Notebook

DoWhy example on the Lalonde dataset¶

Thanks to @mizuy for providing this example. Here we use the Lalonde dataset and apply IPW estimator to it.

In [1]:

import os, sys
sys.path.append(os.path.abspath("../../../"))

import dowhy
from dowhy import CausalModel
from rpy2.robjects import r as R
%load_ext rpy2.ipython

#%R install.packages("Matching")
%R library(Matching)

R[write to console]: Loading required package: MASS

R[write to console]: ## 
##  Matching (Version 4.9-6, Build Date: 2019-04-07)
##  See http://sekhon.berkeley.edu/matching for additional documentation.
##  Please cite software as:
##   Jasjeet S. Sekhon. 2011. ``Multivariate and Propensity Score Matching
##   Software with Automated Balance Optimization: The Matching package for R.''
##   Journal of Statistical Software, 42(7): 1-52. 
##

Out[1]:

array(['Matching', 'MASS', 'tools', 'stats', 'graphics', 'grDevices',
       'utils', 'datasets', 'methods', 'base'], dtype='<U9')

1. Load the data¶

In [2]:

%R data(lalonde)
%R -o lalonde
lalonde = lalonde.astype({'treat':'bool'}, copy=False)

Run DoWhy analysis: model, identify, estimate¶

In [3]:

model=CausalModel(
        data = lalonde,
        treatment='treat',
        outcome='re78',
        common_causes='nodegr+black+hisp+age+educ+married'.split('+'))
identified_estimand = model.identify_effect()
estimate = model.estimate_effect(identified_estimand,
        method_name="backdoor.propensity_score_weighting")
#print(estimate)
print("Causal Estimate is " + str(estimate.value))

WARNING:dowhy.causal_model:Causal Graph not provided. DoWhy will construct a graph based on data inputs.
INFO:dowhy.causal_graph:If this is observed data (not from a randomized experiment), there might always be missing confounders. Adding a node named "Unobserved Confounders" to reflect this.
INFO:dowhy.causal_model:Model to find the causal effect of treatment ['treat'] on outcome ['re78']
INFO:dowhy.causal_identifier:Common causes of treatment and outcome:['educ', 'nodegr', 'married', 'black', 'U', 'age', 'hisp']
WARNING:dowhy.causal_identifier:If this is observed data (not from a randomized experiment), there might always be missing confounders. Causal effect cannot be identified perfectly.

WARN: Do you want to continue by ignoring any unobserved confounders? (use proceed_when_unidentifiable=True to disable this prompt) [y/n] y

INFO:dowhy.causal_identifier:Instrumental variables for treatment and outcome:[]
INFO:dowhy.causal_estimator:INFO: Using Propensity Score Weighting Estimator
INFO:dowhy.causal_estimator:b: re78~treat+educ+nodegr+married+black+age+hisp

Causal Estimate is 1614.0090222453164

/home/amshar/python-environments/vpy36/lib/python3.6/site-packages/sklearn/utils/validation.py:744: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().
  y = column_or_1d(y, warn=True)

Sanity check: compare to manual IPW estimate¶

In [4]:

df = model._data
ps = df['ps']
y = df['re78']
z = df['treat']

ey1 = z*y/ps / sum(z/ps)
ey0 = (1-z)*y/(1-ps) / sum((1-z)/(1-ps))
ate = ey1.sum()-ey0.sum()
print("Causal Estimate is " + str(ate))

# correct -> Causal Estimate is 1634.9868359746906

Causal Estimate is 1639.7820238870836