20 Seaborn concepts with Before-and-After Examples

1. Line Plot (seaborn.lineplot) 📉

Boilerplate Code:

import seaborn as sns

Use Case: Create a line plot to display trends over time or ordered data. 📉

Goal: Visualize continuous data over a range or time. 🎯

Sample Code:

# Example data
data = [1, 2, 3, 4, 5]
sns.lineplot(x=[1, 2, 3, 4, 5], y=[2, 4, 6, 8, 10])

Before Example: You have a sequence of data points but no clear trend visualization. 🤔

Data: x = [1, 2, 3, 4, 5], y = [2, 4, 6, 8, 10]

After Example: With lineplot, the trend of the data is clear! 📉

Output: A line plot connecting the points.

Challenge: 🌟 Try adding markers with markers=True to make each data point stand out.

---

2. Bar Plot (seaborn.barplot) 📊

Boilerplate Code:

import seaborn as sns

Use Case: Create a bar plot to compare categorical data. 📊

Goal: Compare means or other summary statistics across categories. 🎯

Sample Code:

# Example data
sns.barplot(x=['A', 'B', 'C'], y=[10, 20, 30])

Before Example: You have categories and values but no visual way to compare them. 🤔

Data: x = ['A', 'B', 'C'], y = [10, 20, 30]

After Example: With barplot, you can compare the values across categories! 📊

Output: A bar chart showing categorical comparisons.

Challenge: 🌟 Try adding error bars using the ci parameter to represent uncertainty.

---

3. Count Plot (seaborn.countplot) 📊

Boilerplate Code:

import seaborn as sns

Use Case: Create a count plot to show the count of observations in each category. 📊

Goal: Visualize the frequency of each category. 🎯

Sample Code:

# Example data
sns.countplot(x=['cat', 'dog', 'dog', 'cat', 'bird'])

Before Example: You have categorical data but no way to visualize how often each category appears. 🤔

Data: ['cat', 'dog', 'dog', 'cat', 'bird']

After Example: With countplot, you see how frequently each category appears! 📊

Output: A bar plot showing the count of each category.

Challenge: 🌟 Try using hue to split the counts by an additional categorical variable.

---

4. Histogram (seaborn.histplot) 📈

Boilerplate Code:

import seaborn as sns

Use Case: Create a histogram to visualize the distribution of numerical data. 📈

Goal: Display the frequency of data points within bins. 🎯

Sample Code:

# Example data
sns.histplot(data=[1, 2, 2, 3, 3, 4, 5], bins=5)

Before Example: You have numerical data but no way to visualize its distribution. 🤔

Data: [1, 2, 2, 3, 3, 4, 5]

After Example: With histplot, the data distribution becomes clear! 📈

Output: A histogram showing data frequencies.

Challenge: 🌟 Try changing the number of bins to see how it affects the histogram.

---

5. Box Plot (seaborn.boxplot) 📦

Boilerplate Code:

import seaborn as sns

Use Case: Create a box plot to show the distribution of data and detect outliers. 📦

Goal: Summarize the data distribution using quartiles and outliers. 🎯

Sample Code:

# Example data
sns.boxplot(data=[1, 2, 3, 4, 5, 6, 7, 8, 9])

Before Example: You have numerical data but no way to visualize its range, quartiles, and outliers. 🤔

Data: [1, 2, 3, 4, 5, 6, 7, 8, 9]

After Example: With boxplot, you can see the distribution and detect any outliers! 📦

Output: A box plot summarizing the data.

Challenge: 🌟 Try splitting the box plots by categories using x and y arguments.

---

6. Violin Plot (seaborn.violinplot) 🎻

Boilerplate Code:

import seaborn as sns

Use Case: Create a violin plot to show the distribution of data and its probability density. 🎻

Goal: Combine the benefits of box plots and density plots. 🎯

Sample Code:

# Example data
sns.violinplot(data=[1, 2, 3, 4, 5, 6, 7, 8, 9])

Before Example: You have numerical data but want a more detailed look at its distribution than just a box plot. 🤔

Data: [1, 2, 3, 4, 5, 6, 7, 8, 9]

After Example: With violinplot, you visualize both the distribution and probability density! 🎻

Output: A violin plot showing data density and distribution.

Challenge: 🌟 Try splitting violins by categories to compare distributions between groups.

---

7. Scatter Plot (seaborn.scatterplot) 🔵

Boilerplate Code:

import seaborn as sns

Use Case: Create a scatter plot to show the relationship between two numerical variables. 🔵

Goal: Visualize individual data points to detect patterns or correlations. 🎯

Sample Code:

# Example data
sns.scatterplot(x=[1, 2, 3, 4], y=[10, 11, 12, 13])

Before Example: You have two variables but no way to visualize their relationship. 🤔

Data: x = [1, 2, 3, 4], y = [10, 11, 12, 13]

After Example: With scatterplot, the relationship between the variables is visualized! 🔵

Output: A scatter plot showing individual data points.

Challenge: 🌟 Try adding a third dimension by mapping a categorical variable to hue.

---

8. Pair Plot (seaborn.pairplot) 🔀

Boilerplate Code:

import seaborn as sns

Use Case: Create a pair plot to visualize relationships between multiple pairs of variables. 🔀

Goal: Show scatter plots for each pair of variables in a dataset. 🎯

Sample Code:

# Example data
sns.pairplot(sns.load_dataset("iris"))

Before Example: You have multiple variables but no way to visualize their relationships in one view. 🤔

Data: Iris dataset (sepal length, sepal width, petal length, petal width)

After Example: With pairplot, you can view scatter plots for each pair of variables! 🔀

Output: A grid of scatter plots for each pair of variables.

Challenge: 🌟 Try using hue to differentiate categories in the dataset.

---

9. Heatmap (seaborn.heatmap) 🔥

Boilerplate Code:

import seaborn as sns

Use Case: Create a heatmap to visualize data intensity in a matrix. 🔥

Goal: Represent numerical data as color-coded intensity. 🎯

Sample Code:

# Example data
import numpy as np
data = np.array([[1, 2], [3, 4]])
sns.heatmap(data)

Before Example: You have a matrix of numerical data but no clear way to visualize the intensity. 🤔

Data: [[1, 2], [3, 4]]

After Example: With heatmap, you visualize the data intensity using colors! 🔥

Output: A heatmap representing the matrix.

Challenge: 🌟 Try using annot=True to show the data values inside the heatmap cells.

---

10. Joint Plot (seaborn.jointplot) 🌐

Boilerplate Code:

import seaborn as sns

Use Case: Create a joint plot to visualize both the distribution and relationship of two variables. 🌐

Goal: Combine scatter plots and distribution plots into one view. 🎯

Sample Code:

# Example data
sns.jointplot(x=[1, 2, 3], y=[4, 5, 6])

Before Example: You have two variables but no way to visualize both their relationship and distribution. 🤔

Data: x = [1, 2, 3], y = [4, 5, 6]

After Example: With jointplot, you can visualize both the scatter plot and histograms in one! 🌐

Output: A combined scatter and histogram plot.

Challenge: 🌟 Try using different kind options like reg for regression or hex for hexbin plots.

---

11. Regression Plot (seaborn.regplot) 📈

Boilerplate Code:

import seaborn as sns

Use Case: Create a regression plot to display a linear regression line with data points. 📈

Goal: Visualize the relationship between two variables with a regression line. 🎯

Sample Code:

# Example data
sns.regplot(x=[1, 2, 3, 4], y=[5, 6, 7, 8])

Before Example: You have two variables but no way to represent their linear relationship. 🤔

Data: x = [1, 2, 3, 4], y = [5, 6, 7, 8]

After Example: With regplot, you visualize the relationship with a linear regression line! 📈

Output: A scatter plot with a regression line.

Challenge: 🌟 Try adding ci=None to remove the confidence interval or change the order for polynomial regression.

---

12. KDE Plot (seaborn.kdeplot) 📊

Boilerplate Code:

import seaborn as sns

Use Case: Create a Kernel Density Estimate (KDE) plot to visualize the distribution of data. 📊

Goal: Show the probability density function of a variable. 🎯

Sample Code:

# Example data
sns.kdeplot(data=[1, 2, 2, 3, 4, 5])

Before Example: You have numerical data but no smooth representation of its distribution. 🤔

Data: [1, 2, 2, 3, 4, 5]

After Example: With kdeplot, you see a smooth curve representing the data's distribution! 📊

Output: A KDE plot showing the probability density.

Challenge: 🌟 Try combining kdeplot with histplot for a detailed view of the distribution.

---

13. Facet Grid (seaborn.FacetGrid) 📐

Boilerplate Code:

import seaborn as sns

Use Case: Create a facet grid to plot multiple plots based on subsets of data. 📐

Goal: Display the relationship between variables across different subsets of data. 🎯

Sample Code:

# Load example data
df = sns.load_dataset("tips")

# Create FacetGrid
g = sns.FacetGrid(df, col="sex", row="time")
g.map(sns.scatterplot, "total_bill", "tip")

Before Example: You have data split by categories but no way to visualize each subset clearly. 🤔

Data: Tips dataset with sex and time categories.

After Example: With FacetGrid, you get multiple plots for each subset of the data! 📐

Output: A grid of scatter plots showing tips by total bill.

Challenge: 🌟 Try using hue to add another categorical dimension to the grid.

---

14. Swarm Plot (seaborn.swarmplot) 🐝

Boilerplate Code:

import seaborn as sns

Use Case: Create a swarm plot to display categorical scatter plots with non-overlapping points. 🐝

Goal: Show the distribution of data points within categories without overlapping. 🎯

Sample Code:

# Example data
sns.swarmplot(x=["cat", "cat", "dog", "dog"], y=[2, 3, 4, 5])

Before Example: You have categorical data but no way to show individual data points clearly. 🤔

Data: Categories: ["cat", "dog"], Values: [2, 3, 4, 5]

After Example: With swarmplot, you display individual data points without overlap! 🐝

Output: A swarm plot showing the distribution of points.

Challenge: 🌟 Try using hue to color-code points based on an additional variable.

---

15. Point Plot (seaborn.pointplot) 📍

Boilerplate Code:

import seaborn as sns

Use Case: Create a point plot to visualize the relationship between categorical data and numerical data. 📍

Goal: Show mean values of categories with confidence intervals. 🎯

Sample Code:

# Example data
sns.pointplot(x=["A", "B", "C"], y=[4, 5, 6])

Before Example: You have categorical data but no way to visualize the mean values across categories. 🤔

Data: Categories: ["A", "B", "C"], Values: [4, 5, 6]

After Example: With pointplot, the mean values are visualized with error bars! 📍

Output: A point plot showing means with error bars.

Challenge: 🌟 Try adding hue to compare means across different categories.

---

16. LM Plot (seaborn.lmplot) 🧮

Boilerplate Code:

import seaborn as sns

Use Case: Create an lmplot to visualize linear relationships with multiple facets. 🧮

Goal: Combine regression lines and scatter plots with facet grids. 🎯

Sample Code:

# Load example data
df = sns.load_dataset("tips")

# Create lmplot
sns.lmplot(x="total_bill", y="tip", hue="sex", data=df)

Before Example: You want to visualize both scatter plots and linear regression with multiple subsets of data. 🤔

Data: Tips dataset with total bill and tip variables.

After Example: With lmplot, the linear regression lines and scatter plots are displayed! 🧮

Output: A scatter plot with regression lines based on gender.

Challenge: 🌟 Try using col to split the plot by time of day.

---

17. Strip Plot (seaborn.stripplot) 🟡

Boilerplate Code:

import seaborn as sns

Use Case: Create a strip plot to display individual data points with some jitter. 🟡

Goal: Show the distribution of data points along a categorical axis with slight jitter. 🎯

Sample Code:

# Example data
sns.stripplot(x=["cat", "dog", "dog", "cat"], y=[2, 3, 4, 5])

Before Example: You have categorical data and want to display individual points, but they overlap. 🤔

Data: Categories: ["cat", "dog"], Values: [2, 3, 4, 5]

After Example: With stripplot, you can add jitter to make individual points clearer! 🟡

Output: A strip plot with jittered data points.

Challenge: 🌟 Try using hue to differentiate points based on another category.

---

18. PairGrid (seaborn.PairGrid) 🔗

Boilerplate Code:

import seaborn as sns

Use Case: Create a pair grid to visualize pairwise relationships between variables using different plots. 🔗

Goal: Combine multiple plot types (scatter, histograms, etc.) into a grid for deeper analysis. 🎯

Sample Code:

# Load example data
df = sns.load_dataset("iris")

# Create PairGrid
g = sns.PairGrid(df)
g.map_diag(sns.histplot)
g.map_offdiag(sns.scatterplot)

Before Example: You have multiple variables and want to use different plot types to explore their relationships. 🤔

Data: Iris dataset with multiple numerical variables.

After Example: With PairGrid, you combine scatter plots, histograms, and more in one view! 🔗

Output: A grid of plots for each pair of variables.

Challenge: 🌟 Try mapping different plot types to diagonal and off-diagonal elements for a more detailed analysis.

---

19. Joint Grid (seaborn.JointGrid) ⚙️

Boilerplate Code:

import seaborn as sns

Use Case: Create a joint grid to visualize the relationship between two variables using multiple plots. ⚙️

Goal: Create a custom plot that combines scatter plots, KDE, or histograms. 🎯

Sample Code:

# Example data
g = sns.JointGrid(x=[1, 2, 3], y=[4, 5, 6])
g.plot(sns

.scatterplot, sns.histplot)

Before Example: You want to customize the combination of scatter and distribution plots. 🤔

Data: x = [1, 2, 3], y = [4, 5, 6]

After Example: With JointGrid, you customize how different plots are displayed together! ⚙️

Output: A combined scatter plot and histogram.

Challenge: 🌟 Try using plot_kws to add custom styles to the plots.

---

20. Heatmap with Annotations (seaborn.heatmap) 🔢

Boilerplate Code:

import seaborn as sns

Use Case: Create a heatmap with annotations to visualize a matrix of numbers with text annotations. 🔢

Goal: Display numerical data as color-coded intensity and show the exact values. 🎯

Sample Code:

# Example data
import numpy as np
data = np.array([[1, 2], [3, 4]])
sns.heatmap(data, annot=True)

Before Example: You have a matrix of numbers but want both a color map and the exact values. 🤔

Data: [[1, 2], [3, 4]]

After Example: With heatmap and annotations, the data is color-coded and labeled! 🔢

Output: A heatmap showing data values with annotations.

Challenge: 🌟 Try changing the fmt argument to display different formats for annotations.

---

Seaborn documentation

https://seaborn.pydata.org/generated/seaborn.FacetGrid.html