Machine learning Seaborn visualization

Main records seaborn Visual learning notes （ Understand which functions to draw images are available ）.

---

One 、seaborn principle

picture source ：https://www.bilibili.com/video/BV1VX4y1F76x/

• boxenplot： Suitable for big data
• Distribution diagram of numerical variables in different categories ：stripplot;swarmplot;violinplot
• FaceGrid,PairGrid You can customize the drawing function

see seaborn edition ：sns.__version__
Version update ：pip install —upgrade seaborn

Two 、 Variable distribution

1.sns.boxplot()： View the value range of numeric variables

sns.boxplot()： View the value range of numeric variables , Whether there are outliers .

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

print(sns.__version__)  
# print(sns.get_dataset_names())

df = pd.read_excel('D:/1.xlsx')

sns.boxplot(data=df,x="Height")
plt.show()

2.sns.displot()： View the distribution of variables

• sns.displot(kind = hist) # Draw histogram
  Histogram ：sns.histplot（bins,hue,shrink）
  bins： change bin numbers
  hue： Category variable
  shrink： Zoom factor
• sns.displot(kind = kde) # Plotting kernel density estimates （kernel density estimate (KDE)）, It is a method to visualize the distribution of observations in data sets , Similar to histogram .KDE Use a continuous probability density curve of one or more dimensions to represent data .
• sns.displot(kind = ecdf) # Represents the proportion or count of observations below each unique value in the dataset . Compare with histogram or density diagram , Its advantage is that each observation is directly visualized , This means that there is no need to adjust the box dividing or smoothing parameters .

penguins = sns.load_dataset("penguins")
sns.ecdfplot(data=penguins, x="flipper_length_mm")

• sns.countplot(data=df,x=“class”) Number of Statistics

3.sns.jointplot()： Plot the joint distribution and respective distribution of two variables

sns.jointplot(dataset,x,y,kind)

sns.jointplot() Function upgrade ：
JoinGrid, Can pass g.plot() Custom function .g = sns.JoinGrid(); g.plot(sns.histplot,sns.boxplot)

4.sns.pairplot()： Plot the joint distribution of all numerical variables in pairs

sns.pairplot() Function upgrade ：
PairGrid, Can pass g.map() Custom drawing function

Reference material

https://www.bilibili.com/video/BV1VX4y1F76x/

https://blog.csdn.net/qq_45176548/article/details/117305614?utm_medium=distribute.pc_relevant.none-task-blog-2_defaultbaidujs_title~default-1.queryctrv2&spm=1001.2101.3001.4242.2&utm_relevant_index=4