In [2]:

    

import pandas as pd
from matplotlib import pyplot as plt

%matplotlib inline


    

In [13]:

    

empl_data = pd.read_csv('fixtures/training_data.csv')
test_data = pd.read_csv('fixtures/test_data.csv')


    

In [15]:

    

print empl_data.shape
print test_data.shape


    

    




(108, 34)
(12, 34)

In [5]:

    

empl_data.head()


    

    
Out[5]:






  

    

      

      
job_growth
      
job_growth_per
      
job_growth_min_1
      
job_growth_min_3
      
job_growth_min_6
      
job_growth_min_12
      
bls_avg_weeks_unemp_per
      
avg_weeks_unemp
      
avg_weeks_unemp_min_1
      
avg_weeks_unemp_min_3
      
...
      
taxes_min_1
      
taxes_min_3
      
taxes_min_6
      
taxes_min_12
      
outlays
      
spending
      
spending_min_1
      
spending_min_3
      
spending_min_6
      
spending_min_12
    
  
  

    

      
0
      
 0.085638
      
 0.000856
      
-0.075590
      
 0.175644
      
 0.274477
      
 0.050552
      
 19.5
      
-6.8
      
-6.890756
      
-6.790756
      
...
      
 19419
      
-59434
      
-61915
      
-104321
      
 194111
      
-64514
      
-40315
      
-65111
      
-55050
      
-69916
    
    

      
1
      
 0.099825
      
 0.000998
      
 0.085638
      
 0.357112
      
 0.012181
      
-0.064240
      
 19.1
      
-7.2
      
-6.790756
      
-6.590756
      
...
      
  5887
      
-61783
      
-58601
      
 -63905
      
 214660
      
-43965
      
-64514
      
-66908
      
-79764
      
-53288
    
    

      
2
      
 0.191616
      
 0.001916
      
 0.099825
      
-0.075590
      
-0.061617
      
 0.163955
      
 19.5
      
-6.8
      
-7.190756
      
-6.890756
      
...
      
-95459
      
 19419
      
 11021
      
  23761
      
 220483
      
-38142
      
-43965
      
-40315
      
-75885
      
-56112
    
    

      
3
      
 0.426579
      
 0.004266
      
 0.191616
      
 0.085638
      
 0.175644
      
 0.124027
      
 19.6
      
-6.7
      
-6.790756
      
-6.790756
      
...
      
-47571
      
  5887
      
-59434
      
 -80880
      
 219690
      
-38935
      
-38142
      
-64514
      
-65111
      
-80712
    
    

      
4
      
 0.251320
      
 0.002513
      
 0.426579
      
 0.099825
      
 0.357112
      
 0.231902
      
 18.6
      
-7.7
      
-6.690756
      
-7.190756
      
...
      
 81284
      
-95459
      
-61783
      
  18052
      
 188920
      
-69705
      
-38935
      
-43965
      
-66908
      
-63367
    
  

5 rows × 48 columns

In [6]:

    

fig, axs = plt.subplots(1, 3, sharey=True)
empl_data.plot(kind='scatter', x='cpi_idx', y='job_growth', ax=axs[0], figsize=(16,8))
empl_data.plot(kind='scatter', x='cpi_index_min_3', y='job_growth', ax=axs[1], figsize=(16,8))
empl_data.plot(kind='scatter', x='cpi_index_min_12', y='job_growth', ax=axs[2], figsize=(16,8))


    

    
Out[6]:





<matplotlib.axes._subplots.AxesSubplot at 0x10bb63950>






    

In [16]:

    

# follow the usual sklearn pattern: import, instantiate, fit
from sklearn import linear_model
regr = linear_model.Lasso(alpha=.3)


    

In [18]:

    

# using scikit-learn
feature_cols = ['job_growth_min_1',
 'job_growth_min_3',
 'job_growth_min_6',
 'job_growth_min_12',
 'avg_weeks_unemp',
 'emp_pop_ratio',
 'lbr_frc_prtcp',
 'not_in_lbr_frc',
 'totl_emp_payrl',
 'unemp_rate',
 'fed_fund_rate',
 'mortgage_rate',
 'cpi_idx',
 'cpi_index_min_1',
 'cpi_index_min_3',
 'cpi_index_min_6',
 'cpi_index_min_12',
 'export_idx',
 'import_idx',
 'taxes',
 'taxes_min_1',
 'taxes_min_3',
 'taxes_min_6',
 'taxes_min_12',
 'spending',
 'spending_min_1',
 'spending_min_3',
 'spending_min_6',
 'spending_min_12']
X = empl_data[feature_cols]
y = empl_data['job_growth']


    

In [19]:

    

regr.fit(X, y)


    

    
Out[19]:





Lasso(alpha=0.3, copy_X=True, fit_intercept=True, max_iter=1000,
   normalize=False, positive=False, precompute='auto', tol=0.0001,
   warm_start=False)

In [24]:

    

regr.score(X,y)


    

    
Out[24]:





0.24025627272146721

In [20]:

    

regr.coef_


    

    
Out[20]:





array([  0.00000000e+00,   0.00000000e+00,   0.00000000e+00,
         0.00000000e+00,   8.03984485e-03,   0.00000000e+00,
        -0.00000000e+00,   2.87458611e-04,   3.11837998e-05,
        -0.00000000e+00,   0.00000000e+00,  -0.00000000e+00,
        -0.00000000e+00,  -0.00000000e+00,  -0.00000000e+00,
        -0.00000000e+00,  -0.00000000e+00,   0.00000000e+00,
         0.00000000e+00,  -4.54083370e-07,  -4.72029232e-07,
        -1.37079301e-07,  -7.45042452e-07,  -4.20317310e-07,
        -1.79906363e-06,  -1.35544634e-06,  -1.77981275e-07,
         5.65260857e-07,  -5.88612741e-07])

In [22]:

    

X_test = test_data[feature_cols]
y_test = test_data['job_growth']


    

In [23]:

    

regr.score(X_test, y_test)


    

    
Out[23]:





-21.966515138132593

In [ ]: