In [1]:

    

import pandas as pd
%matplotlib inline
import statsmodels.formula.api as smf
import matplotlib.pyplot as plt


    

In [4]:

    

df = pd.read_csv("data-iran.csv")


    

In [6]:

    

df.columns =['Group', 'Favor_Iran_Deal','Obama_Approve' ]


    

In [8]:

    

df.head()


    

    
Out[8]:






  

    

      

      
Group
      
Favor_Iran_Deal
      
Obama_Approve
    
  
  

    

      
0
      
Republicans
      
0.34
      
0.10
    
    

      
1
      
Democrats
      
0.60
      
0.78
    
    

      
2
      
White
      
0.45
      
0.35
    
    

      
3
      
Black
      
0.54
      
0.85
    
    

      
4
      
Ind
      
0.44
      
0.37
    
  

In [10]:

    

df.corr()['Favor_Iran_Deal']


    

    
Out[10]:





Favor_Iran_Deal    1.000000
Obama_Approve      0.913868
Name: Favor_Iran_Deal, dtype: float64

In [12]:

    

lm = smf.ols(formula="Favor_Iran_Deal~Obama_Approve",data=df).fit()


    

In [13]:

    

intercept, slope = lm.params


    

In [14]:

    

lm.summary()


    

    




c:\users\dongjin\envs\03stat\lib\site-packages\scipy\stats\stats.py:1535: UserWarning: kurtosistest only valid for n>=20 ... continuing anyway, n=18
  "anyway, n=%i" % int(n))






    
Out[14]:





OLS Regression Results

  
Dep. Variable:
     
Favor_Iran_Deal
 
  R-squared:         
 
   0.835


  
Model:
                   
OLS
       
  Adj. R-squared:    
 
   0.825


  
Method:
             
Least Squares
  
  F-statistic:       
 
   81.06


  
Date:
             
Mon, 01 Aug 2016
 
  Prob (F-statistic):
 
1.16e-07


  
Time:
                 
23:03:36
     
  Log-Likelihood:    
 
  37.901


  
No. Observations:
      
    18
      
  AIC:               
 
  -71.80


  
Df Residuals:
          
    16
      
  BIC:               
 
  -70.02


  
Df Model:
              
     1
      
                     
     
 
   


  
Covariance Type:
      
nonrobust
    
                     
     
 
   




        
           
coef
     
std err
      
t
      
P>|t|
  
[0.025
    
0.975]
  


  
Intercept
     
    0.3053
 
    0.019
 
   16.036
 
 0.000
 
    0.265
 
    0.346


  
Obama_Approve
 
    0.3556
 
    0.039
 
    9.003
 
 0.000
 
    0.272
 
    0.439




  
Omnibus:
       
 1.636
 
  Durbin-Watson:     
 
   2.194


  
Prob(Omnibus):
 
 0.441
 
  Jarque-Bera (JB):  
 
   0.621


  
Skew:
          
-0.443
 
  Prob(JB):          
 
   0.733


  
Kurtosis:
      
 3.207
 
  Cond. No.          
 
    6.45

In [16]:

    

df.plot(kind="scatter",x="Obama_Approve",y="Favor_Iran_Deal")
plt.plot(df["Obama_Approve"],slope*df["Obama_Approve"]+intercept,"-",color="red") 
plt.xlabel('Approve of Obama')
plt.ylabel('Favors Iran Deal')


    

    
Out[16]:





<matplotlib.text.Text at 0xbac70b8>