Anscombe's quartet

%matplotlib inline

import random

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

import statsmodels.api as sm
import statsmodels.formula.api as smf

sns.set_context("talk")

Anscombe's quartet

Anscombe's quartet comprises of four datasets, and is rather famous. Why? You'll find out in this exercise.  

anascombe = pd.read_csv('data/anscombe.csv')
anascombe.head()  

	dataset	x	y
0	I	10	8.04
1	I	8	6.95
2	I	13	7.58
3	I	9	8.81
4	I	11	8.33

Part 1

For each of the four datasets...

• Compute the mean and variance of both x and y
• Compute the correlation coefficient between x and y
• Compute the linear regression line: y=β0+β1x+ϵy=β0+β1x+ϵ (hint: use statsmodels and look at the Statsmodels notebook)

Part 2

Using Seaborn, visualize all four datasets.

hint: use sns.FacetGrid combined with plt.scatter

Code:

#matplotlib inline

import random

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

import statsmodels.api as sm
import statsmodels.formula.api as smf
import statistics as sta 
import scipy.stats.stats as stats

#part1
#计算x和y的均值和方差
#计算x和y之间的相关系数
#计算线性回归直线:y=β0+β1x+ϵ

#part2
#可视化四个数据集
#use sns.FacetGrid combined with plt.scatter

anscombe = sns.load_dataset("anscombe")
print(anscombe)#打印原数据

Xarray_I = anscombe.x[:10].values    #获取x的值，并打印
Xarray_II = anscombe.x[11:21].values
Xarray_III = anscombe.x[22:32].values
Xarray_IV = anscombe.x[33:43].values

print("Xarray in I:", Xarray_I)
print("Xarray in II:", Xarray_II)
print("Xarray in III:", Xarray_III)
print("Xarray in IV:", Xarray_IV)

Xmean_I = np.mean(Xarray_I)         #计算x的均值，并打印
print("mean of x in I: ", Xmean_I)
Xmean_II = np.mean(Xarray_II)
print("mean of x in II: ", Xmean_II)
Xmean_III = np.mean(Xarray_III)
print("mean of x in III: ", Xmean_III)
Xmean_IV = np.mean(Xarray_IV)
print("mean of x in IV: ", Xmean_IV)

Xvariance_I = sta.variance(Xarray_I)  # 计算x的方差，并打印
print("variance of x in I: ", Xvariance_I)
Xvariance_II = sta.variance(Xarray_II)
print("variance of x in II: ", Xvariance_II)
Xvariance_III = sta.variance(Xarray_III)
print("variance of x in III: ", Xvariance_III)
Xvariance_IV = sta.variance(Xarray_IV)
print("variance of x in IV: ", Xvariance_IV)

print(' ')

Yarray_I = anscombe.y[:10].values  # 获取y的值，并打印
Yarray_II = anscombe.y[11:21].values
Yarray_III = anscombe.y[22:32].values
Yarray_IV = anscombe.y[33:43].values

print("Yarray in I:", Yarray_I)
print("Yarray in II:", Yarray_II)
print("Yarray in III:", Yarray_III)
print("Yarray in IV:", Yarray_IV)

Ymean_I = np.mean(Yarray_I)  # 获取y的均值，并打印
print("mean of y in I: ", Ymean_I)
Ymean_II = np.mean(Yarray_II)
print("mean of y in II: ", Ymean_II)
Ymean_III = np.mean(Yarray_III)
print("mean of y in III: ", Ymean_III)
Ymean_IV = np.mean(Yarray_IV)
print("mean of y in IV: ", Ymean_IV)

Yvariance_I = sta.variance(Yarray_I)  # 获取y的方差，并打印
print("variance of y in I: ", Yvariance_I)
Yvariance_II = sta.variance(Yarray_II)
print("variance of y in II: ", Yvariance_II)
Yvariance_III = sta.variance(Yarray_III)
print("variance of y in III: ", Yvariance_III)
Yvariance_IV = sta.variance(Yarray_IV)
print("variance of y in IV: ", Yvariance_IV)

print('')
cof_I = stats.pearsonr(Xarray_I, Yarray_I)[0]  #计算x与y的相关系数
cof_II = stats.pearsonr(Xarray_II, Yarray_II)[0]
cof_III = stats.pearsonr(Xarray_III, Yarray_III)[0]
cof_IV = stats.pearsonr(Xarray_IV, Yarray_IV)[0]
print("correlation coefficient of I: ", cof_I)
print("correlation coefficient of II: ", cof_II)
print("correlation coefficient of III: ", cof_III)
print("correlation coefficient of IV: ", cof_IV)

print(" ")

X_I = sm.add_constant(Xarray_I)     #计算x与y的线性回归
model_I = sm.OLS(Yarray_I, X_I)
result_I = model_I.fit()
params_I = result_I.params
print("DatasetI: y =", params_I[0], "+", params_I[1], "* x")

X_II = sm.add_constant(Xarray_II)
model_II = sm.OLS(Yarray_II, X_II)
result_II = model_II.fit()
params_II = result_II.params
print("DatasetII: y =", params_II[0], "+", params_II[1], "* x")

X_III = sm.add_constant(Xarray_III)
model_III = sm.OLS(Yarray_III, X_III)
result_III = model_III.fit()
params_III = result_III.params
print("DatasetIII: y =", params_III[0], "+", params_III[1], "* x")

X_IV = sm.add_constant(Xarray_IV)
model_IV = sm.OLS(Yarray_IV, X_IV)
result_IV = model_IV.fit()
params_IV = result_IV.params
print("DatasetIV: y =", params_IV[0], "+", params_IV[1], "* x")

sns.set(style='whitegrid')      #数据可视化，散点图
g = sns.FacetGrid(anscombe, col="dataset", hue="dataset", size=3)
g.map(plt.scatter, 'x', 'y')
plt.show()

Figure: