图6-3-4 统计分布组合图.py

# -*- coding: utf-8 -*-
"""
Created on Mon Jan 28 23:05:03 2019

@author: Jie Zhang，微信公众号【EasyShu】，本代码源自《Python数据可视化之美》
"""


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

#N = 300
#x1 = np.random.normal(1.5, 0.5,N)
#y1 =  np.random.normal(2,0.2, N)
#df=pd.DataFrame({'x':x1,'y':y1})

tips = sns.load_dataset("tips")
df=pd.DataFrame({'x':tips['total_bill'],'y':tips['tip']})

sns.set_style("ticks")
sns.set_context("notebook", font_scale=1.5,
                rc={'font.size': 13, 
                    'axes.labelsize': 16, 'legend.fontsize':16, 
                    'xtick.labelsize': 15,'ytick.labelsize': 15})

#---------------------------------(a) 二维散点与统计直方图 --------------------------------
sns_hist=sns.jointplot(x='x', y='y',  # 设置xy轴，显示columns名称
              data=df,   # 设置数据
              color = '#D31A8A',   # 设置颜色
              s = 60, edgecolor="w",linewidth=0.5,
              # 设置散点大小、边缘线颜色及宽度(只针对scatter）
              kind = 'scatter', #kind="kde","hex","reg"
              space = 0,  # 设置散点图和布局图的间距
              size = 5, ratio = 5,  # 散点图与布局图高度比，整型
              xlim=(0,60),
              marginal_kws=dict(bins=20, rug=False,
                                hist_kws={'edgecolor':'k',
                                          'color':'#7CBC47',
                                          'alpha':1})  # 设置柱状图箱数，是否设置rug
              )

sns_hist.set_axis_labels(xlabel='total bill', ylabel='tip')

#sns_hist.savefig('scatter_hist.pdf') 

#---------------------------------(b) 带趋势线的二维散点与统计直方图----------------------------------------------------
sns_reg=sns.jointplot(x='x', y='y',  # 设置xy轴，显示columns名称
              data=df,   # 设置数据
              color = '#7CBC47',   # 设置颜色
              # 设置散点大小、边缘线颜色及宽度(只针对scatter）
              kind = 'reg', #kind="kde","hex","reg"
              space = 0,  # 设置散点图和布局图的间距
              size = 5, ratio = 5,  # 散点图与布局图高度比，整型
              scatter_kws={"color":"#7CBC47","alpha":0.7,"s":30,'marker':"+"},
              line_kws={"color":"#D31A8A","alpha":1,"lw":4},
              marginal_kws=dict(bins=20, rug=False,
                                hist_kws={'edgecolor':'k',
                                          'color':'#7CBC47',
                                          'alpha':1})  # 设置柱状图箱数，是否设置rug
              )
sns_reg.set_axis_labels(xlabel='total bill', ylabel='tip')
#sns_reg.savefig('scatter_reg.pdf') 

#----------------------------------(c) 二维与一维统计直方图---------------------------------------------------
sns_hex =sns.jointplot(x='x', y='y',  # 设置xy轴，显示columns名称
              data=df,   # 设置数据
              # 设置散点大小、边缘线颜色及宽度(只针对scatter）
              kind = 'hex', #kind="kde","hex","reg"
              color='#D31A8A',linewidth=0.1,
              space = 0,  # 设置散点图和布局图的间距
              size = 5, ratio = 5,  # 散点图与布局图高度比，整型
              xlim=(0,60),
              joint_kws=dict(gridsize=20,edgecolor='w'),  # 主图参数设置
              marginal_kws=dict(bins=20,color='#D31A8A',
                                    hist_kws={'edgecolor':'k','alpha':1}),   # 边缘图设置
              )  # 修改统计注释
sns_hex.set_axis_labels(xlabel='total bill', ylabel='tip')              
#sns_hex.savefig('sns_hex.pdf')   


#--------------------------------(d) 二维与一维核密度估计图--------------------------------------------------
sns_kde =sns.jointplot(x="x", y="y", data=df, kind="kde",color='#D31A8A')
sns_kde.plot_joint(plt.scatter, c="k", s=10, linewidth=1, marker="+") 
sns_kde.set_axis_labels(xlabel='total bill', ylabel='tip') 
#sns_kde.savefig('sns_kde.pdf')  


#-------------------------------(e)二维散点与一维核密度估计图---------------------------------------------
from scipy import stats
g = sns.JointGrid(x="total_bill", y="tip", data=tips, size = 5, ratio = 5)
g = g.plot_joint(plt.scatter,
                color="#00B1B8", s=40, edgecolor="black",linewidth=0.2)
g = g.plot_marginals(sns.kdeplot, shade=True,color="#00B1B8")
g = g.annotate(stats.pearsonr,loc="upper left", fontsize=12)
rsquare = lambda a, b: stats.pearsonr(a, b)[0] ** 2
g = g.annotate(rsquare, template="{stat}: {val:.2f}",
                stat="$R^2$", loc="upper left", fontsize=15)
g.set_axis_labels(xlabel='total bill', ylabel='tip') 
#g.savefig('sns_kde_scatter.pdf')