Seaborn

1. Introduction to Seaborn

1.1 📊 What is Seaborn?

Seaborn is a Python library for statistical data visualization. It is built on Matplotlib and provides a higher-level interface for creating informative statistical graphics with less code. Seaborn supports charts such as: line plots, bar plots, scatter plots, histograms, box plots, heatmaps, count plots, regression plots, multi-plot grids. Install Seaborn using: pip install seaborn Import it using: import seaborn as sns 📌 Seaborn simplifies statistical visualization while still allowing Matplotlib customization.

1.2 Seaborn and Matplotlib Together

Seaborn creates the main visualization, while Matplotlib can customize elements such as: titles, axis labels, axis limits, figure size, legends, grid lines, saving. Example: 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 ) ax.set_title("Relationship Between Total Bill and Tip") ax.set_xlabel("Total Bill") ax.set_ylabel("Tip") plt.show() 📌 We can create a chart with Seaborn and customize it using Matplotlib.

1.3 About sns.load_dataset()

Seaborn provides sample datasets for learning and demonstration. Example: df = sns.load_dataset("tips") The tips dataset includes information such as: total bill, tip amount, day, time, customer-related categories. 📌 In real projects, we usually load our own DataFrame using Pandas.

2. Axes-Level and Figure-Level Functions

2.1 🧩 Axes-Level Functions

Axes-level functions draw one plot on a Matplotlib Axes. Examples: sns.scatterplot(), sns.lineplot(), sns.barplot(), sns.histplot(), sns.boxplot(), sns.countplot(), sns.heatmap(). We can pass an Axes using: ax=ax Example: fig, ax = plt.subplots() sns.scatterplot( data=df, x="total_bill", y="tip", ax=ax ) Axes-level functions plot on one Matplotlib Axes and return that Axes.

2.2 Figure-Level Functions

Figure-level functions manage their own Figure and usually return a Seaborn grid object. Examples: sns.relplot(), sns.catplot(), sns.displot(), sns.lmplot(), sns.jointplot(), sns.pairplot(). They are useful for creating multiple related plots using parameters such as: row, col, col_wrap. Figure-level functions normally manage the Figure through objects such as FacetGrid, PairGrid, or JointGrid. 📌 Axes-level functions create one plot. Figure-level functions can manage an entire group of plots.

3. Line Plot in Seaborn

3.1 📈 What is a Line Plot?

A line plot is used to show trends over time or another continuous sequence. It is created using: sns.lineplot() Common parameters include: x, y, hue, style, size, marker, linewidth, linestyle.

3.2 What is hue?

The hue parameter uses color to represent another variable. Example: hue="event" Each event category will be shown using a different color. Color, size, and style can represent additional variables in relational plots.

3.3 Line Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("fmri") fig, ax = plt.subplots() sns.lineplot( data=df, x="timepoint", y="signal", hue="event", ax=ax ) ax.set_title("Signal Over Time") ax.set_xlabel("Time Point") ax.set_ylabel("Signal") plt.show() 📌 Use hue when you want to compare multiple groups on the same chart.

4. Bar Plot in Seaborn

4.1 📊 What is a Bar Plot?

A bar plot compares a numeric measurement across categories. It is created using: sns.barplot() By default, sns.barplot() calculates the mean of the numeric variable for each category. 📌 A Seaborn bar plot summarizes observations. It does not simply draw every original numeric value.

4.2 Important Parameters

estimator: Determines how the bar height is calculated. Default: estimator="mean" Example using sum: estimator="sum" hue: Creates separate bars using another categorical variable. order: Controls the order of categories. order=["Thur", "Fri", "Sat", "Sun"] hue_order: Controls the order of the hue categories. orient: Controls bar orientation. orient="v" or: orient="h" errorbar: Controls the uncertainty or variability displayed around the estimate.

4.3 Bar Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") fig, ax = plt.subplots() sns.barplot( data=df, x="day", y="total_bill", hue="sex", errorbar="sd", ax=ax ) ax.set_title("Average Total Bill by Day") ax.set_xlabel("Day") ax.set_ylabel("Average Total Bill") plt.show() 📌 Each bar represents a summary value for one category.

5. Scatter Plot in Seaborn

5.1 🔵 What is a Scatter Plot?

A scatter plot shows the relationship between two numeric variables. It is created using: sns.scatterplot() A scatter plot represents each observation as a point and helps reveal possible relationships, clusters, and unusual observations.

5.2 Common Parameters

hue changes point color. size changes point size. style changes marker shape. alpha changes transparency. edgecolor changes point-border color. linewidth changes point-border thickness.

5.3 Scatter Plot Example

Example: 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", hue="sex", style="sex", ax=ax ) ax.set_title("Relationship Between Total Bill and Tip") ax.set_xlabel("Total Bill") ax.set_ylabel("Tip") plt.show() 📌 Using both hue and style can make categories easier to distinguish.

6. Box Plot in Seaborn

6.1 📦 What is a Box Plot?

A box plot shows the distribution of a numeric variable across categories. It displays: median, quartiles, spread, possible outliers. It is created using: sns.boxplot()

6.2 Common Parameters

hue, order, hue_order, width, fliersize, linewidth, showcaps, boxprops, medianprops.

6.3 Box Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") fig, ax = plt.subplots() sns.boxplot( data=df, x="day", y="total_bill", ax=ax ) ax.set_title("Total Bill Distribution by Day") ax.set_xlabel("Day") ax.set_ylabel("Total Bill") plt.show() 📌 Use a box plot to compare distributions and possible outliers across categories.

7. Histogram in Seaborn

7.1 📶 What is a Histogram?

A histogram shows the distribution of one numeric variable. It divides values into intervals called bins. It is created using: sns.histplot()

7.2 What is kde=True?

KDE means Kernel Density Estimate. It adds a smooth curve that estimates the shape of the distribution. kde=True 📌 The histogram shows frequency using bars. The KDE shows an estimated distribution using a smooth curve.

7.3 Histogram Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") fig, ax = plt.subplots() sns.histplot( data=df, x="total_bill", bins=30, kde=True, ax=ax ) ax.set_title("Distribution of Total Bills") ax.set_xlabel("Total Bill") ax.set_ylabel("Frequency") plt.show()

8. Heatmap in Seaborn

8.1 🌡️ What is a Heatmap?

A heatmap displays numeric values as colors in a matrix. It is created using: sns.heatmap() Heatmaps are useful for: correlation matrices, cross-tabulated values, activity by day and hour, values organized by rows and columns.

8.2 Common Parameters

annot=True displays values inside cells. fmt controls annotation format. cmap controls the color map. linewidths controls space between cells. linecolor controls boundary color. square=True creates square cells. annot_kws customizes annotation text.

8.3 Heatmap Example

Example: import pandas as pd import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") tabular_data = pd.crosstab( df["day"], df["sex"] ) fig, ax = plt.subplots() sns.heatmap( data=tabular_data, annot=True, fmt="d", cmap="Blues", ax=ax ) ax.set_title("Customer Count by Day and Category") plt.show() 📌 annot=True displays the values inside the cells.

9. Count Plot in Seaborn

9.1 🔢 What is a Count Plot?

A count plot shows the number of observations in each category. It is created using: sns.countplot() Unlike barplot(), a count plot does not require a numeric Y variable.

9.2 Count Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") fig, ax = plt.subplots() sns.countplot( data=df, x="day", hue="time", ax=ax ) ax.set_title("Number of Observations by Day") ax.set_xlabel("Day") ax.set_ylabel("Count") plt.show() 📌 Use a count plot when you want to count how often each category appears.

10. Figure-Level Plots in Seaborn

10.1 🖼️ What are Figure-Level Plots?

Figure-level plots manage their own Figure. Three main figure-level interfaces are:

Figure-Level Interfaces

Function	Purpose
`sns.relplot()`	Relationships
`sns.catplot()`	Categorical data
`sns.displot()`	Distributions

The three main figure-level interfaces in Seaborn

10.1 🖼️ What are Figure-Level Plots? (continued)

Each one provides access to several related axes-level plots using the kind parameter.

10.2 Creating Facets

Facets are multiple related plots created by splitting data into groups. Use: row, col, col_wrap Example: sns.relplot( data=df, x="total_bill", y="tip", col="day" ) This creates a separate plot for each day. 📌 Facets help compare the same relationship across different groups.

11. Relational Plot

11.1 🔗 What is relplot()?

relplot() is a figure-level function used to visualize statistical relationships. It supports: kind="scatter", kind="line" It combines a FacetGrid with either scatterplot() or lineplot().

11.2 Relational Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.relplot( data=df, x="total_bill", y="tip", kind="scatter", col="day", col_wrap=2 ) g.set_axis_labels("Total Bill", "Tip") plt.show() 📌 col="day" creates a separate plot for each day.

12. Customizing Figure-Level Plots

12.1 🎨 Using the Returned Grid Object

Figure-level functions return objects such as FacetGrid. g = sns.relplot(...) We can use g to customize the full collection of plots.

12.2 Customization Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.relplot( data=df, x="total_bill", y="tip", kind="scatter", row="day" ) g.figure.suptitle( "Tip Relationship by Day", y=1.02 ) g.set( xlim=(0, 60), ylim=(0, 12) ) g.set_axis_labels( "Total Bill", "Tip" ) plt.show() 📌 g.figure.suptitle() sets an overall title. 📌 g.set() applies settings across the facets. 📌 g.set_axis_labels() sets common axis labels.

13. Categorical Plot

13.1 📊 What is catplot()?

catplot() is a figure-level function for categorical visualizations. The kind parameter controls the chart type.

13.2 Common Categorical Plot Types

The kind parameter supports the following categorical plot types:

Common Categorical Plot Types

Kind	Description	Best Use
strip	Shows observations with optional jitter	Viewing individual observations across categories
swarm	Adjusts points to avoid overlap	Viewing individual observations clearly
box	Displays a box plot	Comparing distributions and outliers
violin	Displays distribution shape using density	Examining detailed distribution shape
bar	Displays an estimated value with uncertainty	Comparing summary values
point	Displays estimates as points connected by lines	Comparing changes in estimated values
count	Displays the number of observations	Comparing category frequencies

13.3 Categorical Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.catplot( data=df, x="day", y="total_bill", kind="violin", col="sex" ) g.set_axis_labels( "Day of Week", "Total Bill" ) g.set_titles( "Category: {col_name}" ) plt.show() 📌 Change kind to compare different categorical visualizations.

14. Distribution Plot

14.1 📶 What is displot()?

displot() is a figure-level function for visualizing distributions. Common kind values include: hist, kde, ecdf displot() is the figure-level interface for distribution plots and uses axes-level functions such as histplot() or kdeplot() internally.

14.2 Distribution Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.displot( data=df, x="total_bill", kind="hist", bins=30, kde=True ) g.figure.suptitle( "Distribution of Total Bills", y=1.02 ) g.set_axis_labels( "Total Bill", "Frequency" ) plt.show() 📌 displot() creates its own Figure, so we customize it through the returned grid object instead of an undefined ax.

15. Regression Plot with lmplot()

15.1 📉 What is lmplot()?

lmplot() displays: a scatter plot, a fitted regression line. It combines regression plotting with a grid that can create facets.

15.2 Regression Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.lmplot( data=df, x="total_bill", y="tip", col="day", col_wrap=2 ) g.figure.suptitle( "Relationship Between Total Bill and Tip", y=1.02 ) plt.show()

15.3 Polynomial Regression

The order parameter controls the polynomial order. Example: sns.lmplot( data=df, x="total_bill", y="tip", order=2 ) 📌 order=1 fits a straight line. order=2 fits a quadratic curve.

16. Joint Plot

16.1 🧩 What is jointplot()?

jointplot() displays: the relationship between two variables in the center, individual distributions along the margins. It returns a JointGrid and supports plot types such as scatter, KDE, histogram, hexagonal binning, regression, and residual plots.

16.2 Joint Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.jointplot( data=df, x="total_bill", y="tip", kind="scatter" ) g.figure.suptitle( "Total Bill and Tip", y=1.02 ) plt.show() 📌 The center shows the relationship. The margins show the separate distributions.

17. Pair Plot

17.1 🔲 What is pairplot()?

pairplot() shows pairwise relationships among multiple numeric variables. By default: diagonal plots show individual variable distributions, off-diagonal plots show relationships between pairs of variables. It returns a PairGrid for further customization.

17.2 Pair Plot Example

Example: import seaborn as sns import matplotlib.pyplot as plt df = sns.load_dataset("tips") g = sns.pairplot( data=df, vars=[ "total_bill", "tip", "size" ], hue="time" ) g.figure.suptitle( "Pairwise Relationships", y=1.02 ) plt.show() 📌 Use vars to limit the number of variables and keep the grid readable.

18. Changing Context and Color Palette

18.1 🔠 What is Context?

Context controls the scale of chart elements such as: font size, labels, lines, markers. Set context using: sns.set_context("context_name") Available contexts:

Available Contexts

Context	General Size
paper	Smallest
notebook	Default
talk	Larger
poster	Largest

18.1 🔠 What is Context? (continued)

Example: sns.set_context("talk")

18.2 What is a Color Palette?

A color palette controls the colors used in Seaborn charts. Set a palette using: sns.set_palette("palette_name") Common palettes include: deep, muted, bright, pastel, dark, colorblind. Example: sns.set_palette("colorblind")

18.3 Setting an Overall Theme

We can configure style, context, and palette together. Example: import seaborn as sns import matplotlib.pyplot as plt sns.set_theme( style="whitegrid", context="talk", palette="colorblind" ) df = sns.load_dataset("tips") fig, ax = plt.subplots() sns.scatterplot( data=df, x="total_bill", y="tip", hue="day", ax=ax ) ax.set_title("Total Bill and Tip") plt.show() 📌 Use style for the background and grid, context for element size, and palette for colors.