📊 Data Visualization: Seaborn Cheat Sheet

date

Apr 1, 2023

slug

data-visualization-seaborn-cheat-sheet

status

Published

-) Preparing the Libraries and Datasets

#
# ---- Importing the modules ----
#
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

sns.set_style('whitegrid')


#
# ---- Reading the datasets ----
#

spotify_data = pd.read_csv('./datas/spotify.csv'
                           , index_col='Date'
                           , parse_dates=True)

flight_data = pd.read_csv('./datas/flight_delays.csv'
                         , index_col='Month')

insurance_data = pd.read_csv('./datas/insurance.csv')

iris_data = pd.read_csv('./datas/iris.csv'
                      , index_col='Id')

0) Trend Plots

Trend Plots are used to show patterns of changes, especially over the time, such as hours, days, weeks, years and so on. Styles:

Line Plots

0.1) Line Plots

Line Plots are usefull to show changes of variables over the time. Commands:

All Variables

sns.lineplot(data=data)

Just a Few Variables

sns.lineplot(data=data1, label='label1') sns.lineplot(data=data2, label='label2)

#
# ---- Showing a line plot with all columns ----
#
plt.figure(figsize=(10,7))
plt.title('Songs on Spotify: Number of Views per Date')
plt.xlabel('Date')
plt.ylabel('Number of Views')

sns.lineplot(data=spotify_data)

plt.show()

#
# ---- Line Plot with just a few columns/variables ----
#
plt.figure(figsize=(10,7))

plt.title('Spotify Songs: Number of Views per Date')
plt.xlabel('Date')
plt.ylabel('Number of Views')

sns.lineplot(data=spotify_data['Despacito'], label='Despacito')
sns.lineplot(data=spotify_data['Something Just Like This'], label='Something Just Like This')
sns.lineplot(data=spotify_data['Unforgettable'], label='Unforgettable')

plt.show()

1) Relationship Plots

Relationship Plots are used to show comparisons/relationships between the dataset variables. Styles:

Bar Plot

HeatMap Plot

Scatter Plot

Regression Plot

Lm Plot

Swarm Plot

1.0) Bar Plots

Bar Plots compare quantities corresponding from different groups. Commands:

Simple Bar Plot

sns.barplot(data=data, x=x_data, y=y_data)

Bar Plot With Categorial Variable

sns.barplot(data=data, x=x_data, y=y_data, hue=categorical variable)

#
# ---- Simple Bar Plot ----
#
plt.figure(figsize=(10,7))
plt.title("Month Delays of HA's Flights")

# positive values >> the flight was delayed
# negative values >> the flight was early
sns.barplot(data=flight_data, x=flight_data.index, y='HA')

plt.show()

#
# ---- Bar Plot with Categorical Variable ----
#
plt.figure(figsize=(10,7))
plt.title('Charges by BMI and Smoker or Not People')


sns.barplot(data=insurance_data[0:20]
           , x='bmi'
           , y='charges'
           , hue='smoker')

plt.show()

1.1) HeatMap Plots

HeatMaps show patterns and correlations colouring the variables, only accept numerical variables and categorical variables must be labelled. Commands:

Simple HeatMap

sns.heatmap(data=data, annot=True|False)

#
# ---- Simple HeatMap ----
#
plt.figure(figsize=(10,7))

plt.title('Flights Delays per Month')
plt.xlabel('Flights')
plt.ylabel('Months')

sns.heatmap(data=flight_data, annot=True)

plt.show()

1.2) Scatter Plot

Scatter Plots show how the data are distributed in the dataset. Commands:

Simple Scatter Plot

sns.scatterplot(data=data, x=x_data, y=y_data)

Scatter Plot with Categorical Variable

sns.scatterplot(data=data, x=x_data, y=y_data, hue=categorical_variable)

#
# ---- Simple Scatter Plot ----
#
plt.figure(figsize=(10,7))

plt.title('Charges by BMI')
plt.xlabel('BMI')
plt.ylabel('Charges')

sns.scatterplot(data=insurance_data, x='bmi', y='charges')

plt.show()

#
# ---- Scatter Plot with Categorical Variable ----
#
plt.figure(figsize=(10,7))

plt.title('Charges by BMI and Smoker or Not People')
plt.xlabel('BMI')
plt.ylabel('Charges')

sns.scatterplot(data=insurance_data, x='bmi', y='charges', hue='smoker')

plt.show()

1.3) Regression Plots

Regression Plots are like the scatter ones, but with a regression line displaying the relationship between the x and y variables. Commands:

Regression Plot

sns.regplot(data=data, x=x_data, y=y_data)

#
# ---- Regression Plot ----
#
plt.figure(figsize=(10,7))

plt.title('Charges by BMI')
plt.xlabel('BMI')
plt.ylabel('Charges')

# as higher the BMI is, as more the person pays for charges
sns.regplot(data=insurance_data, x='bmi', y='charges')

plt.show()

1.4) LM Plots

LM Plots are like the regression ones, but with a categorical variable and a regression line for each categorical group. Commands:

LM Plot

sns.lmplot(data=data, x=x_data, y=y_data, hue=categorical_variable)

#
# ---- LM Plot ----
#
#
# we can assume that if the person is a smoker, they will pay
# for more charges as higher the BMI is.
#
# the same is valid for non-smokers, but the charge value is way
# to slow compared to the smoker ones
sns.lmplot(data=insurance_data, x='bmi', y='charges', hue='smoker');

1.5) Swarm Plots

Swarm Plots show the relationship between a categorical variable and a numerical one. Commands:

Swarm Plot

sns.swarmplot(data=data, x=x_data, y=y_data)

#
# ---- Swarm Plot ----
#
plt.figure(figsize=(10,7))

plt.title('Charges by Smoker or Non-Smoker People')
plt.xlabel('Smoker or Not')
plt.ylabel('Charges')

sns.swarmplot(data=insurance_data, x='smoker', y='charges')

plt.show()

2) Distribution Plots

Distribution Plots show how the datas are distributed into the dataset and help the Data Scientist takes guesses of the possible output values for specific inputs. Styles:

Histogram Plots

KDE Plots

Joint Plots

2.0) Histogram Plots

Histogram Plots show the distribution of a single a variable with or without a categorical one. Commands:

Simple Hist Plot

sns.histplot(data=data, x=x_data)

Hist Plot with Categorical Variable

sns.histplot(data=data, x=x_data, hue=categorical_variable)

#
# ---- Simple Hist Plot ----
#
plt.figure(figsize=(10,7))

plt.title('Number of Flowers by Sepal Length (cm)')
plt.xlabel('Count')
plt.ylabel('Sepal Length (cm)')

sns.histplot(data=iris_data, x='Sepal Length (cm)')

plt.show()

#
# ---- Hist Plot with Categorical Variable ----
#
plt.figure(figsize=(10,7))

plt.title('Number of Flowers by Sepal Length (cm) and Species')
plt.xlabel('Sepal Length (cm)')
plt.ylabel('Count')

sns.histplot(data=iris_data, x='Sepal Length (cm)', hue='Species')

plt.show()

#
# ---- Simple Histogram that I liked so badly ----
#
plt.figure(figsize=(10,7))

plt.title('Number of Flowers by Sepal Length (cm)')
plt.xlabel('Count')
plt.ylabel('Sepal Length (cm)')

sns.histplot(data=iris_data)

plt.show()

2.1) KDE Plots

KDE Plots show a smoother distribution (compared to the histograms) of one variable with or without a categorical one. Commands:

Simple KDE Plot

sns.kdeplot(data=data, x=x_data, shade=True|False)

KDE Plot with Categorical Variable

sns.kdeplot(data=data, x=x_data, hue=categorical_variable, shade=True|False)

#
# ---- Simple KDE Plot ----
#
plt.figure(figsize=(10,7))

plt.title('Number of Flowers per Sepal Length (cm)')
plt.xlabel('Sepal Length (cm)')
plt.ylabel('Density')

sns.kdeplot(data=iris_data, x='Sepal Length (cm)', shade=True)

plt.show()

#
# ---- KDE Plot with Categorical Variable ----
#
plt.figure(figsize=(10,7))

plt.title('Number of Flowers by Sepal Length (cm) and Species')
plt.xlabel('Sepal Length (cm)')
plt.ylabel('Density')

sns.kdeplot(data=iris_data, x='Sepal Length (cm)', hue='Species', shade=True)

plt.show()

2.2) Joint Plot

Joint Plots shows a KDE Plot of two numerical variables merged. Commands:

Joint Plot

sns.jointplot(data=data, x=x_data, y=y_data, kind='kde)

#
# ---- Joint Plot ----
#
sns.jointplot(data=iris_data
             , x='Sepal Length (cm)'
             , y='Petal Length (cm)'
             , kind='kde');

Yeah, yeah, I know that this lesson has a little too much information, but you don't have to learn all of these things at once. Use this post to review every time you have to create plots and you will get used to it.

See ya in the next lesson!! 👋