Show Menu
Cheatography

Import the Seaborn Library

from matplotlib import pyplot as plt
import seaborn as sns

import numpy as np
Seab­orn is a extension to Matp­lot­lib with more visually appealing syntax and additional Chart Types. That's why Matp­lot­lib should also be imported.

If we want to calculate aggr­ega­tes we need to import numpy aswell.

Bar Plot

sns.barplot(
 data=df ,
 x="x value column" ,
 y="y value column",

# everything specified below is optional
 ci="sd" 
estimator=np.median | len
hue="column to compare"
 )

plt.show()

# ci="sd" changes the error bar to standard deviation
# estimator is used to specifiy the aggregation and takes any argument that works on a list. (examples provided in code)
# hue adds a nested categorical variable to compare to the "y value column"
If the specified columns need to be aggr­ega­ted first, Seab­orn will perform that aggreg­ation automa­tic­ally. (mean by default)

Seab­orn will, by default, provide an error bar displaying the boot­str­apped confidence interv­al­(95%).
 

Aggregates (with numpy)

Median
np.m­edi­an­(d­f.c­olu­mn_­name)

KDE (Key Densitiy Estimator) Plots

sns.k­de­plo­t­(d­ata­set1, shade­=True)
sns.k­de­plo­t­(d­ata­set2, shade­=True)
...

KDE Plots show the dist­rib­ution of an univ­ari­ate dataset.
univ­ari­ate datasets have only one variable. (e.g.: Temper­ature)

shade defines if the are under the line is shaded

Boxplots

sns.boxplot(
 data=df, 
 x='label', 
 y='value',
 
# optional
 width=0.45
)
plt.show()

# In Seaborn it's also possible to plot multiple Boxplots in one viz
The box represents the inte­rqu­artile range
The line in the middle of the box is the median
The end lines are the first and third quarti­les
The diamonds show outl­iers

Violin Plots

sns.violinplot(
 data=df, 
 x="label", 
 y="value"
)
Two KDE plots that are symmet­rical along the center line. (Just for visual effect)

A white dot represents the median.

The thick black line in the center of each violin represents the interq­uartile range.

The lines that extend from the center are the confidence intervals (95%)
 

Seaborn Styling (Figure Style and Scale)

# Themes: (called prior to plot)
sns.set_style("")
->darkgrid, whitegrid, dark, white, ticks

#Removes Plot Borders (called after plot)
sns.despine() (default: top=True, right=True)
-> bottom, left

#Adjust font- and label size
sns.set_context(context="paper", font_scale=1.4, rc={"grid.linewidth": 0.6} )
-> paper, notebook (default), talk, poster
   In order of relative size
-> Arguments for rc parameter:
{
 'axes.labelsize': 17.6,
 'axes.titlesize': 19.0,
 'font.size': 19.2,
 'grid.linewidth': 1.6,
 'legend.fontsize': 16.0,
 'lines.linewidth': 2.8,
 'lines.markeredgewidth': 0.0,
 'lines.markersize': 11.2,
 'patch.linewidth': 0.48,
 'xtick.labelsize': 16.0,
 'xtick.major.pad': 11.2,
 'xtick.major.width': 1.6,
 'xtick.minor.width': 0.8,
 'ytick.labelsize': 16.0,
 'ytick.major.pad': 11.2,
 'ytick.major.width': 1.6,
 'ytick.minor.width': 0.8
}

Seaborn Styling (Color)

# If you want to quickly see what a palette looks like
 # Save a palette to a variable:
   palette = sns.color_palette("bright")

 # Use palplot and pass in the variable:
   sns.palplot(palette)

# Select a palette in Seaborn:
  sns.set_palette("Paired")

# Default Palettes
-> deep, muted, pastel, bright, dark, colorblind

# More Palettes using Color Brewer:
http://colorbrewer2.org
                       
 

Metadata

  • Languages:
  • Published: 26th November, 2019
  • Rated: 5 stars based on 1 ratings

Comments

No comments yet. Add yours below!

Add a Comment

Your Comment

Please enter your name.

    Please enter your email address

      Please enter your Comment.

          More Cheat Sheets by Justin1209

          NumPy