Customizing

seaborn

Load packages

# numerical calculation & data frames
import numpy as np
import pandas as pd

# visualization
import matplotlib.pyplot as plt
import seaborn as sns
import seaborn.objects as so

# statistics
import statsmodels.api as sm

Customizing

• Seaborn/Customizing the appearance
  
• Reference

Scales and layouts

penguins = sns.load_dataset("penguins")

(
    so.Plot(penguins, x="body_mass_g", y="species", color="island")
    .facet(col="sex")
    .add(so.Dot(), so.Jitter(.5))
    .scale(color="Set2")  # color palettes: "Set2"
    .layout(size=(8, 5))  # plot size
)

diamonds = sns.load_dataset("diamonds")

(
    so.Plot(diamonds, x="carat", y="price", color="carat", marker="cut")
    .add(so.Dots())
    .scale(
        color=so.Continuous("crest", norm=(0, 3), trans="sqrt"),
    )
)

Choosing color palettes

Legends and ticks

(
    so.Plot(penguins, x="species")
    .add(so.Bar(), so.Count())
    .scale(x=so.Nominal(order=["Adelie", "Gentoo", "Chinstrap"])) # x축의 카테고리 순서를 변경
)

(
    so.Plot(diamonds, x="carat", y="price", color="carat")
    .add(so.Dots())
    .scale(
        x=so.Continuous().tick(every=0.5),
        y=so.Continuous().label(like="${x:.0f}"), # %표시: like="{x:.1%}" 
        color=so.Continuous().tick(at=[1, 2, 3, 4]),
    )
)

Limits, labels, and titles

Plot has a number of methods for simple customization, including Plot.label(), Plot.limit(), and Plot.share():

penguins = sns.load_dataset("penguins")

(
    so.Plot(penguins, x="body_mass_g", y="species", color="island")
    .facet(col="sex")
    .add(so.Dot(), so.Jitter(.5))
    .share(x=False)
    .limit(y=(2.5, -.5))
    .label(
        x="Body mass (g)", y="",
        color=str.capitalize,
        title="{} penguins".format,
    )
)

Themes

tips = sns.load_dataset("tips")
p = (
    so.Plot(tips, x="total_bill", y="tip", color="time")
    .add(so.Dot())
)
p.theme({"axes.facecolor": "white", 
         "axes.edgecolor": "0.8", 
         'axes.spines.top': False,
         'axes.spines.right': False})

seaborn.axes_style

from seaborn import axes_style
p.theme({**axes_style("whitegrid")})

Seaborn: controlling figure aesthetics

from matplotlib import style
p.theme({**style.library["fivethirtyeight"]})

Custom plots

sbcustom.py로 working directory에 저장 또는 conda 환경 폴더에 저장

"""
Seaborn.objects statistical plotting custom functions.
boxplot, rangeplot
"""

import seaborn as sns
import seaborn.objects as so


def boxplot(df, x, y, color=None, alpha=0.1, marker="<"):
    
    return (
        so.Plot(df, x=x, y=y, color=color)
        .add(so.Dots(alpha=alpha, color=".6"), so.Jitter(), so.Dodge())
        .add(so.Range(), so.Est("median", errorbar=("pi", 50)), so.Dodge())
        .add(so.Dot(pointsize=8, marker=marker), so.Agg("median"), so.Dodge())
        .scale(color="Dark2")
        .theme({**sns.axes_style("whitegrid")})
    )


def rangeplot(df, x, y, color=None, alpha=0.1):

    return (
        so.Plot(df, x=x, y=y, color=color)
        .add(so.Range(), so.Est("median", errorbar=("pi", 50)), so.Dodge())
        .add(so.Dots(pointsize=8, marker="<"), so.Agg("median"), so.Dodge())
        .scale(color="Dark2")
        .theme({**sns.axes_style("whitegrid")})
    )

from sbcustom import boxplot, rangeplot
rangeplot(penguins, x="species", y="body_mass_g")

boxplot(penguins, x="species", y="body_mass_g", color="island")

Snippets

Snippets: Configure User Snippets

"seaborn.obj": {
        "prefix": "sbj",
        "body": [
            "(",
            "\tso.Plot($1, x='$2', y='$3')",
            "\t.add($0)",
            ")",
        ],
        "description": "plot seaborn.objects"
    },

Multiple plots in a Jupyter Notebook

display()
.show() (pyplot으로 변환; 미진한 면이 있음)

p1 = rangeplot(penguins, x="species", y="body_mass_g")
p2 = boxplot(penguins, x="species", y="body_mass_g", color="island")

display(p1, p2)  # from IPython.display import display

p1.show(); p2.show()