Seaborn entry series (2)-barplot countplot pointplot

Seaborn is a Python visualization library based on matplotlib. It provides an advanced interface to draw attractive statistical graphics. Seaborn is actually a more advanced API encapsulation on the basis of matplotlib, which makes drawing easier, and it does not require a lot of adjustments to make your graphs refined.

Note: All codes are implemented in IPython notebook

barplot (bar graph)

bar graph represents the estimated value of the central trend of the numerical variable and the height of each rectangle, and uses error bars to provide uncertainty about the estimated value Some instructions. The specific usage is as follows:

seaborn.barplot(x=None, y=None, hue=None, data=None, order=None, hue_order=None, estimator=(function mean), ci=95, n_boot=1000, units=None , orient=None, color=None, palette=None, saturation=0.75, errcolor=&39;.26&39;, errwidth=None, capsize=None, dodge=True, ax=None, **kwargs)

Specific examples to learn the usage of some of the parameters inside:

%matplotlib inline
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
 is used to solve the problem of not displaying Chinese 
plt.rc("font",family="SimHei", size="12") 
sns.set_style("whitegrid") 

The content of the data set used in this article is as follows:

data.head(5) data is a dataframe 

x, y (str): the column name in the dataframe

data: dataframe or Array

sns.barplot(x="color", y="age", data=data)

Explanation of the image: Seaborn will classify the value in the "color" column and then follow the method of estimator parameters (default Calculate the corresponding value for the average value, and the value calculated by is used as the value displayed in the bar graph (the error bar on the bar graph indicates the error of the various values ​​of relative to the value displayed in the bar graph )

hue (str): the column name of the dataframe, according to the value in the column name to form a classified bar graph

sns.barplot(x="color" y="age",data=data,hu e="gender")

order, hue_order (lists of strings): used to control the order of the bar graph

fig,axes=plt.subplots(1,2)
sns.barplot(x="gender", y="age" , data=data, ax=axes[0])
sns.barplot(x="gender", y="age", data=data, ax=axes[1], order=["女","男"])

estimator: controls what value the bar graph takes in each column of data

fig,axes=plt.subplots(1,2)
 on the left, the default is the average value 
sns .barplot(x="gender", y="age", data=data, ax=axes[0]) 
 right picture, median 
sns.barplot(x="gender", y="age", estimator =np.median, data=data, ax=axes[1]) 

ci (float): Confidence interval (between 0-100), if you fill in "sd", the error bar uses the standard error. (The default is 95)

fig,axes=plt.subplots(1,2)
sns.barplot(x="color", y="age", data=data, ci=0, ax=axes[0]) left 
sns.barplot(x="color", y="age", data=data, ci="sd", ax=axes[1]) The picture on the right 

capsize(float): sets the error bar caps ( up and down two The width of the root horizontal line )

fig,axes=plt.subplots(1,2)
sns.barplot(x="color", y="age", data=data, ax=axes[0], capsize=.2 ) The left picture 
sns.barplot(x="color", y="age", data=data, ax=axes[1], capsize=.5) The right picture 

palette: palette, control different colors style

fig,axes =plt.subplots(2,1)
sns.barplot(x="color", y="age", data=data, ax=axes[0]) The picture above 
sns.barplot(x="color", y= "age", data=data, palette="Set3", ax=axes[1]) The following figure 

X, Y-axis interchange

fig,axes=plt.subplots(1,2)
sns.barplot(x="age" ,y="color",data=data,ax=axes[0]) left image 
sns.barplot(x="color",y="age",data=data,ax=axes[1]) right image 

countplot Getting started

A count chart can be thought of as a classification histogram , rather than a quantitative variable. The basic api and options are the same as barplot(), so you can compare counts in nested variables. (The working principle is that classifies the input data , and the bar graph shows the number of categories ) The specific usage is as follows:

seaborn.countplot(x=None, y=None, hue=None, data=None, order=None , hue_order=None, orient=None, color=None, palette=None, saturation=0.75, dodge=True, ax=None, **kwargs)

Note: The parameters of countplot are basically the same as those of barplot, which can be compared and remembered. One difference is that x and y cannot be input at the same time in countplot, and there is no confidence interval for countplot.

Experience it based on an example:

fig,axes=plt.subplots(1,2)
sns.countplot(x="gender", data=data, ax=axes[0]) Left image 
sns.countplot(y="gender" , data=data, ax=axes[1]) right 

fig,axes=plt.subplots(1,2)
sns.countplot(x="gender", hue="smoker", data=data, ax=axes[ 0]) Left image 
sns.countplot(y="gender", hue="smoker", data=data, ax=axes[1]) Right image 

fig,axes=plt.subplots(2,1)
sns.countplot( x="color", data=data, ax=axes[0]) The above figure 
sns.countplot(x="color", data=data, palette="Set3", ax=axes[1]) The figure below 

sns.pointplot(x="smoker", y="age", data=data, join=False)

sns.pointplot(x=" smoker", y="age", data=data, hue="gender")

sns.pointplot(x="smoker", y="age", data=data, hue="gender", dodge=True)

sns.pointplot(x="smoker", y="age", data=data, hue="gender", dodge=True,
 markers=["*","x"], linestyles=["-." ,"--"])

sns.pointplot(x="age",y="color",data=data)

sns.pointplot(x="age",y="color",data=data,color="bb3f3f")