pandas and statsmodels(2018.6.6)

题目网址：

https://nbviewer.jupyter.org/github/schmit/cme193-ipython-notebooks-lecture/blob/master/Exercises.ipynb

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)

代码实现：

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")  
  
df = pd.read_csv('anscombe.csv')  
x_mean = df.groupby('dataset')['x'].mean()  
x_var = df.groupby('dataset')['x'].var()  
  
y_mean = df.groupby('dataset')['y'].mean()  
y_var = df.groupby('dataset')['y'].var()  
  
print( x_mean , x_var )  
print( round(y_mean,3) , round(y_var,3) )  
   
c = df.groupby('dataset')['x'].corr(df['y']) #get correlation  
print("the correlation coefficient is :" , c )  
x = df['x']  
y = df['y']  
data = df['dataset']  
  
for i in range(0,4): #sep the different dataset    
    flag = 'I'  
    if( i == 0 ):  
        flag = 'I'  
    elif ( i == 1 ):  
        flag = 'II'  
    elif ( i == 2 ):  
        flag = 'III'  
    else:  
        flag = 'IV'  
      
    X = []  
    Y = []  
    for ele in range(len(data)):  
        if ( data[ele] == flag):  
            X.append( x[ele] )  
            Y.append( y[ele])  
       
    X = sm.add_constant(X)    
    ols = sm.OLS(Y, X)   
    reg_func = ols.fit()   
    print('dataset '+str(i+1), "y = "+str(reg_func.params[0])+"+"+str(reg_func.params[1])+"x")   

Part 2

Using Seaborn, visualize all four datasets.

hint: use sns.FacetGrid combined with plt.scatter

代码实现：

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")  
  
df = pd.read_csv('anscombe.csv')  
x_mean = df.groupby('dataset')['x'].mean()  
x_var = df.groupby('dataset')['x'].var()  
  
y_mean = df.groupby('dataset')['y'].mean()  
y_var = df.groupby('dataset')['y'].var()  
  
print( x_mean , x_var )  
print( round(y_mean,3) , round(y_var,3) )  
  
c = df.groupby('dataset')['x'].corr(df['y']) #get correlation  
print("the correlation coefficient is :" , c )  
x = df['x']  
y = df['y']  
data = df['dataset']  

for i in range(0,4): #sep the different dataset    
    flag = 'I'  
    if( i == 0 ):  
        flag = 'I'  
    elif ( i == 1 ):  
        flag = 'II'  
    elif ( i == 2 ):  
        flag = 'III'  
    else:  
        flag = 'IV'  
      
    X = []  
    Y = []  
    for ele in range(len(data)):  
        if ( data[ele] == flag):  
            X.append( x[ele] )  
            Y.append( y[ele])  
      
    X = sm.add_constant(X)    
    ols = sm.OLS(Y, X)   
    reg_func = ols.fit()   
    print('dataset '+str(i+1), "y = "+str(reg_func.params[0])+"+"+str(reg_func.params[1])+"x")   

#part 2  
g = sns.FacetGrid(df, col="dataset")   
g.map(plt.scatter, "x","y")  
#g.map(sns.distplot, 'mpg')  
plt.show()