{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Chapter 5: Explanation\n",
"\n",
"Load the modules."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import scipy as sp\n",
"import statsmodels.api as sm\n",
"import statsmodels.formula.api as smf"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Simple Meaning\n",
"\n",
"Read in the data. Drop the unused variable."
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" Species | \n",
" Area | \n",
" Elevation | \n",
" Nearest | \n",
" Scruz | \n",
" Adjacent | \n",
"
\n",
" \n",
" \n",
" \n",
" Baltra | \n",
" 58 | \n",
" 25.09 | \n",
" 346 | \n",
" 0.6 | \n",
" 0.6 | \n",
" 1.84 | \n",
"
\n",
" \n",
" Bartolome | \n",
" 31 | \n",
" 1.24 | \n",
" 109 | \n",
" 0.6 | \n",
" 26.3 | \n",
" 572.33 | \n",
"
\n",
" \n",
" Caldwell | \n",
" 3 | \n",
" 0.21 | \n",
" 114 | \n",
" 2.8 | \n",
" 58.7 | \n",
" 0.78 | \n",
"
\n",
" \n",
" Champion | \n",
" 25 | \n",
" 0.10 | \n",
" 46 | \n",
" 1.9 | \n",
" 47.4 | \n",
" 0.18 | \n",
"
\n",
" \n",
" Coamano | \n",
" 2 | \n",
" 0.05 | \n",
" 77 | \n",
" 1.9 | \n",
" 1.9 | \n",
" 903.82 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Species Area Elevation Nearest Scruz Adjacent\n",
"Baltra 58 25.09 346 0.6 0.6 1.84\n",
"Bartolome 31 1.24 109 0.6 26.3 572.33\n",
"Caldwell 3 0.21 114 2.8 58.7 0.78\n",
"Champion 25 0.10 46 1.9 47.4 0.18\n",
"Coamano 2 0.05 77 1.9 1.9 903.82"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"gala = pd.read_csv(\"data/gala.csv\", index_col=0)\n",
"gala.drop('Endemics', axis=1, inplace=True)\n",
"gala.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Fit the basic model"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | Species | R-squared: | 0.766 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.717 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 15.70 | \n",
"
\n",
"\n",
" Date: | Fri, 07 Sep 2018 | Prob (F-statistic): | 6.84e-07 | \n",
"
\n",
"\n",
" Time: | 15:26:43 | Log-Likelihood: | -162.54 | \n",
"
\n",
"\n",
" No. Observations: | 30 | AIC: | 337.1 | \n",
"
\n",
"\n",
" Df Residuals: | 24 | BIC: | 345.5 | \n",
"
\n",
"\n",
" Df Model: | 5 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 7.0682 | 19.154 | 0.369 | 0.715 | -32.464 | 46.601 | \n",
"
\n",
"\n",
" Area | -0.0239 | 0.022 | -1.068 | 0.296 | -0.070 | 0.022 | \n",
"
\n",
"\n",
" Elevation | 0.3195 | 0.054 | 5.953 | 0.000 | 0.209 | 0.430 | \n",
"
\n",
"\n",
" Nearest | 0.0091 | 1.054 | 0.009 | 0.993 | -2.166 | 2.185 | \n",
"
\n",
"\n",
" Scruz | -0.2405 | 0.215 | -1.117 | 0.275 | -0.685 | 0.204 | \n",
"
\n",
"\n",
" Adjacent | -0.0748 | 0.018 | -4.226 | 0.000 | -0.111 | -0.038 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 12.683 | Durbin-Watson: | 2.476 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.002 | Jarque-Bera (JB): | 13.498 | \n",
"
\n",
"\n",
" Skew: | 1.136 | Prob(JB): | 0.00117 | \n",
"
\n",
"\n",
" Kurtosis: | 5.374 | Cond. No. | 1.90e+03 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 1.9e+03. This might indicate that there are
strong multicollinearity or other numerical problems."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: Species R-squared: 0.766\n",
"Model: OLS Adj. R-squared: 0.717\n",
"Method: Least Squares F-statistic: 15.70\n",
"Date: Fri, 07 Sep 2018 Prob (F-statistic): 6.84e-07\n",
"Time: 15:26:43 Log-Likelihood: -162.54\n",
"No. Observations: 30 AIC: 337.1\n",
"Df Residuals: 24 BIC: 345.5\n",
"Df Model: 5 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"Intercept 7.0682 19.154 0.369 0.715 -32.464 46.601\n",
"Area -0.0239 0.022 -1.068 0.296 -0.070 0.022\n",
"Elevation 0.3195 0.054 5.953 0.000 0.209 0.430\n",
"Nearest 0.0091 1.054 0.009 0.993 -2.166 2.185\n",
"Scruz -0.2405 0.215 -1.117 0.275 -0.685 0.204\n",
"Adjacent -0.0748 0.018 -4.226 0.000 -0.111 -0.038\n",
"==============================================================================\n",
"Omnibus: 12.683 Durbin-Watson: 2.476\n",
"Prob(Omnibus): 0.002 Jarque-Bera (JB): 13.498\n",
"Skew: 1.136 Prob(JB): 0.00117\n",
"Kurtosis: 5.374 Cond. No. 1.90e+03\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"[2] The condition number is large, 1.9e+03. This might indicate that there are\n",
"strong multicollinearity or other numerical problems.\n",
"\"\"\""
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lmod = smf.ols(formula='Species ~ Area + Elevation + Nearest + Scruz + Adjacent', data=gala).fit()\n",
"lmod.summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"See how coefficient is different when only one predictor is used:"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | Species | R-squared: | 0.545 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.529 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 33.59 | \n",
"
\n",
"\n",
" Date: | Fri, 07 Sep 2018 | Prob (F-statistic): | 3.18e-06 | \n",
"
\n",
"\n",
" Time: | 15:26:43 | Log-Likelihood: | -172.49 | \n",
"
\n",
"\n",
" No. Observations: | 30 | AIC: | 349.0 | \n",
"
\n",
"\n",
" Df Residuals: | 28 | BIC: | 351.8 | \n",
"
\n",
"\n",
" Df Model: | 1 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 11.3351 | 19.205 | 0.590 | 0.560 | -28.005 | 50.675 | \n",
"
\n",
"\n",
" Elevation | 0.2008 | 0.035 | 5.795 | 0.000 | 0.130 | 0.272 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 14.860 | Durbin-Watson: | 2.021 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.001 | Jarque-Bera (JB): | 29.281 | \n",
"
\n",
"\n",
" Skew: | 0.878 | Prob(JB): | 4.38e-07 | \n",
"
\n",
"\n",
" Kurtosis: | 7.510 | Cond. No. | 741. | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: Species R-squared: 0.545\n",
"Model: OLS Adj. R-squared: 0.529\n",
"Method: Least Squares F-statistic: 33.59\n",
"Date: Fri, 07 Sep 2018 Prob (F-statistic): 3.18e-06\n",
"Time: 15:26:43 Log-Likelihood: -172.49\n",
"No. Observations: 30 AIC: 349.0\n",
"Df Residuals: 28 BIC: 351.8\n",
"Df Model: 1 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"Intercept 11.3351 19.205 0.590 0.560 -28.005 50.675\n",
"Elevation 0.2008 0.035 5.795 0.000 0.130 0.272\n",
"==============================================================================\n",
"Omnibus: 14.860 Durbin-Watson: 2.021\n",
"Prob(Omnibus): 0.001 Jarque-Bera (JB): 29.281\n",
"Skew: 0.878 Prob(JB): 4.38e-07\n",
"Kurtosis: 7.510 Cond. No. 741.\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"smf.ols(formula='Species ~ Elevation', data=gala).fit().summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Set up a plot with univariate regression line added"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[]"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(gala.Elevation, gala.Species)\n",
"plt.plot([0,1750],[11.3, 11.3+1750*0.2008])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Mean of all variables"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([ 1. , 261.70866667, 368.03333333, 10.06 ,\n",
" 56.97666667, 261.09833333])"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x0 = np.append(1,gala.iloc[:,1:].mean())\n",
"x0"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Find the predicted output when Elevation varies from its minimum to its maximum value."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[-24.35372847730425, 512.985999339698]"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x0[2] = np.min(gala.Elevation)\n",
"yl = np.dot(lmod.params, x0)\n",
"x0[2] = np.max(gala.Elevation)\n",
"yh = np.dot(lmod.params, x0)\n",
"[yl, yh]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Add this covariate adjusted prediction to the plot and display."
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[]"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(gala.Elevation, gala.Species)\n",
"plt.plot([0,1750],[11.3, 11.3+1750*0.2008])\n",
"plt.plot([min(gala.Elevation),max(gala.Elevation)],[yl,yh])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Observational Data\n",
"\n",
"Read in the New Hampshire voting data"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"newhamp = pd.read_csv(\"data/newhamp.csv\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Sum of votes for the candidates by voting method."
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" Obama | \n",
" Clinton | \n",
"
\n",
" \n",
" votesys | \n",
" | \n",
" | \n",
"
\n",
" \n",
" \n",
" \n",
" D | \n",
" 86353 | \n",
" 96890 | \n",
"
\n",
" \n",
" H | \n",
" 16926 | \n",
" 14471 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Obama Clinton\n",
"votesys \n",
"D 86353 96890\n",
"H 16926 14471"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"newhamp.groupby('votesys').agg({'Obama': sum, 'Clinton': sum})"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Set up a binary variable"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"newhamp['trt'] = np.where(newhamp.votesys == 'H',1,0)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Proportion voting for obama by voting method. See there is a significant difference of 4.25%"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | pObama | R-squared: | 0.083 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.080 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 24.93 | \n",
"
\n",
"\n",
" Date: | Fri, 07 Sep 2018 | Prob (F-statistic): | 1.06e-06 | \n",
"
\n",
"\n",
" Time: | 15:26:43 | Log-Likelihood: | 350.38 | \n",
"
\n",
"\n",
" No. Observations: | 276 | AIC: | -696.8 | \n",
"
\n",
"\n",
" Df Residuals: | 274 | BIC: | -689.5 | \n",
"
\n",
"\n",
" Df Model: | 1 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 0.3525 | 0.005 | 68.148 | 0.000 | 0.342 | 0.363 | \n",
"
\n",
"\n",
" trt | 0.0425 | 0.009 | 4.993 | 0.000 | 0.026 | 0.059 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 5.996 | Durbin-Watson: | 1.579 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.050 | Jarque-Bera (JB): | 5.738 | \n",
"
\n",
"\n",
" Skew: | 0.312 | Prob(JB): | 0.0568 | \n",
"
\n",
"\n",
" Kurtosis: | 3.330 | Cond. No. | 2.43 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: pObama R-squared: 0.083\n",
"Model: OLS Adj. R-squared: 0.080\n",
"Method: Least Squares F-statistic: 24.93\n",
"Date: Fri, 07 Sep 2018 Prob (F-statistic): 1.06e-06\n",
"Time: 15:26:43 Log-Likelihood: 350.38\n",
"No. Observations: 276 AIC: -696.8\n",
"Df Residuals: 274 BIC: -689.5\n",
"Df Model: 1 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"Intercept 0.3525 0.005 68.148 0.000 0.342 0.363\n",
"trt 0.0425 0.009 4.993 0.000 0.026 0.059\n",
"==============================================================================\n",
"Omnibus: 5.996 Durbin-Watson: 1.579\n",
"Prob(Omnibus): 0.050 Jarque-Bera (JB): 5.738\n",
"Skew: 0.312 Prob(JB): 0.0568\n",
"Kurtosis: 3.330 Cond. No. 2.43\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"smf.ols(formula='pObama ~ trt',data=newhamp).fit().summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Adjust by previous votes for Dean. Treatment effect becomes insignificant."
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | pObama | R-squared: | 0.419 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.415 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 98.40 | \n",
"
\n",
"\n",
" Date: | Fri, 07 Sep 2018 | Prob (F-statistic): | 6.61e-33 | \n",
"
\n",
"\n",
" Time: | 15:26:43 | Log-Likelihood: | 413.28 | \n",
"
\n",
"\n",
" No. Observations: | 276 | AIC: | -820.6 | \n",
"
\n",
"\n",
" Df Residuals: | 273 | BIC: | -809.7 | \n",
"
\n",
"\n",
" Df Model: | 2 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 0.2211 | 0.011 | 19.655 | 0.000 | 0.199 | 0.243 | \n",
"
\n",
"\n",
" trt | -0.0048 | 0.008 | -0.613 | 0.541 | -0.020 | 0.011 | \n",
"
\n",
"\n",
" Dean | 0.5229 | 0.042 | 12.555 | 0.000 | 0.441 | 0.605 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 10.671 | Durbin-Watson: | 1.915 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.005 | Jarque-Bera (JB): | 14.156 | \n",
"
\n",
"\n",
" Skew: | 0.304 | Prob(JB): | 0.000843 | \n",
"
\n",
"\n",
" Kurtosis: | 3.928 | Cond. No. | 14.7 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: pObama R-squared: 0.419\n",
"Model: OLS Adj. R-squared: 0.415\n",
"Method: Least Squares F-statistic: 98.40\n",
"Date: Fri, 07 Sep 2018 Prob (F-statistic): 6.61e-33\n",
"Time: 15:26:43 Log-Likelihood: 413.28\n",
"No. Observations: 276 AIC: -820.6\n",
"Df Residuals: 273 BIC: -809.7\n",
"Df Model: 2 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"Intercept 0.2211 0.011 19.655 0.000 0.199 0.243\n",
"trt -0.0048 0.008 -0.613 0.541 -0.020 0.011\n",
"Dean 0.5229 0.042 12.555 0.000 0.441 0.605\n",
"==============================================================================\n",
"Omnibus: 10.671 Durbin-Watson: 1.915\n",
"Prob(Omnibus): 0.005 Jarque-Bera (JB): 14.156\n",
"Skew: 0.304 Prob(JB): 0.000843\n",
"Kurtosis: 3.928 Cond. No. 14.7\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"smf.ols(formula='pObama ~ trt + Dean',data=newhamp).fit().summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Check that proportion voting for Dean changes by voting method."
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"OLS Regression Results\n",
"\n",
" Dep. Variable: | Dean | R-squared: | 0.235 | \n",
"
\n",
"\n",
" Model: | OLS | Adj. R-squared: | 0.232 | \n",
"
\n",
"\n",
" Method: | Least Squares | F-statistic: | 84.21 | \n",
"
\n",
"\n",
" Date: | Fri, 07 Sep 2018 | Prob (F-statistic): | 1.11e-17 | \n",
"
\n",
"\n",
" Time: | 15:26:43 | Log-Likelihood: | 310.13 | \n",
"
\n",
"\n",
" No. Observations: | 276 | AIC: | -616.3 | \n",
"
\n",
"\n",
" Df Residuals: | 274 | BIC: | -609.0 | \n",
"
\n",
"\n",
" Df Model: | 1 | | | \n",
"
\n",
"\n",
" Covariance Type: | nonrobust | | | \n",
"
\n",
"
\n",
"\n",
"\n",
" | coef | std err | t | P>|t| | [0.025 | 0.975] | \n",
"
\n",
"\n",
" Intercept | 0.2513 | 0.006 | 41.986 | 0.000 | 0.240 | 0.263 | \n",
"
\n",
"\n",
" trt | 0.0903 | 0.010 | 9.177 | 0.000 | 0.071 | 0.110 | \n",
"
\n",
"
\n",
"\n",
"\n",
" Omnibus: | 14.901 | Durbin-Watson: | 1.206 | \n",
"
\n",
"\n",
" Prob(Omnibus): | 0.001 | Jarque-Bera (JB): | 15.737 | \n",
"
\n",
"\n",
" Skew: | 0.555 | Prob(JB): | 0.000383 | \n",
"
\n",
"\n",
" Kurtosis: | 3.370 | Cond. No. | 2.43 | \n",
"
\n",
"
Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
],
"text/plain": [
"\n",
"\"\"\"\n",
" OLS Regression Results \n",
"==============================================================================\n",
"Dep. Variable: Dean R-squared: 0.235\n",
"Model: OLS Adj. R-squared: 0.232\n",
"Method: Least Squares F-statistic: 84.21\n",
"Date: Fri, 07 Sep 2018 Prob (F-statistic): 1.11e-17\n",
"Time: 15:26:43 Log-Likelihood: 310.13\n",
"No. Observations: 276 AIC: -616.3\n",
"Df Residuals: 274 BIC: -609.0\n",
"Df Model: 1 \n",
"Covariance Type: nonrobust \n",
"==============================================================================\n",
" coef std err t P>|t| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"Intercept 0.2513 0.006 41.986 0.000 0.240 0.263\n",
"trt 0.0903 0.010 9.177 0.000 0.071 0.110\n",
"==============================================================================\n",
"Omnibus: 14.901 Durbin-Watson: 1.206\n",
"Prob(Omnibus): 0.001 Jarque-Bera (JB): 15.737\n",
"Skew: 0.555 Prob(JB): 0.000383\n",
"Kurtosis: 3.370 Cond. No. 2.43\n",
"==============================================================================\n",
"\n",
"Warnings:\n",
"[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
"\"\"\""
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"smf.ols(formula='Dean ~ trt',data=newhamp).fit().summary()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Matching\n",
"\n",
"Uses code from [match module](http://www.kellieottoboni.com/pscore_match/api/match.html)"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" votesys | \n",
" Obama | \n",
" Clinton | \n",
" dem | \n",
" povrate | \n",
" pci | \n",
" Dean | \n",
" Kerry | \n",
" white | \n",
" absentee | \n",
" population | \n",
" pObama | \n",
" trt | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" D | \n",
" 371 | \n",
" 362 | \n",
" 979 | \n",
" 0.0653 | \n",
" 25940 | \n",
" 0.27820 | \n",
" 0.32030 | \n",
" 0.98312 | \n",
" 0.059857 | \n",
" 4693.0 | \n",
" 0.378958 | \n",
" 0 | \n",
"
\n",
" \n",
" 1 | \n",
" D | \n",
" 345 | \n",
" 333 | \n",
" 913 | \n",
" 0.0380 | \n",
" 19773 | \n",
" 0.24398 | \n",
" 0.36747 | \n",
" 0.97349 | \n",
" 0.050449 | \n",
" 4266.0 | \n",
" 0.377875 | \n",
" 0 | \n",
"
\n",
" \n",
" 2 | \n",
" D | \n",
" 375 | \n",
" 570 | \n",
" 1305 | \n",
" 0.0428 | \n",
" 19986 | \n",
" 0.20096 | \n",
" 0.41627 | \n",
" 0.96739 | \n",
" 0.043649 | \n",
" 7006.0 | \n",
" 0.287356 | \n",
" 0 | \n",
"
\n",
" \n",
" 3 | \n",
" H | \n",
" 92 | \n",
" 89 | \n",
" 268 | \n",
" 0.0669 | \n",
" 25627 | \n",
" 0.28495 | \n",
" 0.33333 | \n",
" 0.97892 | \n",
" 0.107356 | \n",
" 1033.0 | \n",
" 0.343284 | \n",
" 1 | \n",
"
\n",
" \n",
" 4 | \n",
" D | \n",
" 668 | \n",
" 595 | \n",
" 1633 | \n",
" 0.0332 | \n",
" 32667 | \n",
" 0.24937 | \n",
" 0.37781 | \n",
" 0.97986 | \n",
" 0.074706 | \n",
" 7033.0 | \n",
" 0.409063 | \n",
" 0 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" votesys Obama Clinton dem povrate pci Dean Kerry white \\\n",
"0 D 371 362 979 0.0653 25940 0.27820 0.32030 0.98312 \n",
"1 D 345 333 913 0.0380 19773 0.24398 0.36747 0.97349 \n",
"2 D 375 570 1305 0.0428 19986 0.20096 0.41627 0.96739 \n",
"3 H 92 89 268 0.0669 25627 0.28495 0.33333 0.97892 \n",
"4 D 668 595 1633 0.0332 32667 0.24937 0.37781 0.97986 \n",
"\n",
" absentee population pObama trt \n",
"0 0.059857 4693.0 0.378958 0 \n",
"1 0.050449 4266.0 0.377875 0 \n",
"2 0.043649 7006.0 0.287356 0 \n",
"3 0.107356 1033.0 0.343284 1 \n",
"4 0.074706 7033.0 0.409063 0 "
]
},
"execution_count": 15,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from match import Match\n",
"newhamp.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Construct matched pairs"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'{76: 84, 242: 222, 62: 245, 265: 85, 73: 230, 61: 202, 74: 219, 71: 91, 188: 206, 51: 12, 173: 241, 100: 99, 121: 80, 260: 267, 252: 48, 270: 239, 161: 26, 21: 220, 83: 82, 33: 261, 66: 44, 213: 221, 209: 184, 87: 166, 64: 31, 53: 164, 54: 81, 50: 94, 22: 165, 18: 90, 34: 13, 103: 41, 115: 237, 79: 108, 179: 8, 154: 19, 229: 39, 86: 273, 185: 211, 275: 40, 258: 232, 3: 212, 47: 42, 88: 150, 92: 2, 78: 153, 157: 262, 93: 136, 105: 110, 191: 186, 49: 128, 192: 183, 114: 224, 111: 30, 16: 168, 69: 236, 20: 244, 63: 160, 17: 5, 29: 196, 45: 181, 109: 203, 137: 25, 102: 208, 243: 198}'"
]
},
"execution_count": 16,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"np.random.seed(100)\n",
"mp = Match(newhamp.votesys, newhamp.Dean, caliper=0.01)\n",
"str(mp)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Look at a pair. See that Dean voting proportion is very similar."
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" votesys | \n",
" Obama | \n",
" Clinton | \n",
" dem | \n",
" povrate | \n",
" pci | \n",
" Dean | \n",
" Kerry | \n",
" white | \n",
" absentee | \n",
" population | \n",
" pObama | \n",
" trt | \n",
"
\n",
" \n",
" \n",
" \n",
" 51 | \n",
" H | \n",
" 90 | \n",
" 69 | \n",
" 228 | \n",
" 0.0185 | \n",
" 24277 | \n",
" 0.22941 | \n",
" 0.29412 | \n",
" 0.96731 | \n",
" 0.040650 | \n",
" 692.0 | \n",
" 0.394737 | \n",
" 1 | \n",
"
\n",
" \n",
" 12 | \n",
" D | \n",
" 589 | \n",
" 513 | \n",
" 1433 | \n",
" 0.0657 | \n",
" 24867 | \n",
" 0.22903 | \n",
" 0.39114 | \n",
" 0.97644 | \n",
" 0.074593 | \n",
" 6121.0 | \n",
" 0.411026 | \n",
" 0 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" votesys Obama Clinton dem povrate pci Dean Kerry white \\\n",
"51 H 90 69 228 0.0185 24277 0.22941 0.29412 0.96731 \n",
"12 D 589 513 1433 0.0657 24867 0.22903 0.39114 0.97644 \n",
"\n",
" absentee population pObama trt \n",
"51 0.040650 692.0 0.394737 1 \n",
"12 0.074593 6121.0 0.411026 0 "
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"newhamp.iloc[[51,12],]"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"dict_keys([76, 242, 62, 265, 73, 61, 74, 71, 188, 51, 173, 100, 121, 260, 252, 270, 161, 21, 83, 33, 66, 213, 209, 87, 64, 53, 54, 50, 22, 18, 34, 103, 115, 79, 179, 154, 229, 86, 185, 275, 258, 3, 47, 88, 92, 78, 157, 93, 105, 191, 49, 192, 114, 111, 16, 69, 20, 63, 17, 29, 45, 109, 137, 102, 243])"
]
},
"execution_count": 18,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"segs = np.zeros((len(mp), 2, 2), float)\n",
"mp.keys()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Draw line segments for matched pairs. This is much more complicated than the R code. "
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"from matplotlib.collections import LineCollection\n",
"\n",
"fig, ax = plt.subplots()\n",
"segs = np.zeros((len(mp), 2, 2), float)\n",
"segs[:,0,0] = newhamp.Dean.loc[mp.keys()]\n",
"segs[:,1,0] = newhamp.Dean.loc[mp.values()]\n",
"segs[:,0,1] = newhamp.pObama.loc[mp.keys()]\n",
"segs[:,1,1] = newhamp.pObama.loc[mp.values()]\n",
"\n",
"plt.scatter(newhamp.Dean, newhamp.pObama, c=newhamp.trt, s=2)\n",
"line_segments = LineCollection(segs)\n",
"ax.add_collection(line_segments)\n",
"plt.show(ax)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Extract and test the pairwise differences."
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Ttest_1sampResult(statistic=-1.334272985783414, pvalue=0.18684260598001726)"
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pdiff = newhamp.pObama.loc[mp.keys()].ravel() - newhamp.pObama.loc[mp.values()].ravel()\n",
"sp.stats.ttest_1samp(pdiff,0)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"See that these difference do not varying according Dean voting proportion."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Text(0,0.5,'Digital vs. manual difference')"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(newhamp.Dean.loc[mp.keys()], pdiff)\n",
"plt.axhline(0)\n",
"plt.xlabel('Proportion voted for Dean')\n",
"plt.ylabel('Digital vs. manual difference')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Construct plot that demonstrates how difference goes away when covariate adjustment is applied and that is functionally similar to matching."
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fig, ax = plt.subplots()\n",
"plt.scatter(newhamp.Dean, newhamp.pObama, c=newhamp.trt, s=2)\n",
"line_segments = LineCollection(segs)\n",
"ax.add_collection(line_segments)\n",
"plt.axhline(0.353)\n",
"plt.axhline(0.353+0.042,linestyle=\"dashed\",c='red')\n",
"plt.plot([0.1, 0.6],[0.5229*0.1+0.221, 0.5229*0.6+0.221])\n",
"plt.plot([0.1, 0.6],[0.5229*0.1+0.216, 0.5229*0.6+0.216], linestyle=\"dashed\")\n",
"plt.show(ax)"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"data": {
"application/json": {
"Software versions": [
{
"module": "Python",
"version": "3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)]"
},
{
"module": "IPython",
"version": "6.5.0"
},
{
"module": "OS",
"version": "Darwin 17.7.0 x86_64 i386 64bit"
},
{
"module": "pandas",
"version": "0.23.4"
},
{
"module": "numpy",
"version": "1.15.1"
},
{
"module": "matplotlib",
"version": "2.2.3"
},
{
"module": "seaborn",
"version": "0.9.0"
},
{
"module": "scipy",
"version": "1.1.0"
},
{
"module": "patsy",
"version": "0.5.0"
},
{
"module": "statsmodels",
"version": "0.9.0"
}
]
},
"text/html": [
"Software | Version |
---|
Python | 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)] |
IPython | 6.5.0 |
OS | Darwin 17.7.0 x86_64 i386 64bit |
pandas | 0.23.4 |
numpy | 1.15.1 |
matplotlib | 2.2.3 |
seaborn | 0.9.0 |
scipy | 1.1.0 |
patsy | 0.5.0 |
statsmodels | 0.9.0 |
Fri Sep 07 15:26:44 2018 BST |
"
],
"text/latex": [
"\\begin{tabular}{|l|l|}\\hline\n",
"{\\bf Software} & {\\bf Version} \\\\ \\hline\\hline\n",
"Python & 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE\\_401/final)] \\\\ \\hline\n",
"IPython & 6.5.0 \\\\ \\hline\n",
"OS & Darwin 17.7.0 x86\\_64 i386 64bit \\\\ \\hline\n",
"pandas & 0.23.4 \\\\ \\hline\n",
"numpy & 1.15.1 \\\\ \\hline\n",
"matplotlib & 2.2.3 \\\\ \\hline\n",
"seaborn & 0.9.0 \\\\ \\hline\n",
"scipy & 1.1.0 \\\\ \\hline\n",
"patsy & 0.5.0 \\\\ \\hline\n",
"statsmodels & 0.9.0 \\\\ \\hline\n",
"\\hline \\multicolumn{2}{|l|}{Fri Sep 07 15:26:44 2018 BST} \\\\ \\hline\n",
"\\end{tabular}\n"
],
"text/plain": [
"Software versions\n",
"Python 3.7.0 64bit [Clang 4.0.1 (tags/RELEASE_401/final)]\n",
"IPython 6.5.0\n",
"OS Darwin 17.7.0 x86_64 i386 64bit\n",
"pandas 0.23.4\n",
"numpy 1.15.1\n",
"matplotlib 2.2.3\n",
"seaborn 0.9.0\n",
"scipy 1.1.0\n",
"patsy 0.5.0\n",
"statsmodels 0.9.0\n",
"Fri Sep 07 15:26:44 2018 BST"
]
},
"execution_count": 23,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%load_ext version_information\n",
"%version_information pandas, numpy, matplotlib, seaborn, scipy, patsy, statsmodels"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.0"
}
},
"nbformat": 4,
"nbformat_minor": 2
}