraries like Matplotlib and Seaborn, Plotly allows you to zoom, pan, and hover over data points to inspect values directly in the plot.

Advantages of Plotly include:

Interactive Plots: Plots in Plotly are interactive out of the box, making them ideal for exploring data.
Web Integration: Plotly graphs can easily be integrated into web applications, such as Dash.
High-quality Visualizations: Plotly can generate a wide range of high-quality, aesthetically appealing plots.

Importing `plotly`

Plotly’s most commonly used module for creating visualizations is plotly.express. Here’s how to import it:

PYTHON

import plotly.express as px

plotly.graph_objects is another module in Plotly that provides more flexibility for creating complex visualizations. However, we will primarily focus on plotly.express as it simplifies the syntax for most common plots.

Creating Simple Plots with Plotly Express

Line Plot

PYTHON

import pandas as pd

# Sample data

data = pd.DataFrame({
"Date": pd.date_range(start="2024-01-01", periods=10, freq="D"),
"Value": [10, 12, 13, 15, 16, 18, 19, 20, 21, 22]
})

# Create a line plot

fig = px.line(data, x="Date", y="Value", title="Simple Line Plot")
fig.write_html("plot.html")

In this code:

px.line() creates the line plot.
The x and y arguments specify which columns to plot.
fig.write_html() saved the plot as a HTML file.

Scatter Plot

PYTHON

import numpy as np

# Sample data

data = pd.DataFrame({
"X": np.random.rand(100),
"Y": np.random.rand(100)
})

# Create a scatter plot

fig = px.scatter(data, x="X", y="Y", title="Scatter Plot")
fig.write_html("plot.html")

Bar Chart

PYTHON

categories = ['A', 'B', 'C', 'D', 'E']
values = [23, 45, 56, 78, 33]
data = pd.DataFrame({"Category": categories, "Value": values})

# Create a bar chart

fig = px.bar(data, x="Category", y="Value", title="Bar Chart")
fig.write_html("plot.html")

Pie Chart

PYTHON

# Sample data for Pie Chart

categories = ['Red', 'Blue', 'Green']
values = [25, 50, 25]
data = pd.DataFrame({"Category": categories, "Value": values})

# Create a pie chart

fig = px.pie(data, names="Category", values="Value", title="Pie Chart")
fig.write_html("plot.html")

Customizing Plots in Plotly Express

Plotly Express automatically makes plots interactive, but you can also customize your plots to make them more informative and visually appealing.

Adding Titles and Labels

You can modify the title and axis labels of your plot:

PYTHON

fig.update_layout(
title="Updated Plot Title",
xaxis_title="Custom X Axis Label",
yaxis_title="Custom Y Axis Label"
)
fig.write_html("plot.html")

Changing Colors

You can change the color of data points or bars based on a categorical variable:

PYTHON

# Adding a color dimension

data['Color'] = np.random.choice(['Red', 'Blue', 'Green'], size=100)

fig = px.scatter(data, x="X", y="Y", color="Color", title="Colored Scatter Plot")
fig.write_html("plot.html")

In this example, the color argument differentiates data points by color based on the Color column.

Plotly Express vs. Plotly Graph Objects

While plotly.express is great for creating quick, simple plots, there are cases when you might need more control over the plot’s components. This is where plotly.graph_objects comes in.

Plotly Graph Objects (go) is a lower-level interface that gives you finer control over the layout and elements of your plot. With go, you can manually define traces (such as lines, bars, and scatter plots), customize plot attributes, and handle more complex visualizations.

When to use plotly.graph_objects:

Multiple Traces: When you need to add different types of plots (like a line and scatter plot) in the same figure.
Advanced Customization: For precise control over each plot element (e.g., customizing legends, adding annotations).
Complex Layouts: When you need subplots or advanced arrangements of figures.

For example, if you wanted to combine a line and scatter plot on the same figure, you would use plotly.graph_objects:

PYTHON

import plotly.graph_objects as go

# Create a figure with both a scatter and line trace

fig = go.Figure()

# Scatter plot trace

fig.add_trace(go.Scatter(x=data["X"], y=data["Y"], mode='markers', name="Scatter"))

# Line plot trace

fig.add_trace(go.Scatter(x=data["X"], y=data["Y"], mode='lines', name="Line"))

# Save the plot

fig.write_html("plot.html")

While plotly.express handles this type of task easily with fewer lines of code, go offers more flexibility for complex customizations.

Interactive Features of Plotly Express

Plotly plots are interactive by default. These features include:

Zooming and Panning: Users can zoom into a region of the plot by dragging the mouse, and pan across it.
Hovering: When you hover over data points, Plotly shows additional information (e.g., exact values).
Saving and Exporting: You can save your plot as an image or an interactive HTML file.

Visualizations

Overview

Questions

Objectives

Introduction

PYTHON

Basic plots with matplotlib

Creating a figure and axes

2 ways to create a matplotlib plot

PYTHON

PYTHON

Customizing Plots

PYTHON

Line plot

Show me the solution

PYTHON

Types of plot

PYTHON

PYTHON

PYTHON

PYTHON

PYTHON

Multiple Subplots

PYTHON

Saving plots to files

PYTHON

Multiple subplots

Show me the solution

PYTHON

Quick and appealing plots with Seaborn

Creating Seaborn plots

PYTHON

PYTHON

PYTHON

PYTHON

Line plot

Show me the solution

PYTHON

Line plot with a loop

Show me the solution

PYTHON

Customizing Seaborn plots

PYTHON

Interactive plots with Plotly

Importing plotly

PYTHON

Creating Simple Plots with Plotly Express

Line Plot

PYTHON

Scatter Plot

PYTHON

Bar Chart

PYTHON

Pie Chart

PYTHON

Customizing Plots in Plotly Express

Adding Titles and Labels

PYTHON

Changing Colors

PYTHON

Plotly Express vs. Plotly Graph Objects

PYTHON

Interactive Features of Plotly Express

Basic plots with `matplotlib`

Importing `plotly`