Python for Scientific Computing in Economics

In [1]:

    

import this


    

    




The Zen of Python, by Tim Peters

Beautiful is better than ugly.
Explicit is better than implicit.
Simple is better than complex.
Complex is better than complicated.
Flat is better than nested.
Sparse is better than dense.
Readability counts.
Special cases aren't special enough to break the rules.
Although practicality beats purity.
Errors should never pass silently.
Unless explicitly silenced.
In the face of ambiguity, refuse the temptation to guess.
There should be one-- and preferably only one --obvious way to do it.
Although that way may not be obvious at first unless you're Dutch.
Now is better than never.
Although never is often better than *right* now.
If the implementation is hard to explain, it's a bad idea.
If the implementation is easy to explain, it may be a good idea.
Namespaces are one honking great idea -- let's do more of those!

In [2]:

    

print("Hello, World!")


    

    




Hello, World!

In [3]:

    

# Import relevant libraries from the SciPy Stack
import numpy as np

# Specify parametrization
num_agents = 1000
num_covars = 3

betas_true = np.array([0.22, 0.30, -0.1]).T

# Set a seed to ensure recomputability in light of randomness
np.random.seed(4292367295)

# Sample exogenous agent characteristics from a uniform distribution in 
# a given shape
X = np.random.rand(num_agents, num_covars)

# Sample random disturbances from a standard normal distribution and rescale
eps = np.random.normal(scale=0.1, size=num_agents)

# Construct endogenous agent characteristic
Y = np.dot(X, betas_true) + eps


    

In [4]:

    

# Import relevant libraries from the SciPy Stack
import statsmodels.api as sm

# Specify and fit the model
rslt = sm.OLS(Y, X).fit()


    

In [5]:

    

# Provide some summary information
print(rslt.summary())


    

    




                            OLS Regression Results                            
==============================================================================
Dep. Variable:                      y   R-squared:                       0.835
Model:                            OLS   Adj. R-squared:                  0.834
Method:                 Least Squares   F-statistic:                     1682.
Date:                Wed, 03 Feb 2016   Prob (F-statistic):               0.00
Time:                        17:49:15   Log-Likelihood:                 864.68
No. Observations:                1000   AIC:                            -1723.
Df Residuals:                     997   BIC:                            -1709.
Df Model:                           3                                         
Covariance Type:            nonrobust                                         
==============================================================================
                 coef    std err          t      P>|t|      [95.0% Conf. Int.]
------------------------------------------------------------------------------
x1             0.2281      0.009     24.624      0.000         0.210     0.246
x2             0.2841      0.009     30.353      0.000         0.266     0.302
x3            -0.0898      0.009     -9.484      0.000        -0.108    -0.071
==============================================================================
Omnibus:                        7.301   Durbin-Watson:                   1.964
Prob(Omnibus):                  0.026   Jarque-Bera (JB):                7.417
Skew:                           0.202   Prob(JB):                       0.0245
Kurtosis:                       2.877   Cond. No.                         3.22
==============================================================================

Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

In [6]:

    

# Import relevant libraries from the SciPy Stack
import matplotlib.pyplot as plt

# Initialize canvas
ax = plt.figure(figsize=(12, 8)).add_subplot(111, axisbg='white')

# Plot actual and fitted values
ax.plot(np.dot(X, rslt.params), Y, 'o', label='True')
ax.plot(np.dot(X, rslt.params), rslt.fittedvalues, 'r--.', label="Predicted")

# Set axis labels and ranges
ax.set_xlabel(r'$X\hat{\beta}$', fontsize=20)
ax.set_ylabel(r'$Y$', fontsize=20)

# Remove first element on y-axis
ax.yaxis.get_major_ticks()[0].set_visible(False)

# Add legend
plt.legend(loc='upper center', bbox_to_anchor=(0.50, -0.10),
    fancybox=False, frameon=False, shadow=False, ncol=2, fontsize=20)

# Add title
plt.suptitle('Synthetic Sample', fontsize=20)

# Save figure
plt.savefig('images/scatterplot.png', bbox_inches='tight', format='png')


    

In [7]:

    

from IPython.display import Image
Image(filename='images/scatterplot.png', width=700, height=700)


    

    
Out[7]:

In [8]:

    

import urllib; from IPython.core.display import HTML
HTML(urllib.urlopen('http://bit.ly/1K5apRH').read())


    

    
Out[8]: