{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# R tutorial for [*Lung cancer incidence decreases with elevation: evidence for oxygen as an inhaled carcinogen*](https://doi.org/10.7717/peerj.705)\n",
    "\n",
    "This tutorial by Daniel Himmelstein recreates some of the main findings from our study. It is written in R. This file is a Jupyter notebook, but you can copy invidual cells into an interactive R terminal."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Packages\n",
    "\n",
    "In general, we use the double colon operator to access functions from packages. So for example to specify the `read_tsv` function in the `readr` package, we use `readr::read_tsv()`. This convention avoids polluting our namespace and clarifies function provenance.\n",
    "\n",
    "This notebook will make use of the following packages, which you may need to install:\n",
    "\n",
    "+ dplyr\n",
    "+ readr\n",
    "+ ggplot2\n",
    "+ glmnet\n",
    "+ broom\n",
    "\n",
    "We do oad dplyr explicitly in order to attach the [pipe operator](http://www.r-statistics.com/2014/08/simpler-r-coding-with-pipes-the-present-and-future-of-the-magrittr-package/) (`%>%`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "library(dplyr, warn = FALSE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read and process data\n",
    "\n",
    "Here we read the county-level dataset compiled for our study. Each row is a county in the United States."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>name</th><th scope=col>state</th><th scope=col>fips</th><th scope=col>metro</th><th scope=col>immigration</th><th scope=col>native</th><th scope=col>white</th><th scope=col>black</th><th scope=col>male</th><th scope=col>population</th><th scope=col>elevation</th><th scope=col>high_temp</th><th scope=col>low_temp</th><th scope=col>diurnal_temp</th><th scope=col>particulate</th><th scope=col>precipitation</th><th scope=col>sunlight</th><th scope=col>uvb</th><th scope=col>smoking</th><th scope=col>smoking_lower</th><th scope=col>smoking_upper</th><th scope=col>female_smoking</th><th scope=col>male_smoking</th><th scope=col>radon</th><th scope=col>drinking</th><th scope=col>meat</th><th scope=col>diabetes</th><th scope=col>obesity</th><th scope=col>education</th><th scope=col>income</th><th scope=col>mammogram</th><th scope=col>all_cancer</th><th scope=col>all_cancer_male</th><th scope=col>all_cancer_female</th><th scope=col>colorectal</th><th scope=col>colorectal_lower</th><th scope=col>colorectal_upper</th><th scope=col>breast</th><th scope=col>breast_lower</th><th scope=col>breast_upper</th><th scope=col>prostate</th><th scope=col>prostate_lower</th><th scope=col>prostate_upper</th><th scope=col>lung</th><th scope=col>lung_lower</th><th scope=col>lung_upper</th><th scope=col>female_lung</th><th scope=col>male_lung</th><th scope=col>over_65_lung</th><th scope=col>under_65_lung</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>1</th><td>Autauga, AL</td><td>AL</td><td>01001</td><td>1</td><td>25.30968</td><td>0.9</td><td>81.5</td><td>17.3</td><td>48.59289</td><td>43671</td><td>0.11012</td><td>22.82228</td><td>12.95648</td><td>9.865793</td><td>15.41521</td><td>3.703882</td><td>16875.17</td><td>1180.577</td><td>48.6</td><td>43.1</td><td>54.1</td><td>41.05</td><td>56.9</td><td>0.74</td><td>14</td><td>74</td><td>10.86</td><td>29.8</td><td>21.7</td><td>53.255</td><td>69.7</td><td>460.6</td><td>548.1</td><td>407.2</td><td>57.1</td><td>47.6</td><td>67.8</td><td>117</td><td>99.3</td><td>137</td><td>119</td><td>97.5</td><td>143.6</td><td>81.5</td><td>70.2</td><td>94.1</td><td>55.8</td><td>119.3</td><td>398.1</td><td>35.7</td></tr>\n",
       "\t<tr><th scope=row>2</th><td>Baldwin, AL</td><td>AL</td><td>01003</td><td>0</td><td>24.94654</td><td>1.1</td><td>88.1</td><td>10.5</td><td>49.0318</td><td>140415</td><td>0.04018</td><td>23.87101</td><td>15.95164</td><td>7.919362</td><td>13.35193</td><td>4.454666</td><td>17407.23</td><td>1256.903</td><td>49.9</td><td>45</td><td>54.8</td><td>43.75</td><td>56.6</td><td>0.58</td><td>18</td><td>69</td><td>7.96</td><td>25.9</td><td>26.8</td><td>50.147</td><td>69.9</td><td>444.9</td><td>527.7</td><td>374.8</td><td>40.7</td><td>36.9</td><td>44.7</td><td>120.2</td><td>111.1</td><td>129.9</td><td>155.1</td><td>144.7</td><td>166.2</td><td>67.9</td><td>63.1</td><td>73</td><td>55</td><td>82.8</td><td>363.7</td><td>25.1</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|llllllllllllllllllllllllllllllllllllllllllllllllll}\n",
       "  & name & state & fips & metro & immigration & native & white & black & male & population & elevation & high_temp & low_temp & diurnal_temp & particulate & precipitation & sunlight & uvb & smoking & smoking_lower & smoking_upper & female_smoking & male_smoking & radon & drinking & meat & diabetes & obesity & education & income & mammogram & all_cancer & all_cancer_male & all_cancer_female & colorectal & colorectal_lower & colorectal_upper & breast & breast_lower & breast_upper & prostate & prostate_lower & prostate_upper & lung & lung_lower & lung_upper & female_lung & male_lung & over_65_lung & under_65_lung\\\\\n",
       "\\hline\n",
       "\t1 & Autauga, AL & AL & 01001 & 1 & 25.30968 & 0.9 & 81.5 & 17.3 & 48.59289 & 43671 & 0.11012 & 22.82228 & 12.95648 & 9.865793 & 15.41521 & 3.703882 & 16875.17 & 1180.577 & 48.6 & 43.1 & 54.1 & 41.05 & 56.9 & 0.74 & 14 & 74 & 10.86 & 29.8 & 21.7 & 53.255 & 69.7 & 460.6 & 548.1 & 407.2 & 57.1 & 47.6 & 67.8 & 117 & 99.3 & 137 & 119 & 97.5 & 143.6 & 81.5 & 70.2 & 94.1 & 55.8 & 119.3 & 398.1 & 35.7\\\\\n",
       "\t2 & Baldwin, AL & AL & 01003 & 0 & 24.94654 & 1.1 & 88.1 & 10.5 & 49.0318 & 140415 & 0.04018 & 23.87101 & 15.95164 & 7.919362 & 13.35193 & 4.454666 & 17407.23 & 1256.903 & 49.9 & 45 & 54.8 & 43.75 & 56.6 & 0.58 & 18 & 69 & 7.96 & 25.9 & 26.8 & 50.147 & 69.9 & 444.9 & 527.7 & 374.8 & 40.7 & 36.9 & 44.7 & 120.2 & 111.1 & 129.9 & 155.1 & 144.7 & 166.2 & 67.9 & 63.1 & 73 & 55 & 82.8 & 363.7 & 25.1\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/plain": [
       "Source: local data frame [2 x 50]\n",
       "\n",
       "         name state  fips metro immigration native white black     male\n",
       "        (chr) (chr) (chr) (int)       (dbl)  (dbl) (dbl) (dbl)    (dbl)\n",
       "1 Autauga, AL    AL 01001     1    25.30968    0.9  81.5  17.3 48.59289\n",
       "2 Baldwin, AL    AL 01003     0    24.94654    1.1  88.1  10.5 49.03180\n",
       "Variables not shown: population (int), elevation (dbl), high_temp (dbl),\n",
       "  low_temp (dbl), diurnal_temp (dbl), particulate (dbl), precipitation (dbl),\n",
       "  sunlight (dbl), uvb (dbl), smoking (dbl), smoking_lower (dbl), smoking_upper\n",
       "  (dbl), female_smoking (dbl), male_smoking (dbl), radon (dbl), drinking (int),\n",
       "  meat (int), diabetes (dbl), obesity (dbl), education (dbl), income (dbl),\n",
       "  mammogram (dbl), all_cancer (dbl), all_cancer_male (dbl), all_cancer_female\n",
       "  (dbl), colorectal (dbl), colorectal_lower (dbl), colorectal_upper (dbl),\n",
       "  breast (dbl), breast_lower (dbl), breast_upper (dbl), prostate (dbl),\n",
       "  prostate_lower (dbl), prostate_upper (dbl), lung (dbl), lung_lower (dbl),\n",
       "  lung_upper (dbl), female_lung (dbl), male_lung (dbl), over_65_lung (dbl),\n",
       "  under_65_lung (dbl)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = 'https://github.com/dhimmel/elevcan/raw/0330d37345c568250b05aedc3161183589a1edba/data/county-data.txt'\n",
    "county_df = readr::read_tsv(url)\n",
    "head(county_df, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "3142"
      ],
      "text/latex": [
       "3142"
      ],
      "text/markdown": [
       "3142"
      ],
      "text/plain": [
       "[1] 3142"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# number of counties\n",
    "nrow(county_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Filter counties\n",
    "\n",
    "Our analysis focused on 10 states that varied substantially in elevation. When designing an experiment, always try to maximize the variation in your variable of interest. We [filtered many counties](https://peerj.com/articles/705/#p-17) for quality control reasons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "259"
      ],
      "text/latex": [
       "259"
      ],
      "text/markdown": [
       "259"
      ],
      "text/plain": [
       "[1] 259"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "states <- c('AZ', 'CA', 'CO', 'ID', 'MT', 'NV', 'NM', 'OR', 'UT', 'WA', 'WY')\n",
    "\n",
    "county_df = county_df %>%\n",
    "  dplyr::filter(state %in% states) %>%\n",
    "  dplyr::filter(population >= 1e4) %>%\n",
    "  dplyr::filter(native <= 25) %>%\n",
    "  dplyr::filter(immigration <= 40)\n",
    "\n",
    "nrow(county_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Add additional variables\n",
    "\n",
    "We calculate a weight for each county which we use as weights in the regression. The weights account for the increased confidence we have in the measurements for high-population counties. We also calculate the rate of all other cancers besides lung (`no_lung`) to use as a predictor in our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "county_df = county_df %>%\n",
    "  dplyr::mutate(no_lung = all_cancer - lung) %>%\n",
    "  dplyr::mutate(weight = pmin(sqrt(population), 500))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create a lung cancer specific dataset\n",
    "\n",
    "We identified 11 covariates in addition to elevation to assess for lung cancer. Below we create a dataset with only the relevant variables and remove counties with any missing data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "253"
      ],
      "text/latex": [
       "253"
      ],
      "text/markdown": [
       "253"
      ],
      "text/plain": [
       "[1] 253"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "covariates = c(\n",
    "  'metro',\n",
    "  'white',\n",
    "  'black',\n",
    "  'education',\n",
    "  'income',\n",
    "  'obesity',\n",
    "  'no_lung',\n",
    "  'male',\n",
    "  'smoking',\n",
    "  'radon',\n",
    "  'particulate')\n",
    "\n",
    "lung_df = county_df %>%\n",
    "  dplyr::select(name, lung, elevation, one_of(covariates), weight) %>%\n",
    "  na.omit()\n",
    "\n",
    "nrow(lung_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot lung cancer versus elevation\n",
    "\n",
    "This bivariate scatterplot shows the same data as [Figure 4A](https://doi.org/10.7717/peerj.705/fig-4)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {
      "image/svg+xml": {
       "isolated": true
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lung_df %>%\n",
    "  ggplot2::ggplot(ggplot2::aes(x = elevation, y = lung)) +\n",
    "  ggplot2::geom_smooth(ggplot2::aes(weight = weight), method = 'lm') + \n",
    "  ggplot2::geom_point(ggplot2::aes(alpha = weight)) +\n",
    "  ggplot2::xlab('Elevation (meters)') + ggplot2::ylab('Lung cancer Incidence') +\n",
    "  ggplot2::theme_bw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best-subset linear model\n",
    "\n",
    "We used best-subset regression to evaluate all possible predictor combinations. Here we take the predictors from the BIC-minimizing model and fit the linear model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Fit the model\n",
    "model = lm('lung ~ smoking + elevation + no_lung + black + education', data = lung_df, weight = weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>term</th><th scope=col>estimate</th><th scope=col>std.error</th><th scope=col>statistic</th><th scope=col>p.value</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>1</th><td>(Intercept)</td><td>-14.85408</td><td>6.779758</td><td>-2.190945</td><td>0.0293905</td></tr>\n",
       "\t<tr><th scope=row>2</th><td>smoking</td><td>1.440257</td><td>0.09678763</td><td>14.88059</td><td>3.460723e-36</td></tr>\n",
       "\t<tr><th scope=row>3</th><td>elevation</td><td>-7.234103</td><td>0.793121</td><td>-9.121059</td><td>2.688957e-17</td></tr>\n",
       "\t<tr><th scope=row>4</th><td>no_lung</td><td>0.05112451</td><td>0.01708235</td><td>2.992826</td><td>0.003044367</td></tr>\n",
       "\t<tr><th scope=row>5</th><td>black</td><td>0.6709068</td><td>0.1719983</td><td>3.90066</td><td>0.000123714</td></tr>\n",
       "\t<tr><th scope=row>6</th><td>education</td><td>-0.4334022</td><td>0.05503473</td><td>-7.875067</td><td>1.077715e-13</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|lllll}\n",
       "  & term & estimate & std.error & statistic & p.value\\\\\n",
       "\\hline\n",
       "\t1 & (Intercept) & -14.85408 & 6.779758 & -2.190945 & 0.0293905\\\\\n",
       "\t2 & smoking & 1.440257 & 0.09678763 & 14.88059 & 3.460723e-36\\\\\n",
       "\t3 & elevation & -7.234103 & 0.793121 & -9.121059 & 2.688957e-17\\\\\n",
       "\t4 & no_lung & 0.05112451 & 0.01708235 & 2.992826 & 0.003044367\\\\\n",
       "\t5 & black & 0.6709068 & 0.1719983 & 3.90066 & 0.000123714\\\\\n",
       "\t6 & education & -0.4334022 & 0.05503473 & -7.875067 & 1.077715e-13\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/plain": [
       "         term     estimate  std.error statistic      p.value\n",
       "1 (Intercept) -14.85407748 6.77975797 -2.190945 2.939050e-02\n",
       "2     smoking   1.44025690 0.09678763 14.880588 3.460723e-36\n",
       "3   elevation  -7.23410337 0.79312100 -9.121059 2.688957e-17\n",
       "4     no_lung   0.05112451 0.01708235  2.992826 3.044367e-03\n",
       "5       black   0.67090677 0.17199826  3.900660 1.237140e-04\n",
       "6   education  -0.43340219 0.05503473 -7.875067 1.077715e-13"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# View the model coefficients\n",
    "coef_df = broom::tidy(model)\n",
    "coef_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>r.squared</th><th scope=col>adj.r.squared</th><th scope=col>sigma</th><th scope=col>statistic</th><th scope=col>p.value</th><th scope=col>df</th><th scope=col>logLik</th><th scope=col>AIC</th><th scope=col>BIC</th><th scope=col>deviance</th><th scope=col>df.residual</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>1</th><td>0.7024248</td><td>0.6964011</td><td>131.2217</td><td>116.6085</td><td>6.049589e-63</td><td>6</td><td>-897.7659</td><td>1809.532</td><td>1834.266</td><td>4253129</td><td>247</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|lllllllllll}\n",
       "  & r.squared & adj.r.squared & sigma & statistic & p.value & df & logLik & AIC & BIC & deviance & df.residual\\\\\n",
       "\\hline\n",
       "\t1 & 0.7024248 & 0.6964011 & 131.2217 & 116.6085 & 6.049589e-63 & 6 & -897.7659 & 1809.532 & 1834.266 & 4253129 & 247\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/plain": [
       "  r.squared adj.r.squared    sigma statistic      p.value df    logLik      AIC\n",
       "1 0.7024248     0.6964011 131.2217  116.6085 6.049589e-63  6 -897.7659 1809.532\n",
       "       BIC deviance df.residual\n",
       "1 1834.266  4253129         247"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Summarize the overall model fit\n",
    "broom::glance(model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Compute standardized coefficients\n",
    "\n",
    "Standardized coefficients allow us to compare effect sizes between variables. These standardized coefficients are reported in [Figure 3A](https://doi.org/10.7717/peerj.705/fig-3)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "weighted.sd <- function(x, w) {\n",
    "  # Weighted standard deviation\n",
    "  # http://stats.stackexchange.com/a/6536/74908\n",
    "  numerator = sum(w * (x - weighted.mean(x, w)) ^ 2)\n",
    "  M = sum(w != 0)\n",
    "  denominator = (M - 1) / M * sum(w)\n",
    "  return(sqrt(numerator / denominator))\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<dl class=dl-horizontal>\n",
       "\t<dt>smoking</dt>\n",
       "\t\t<dd>0.585479432787212</dd>\n",
       "\t<dt>elevation</dt>\n",
       "\t\t<dd>-0.354751971504714</dd>\n",
       "\t<dt>no_lung</dt>\n",
       "\t\t<dd>0.121633937567613</dd>\n",
       "\t<dt>black</dt>\n",
       "\t\t<dd>0.148946139270269</dd>\n",
       "\t<dt>education</dt>\n",
       "\t\t<dd>-0.303981290117252</dd>\n",
       "</dl>\n"
      ],
      "text/latex": [
       "\\begin{description*}\n",
       "\\item[smoking] 0.585479432787212\n",
       "\\item[elevation] -0.354751971504714\n",
       "\\item[no_lung] 0.121633937567613\n",
       "\\item[black] 0.148946139270269\n",
       "\\item[education] -0.303981290117252\n",
       "\\end{description*}\n"
      ],
      "text/markdown": [
       "smoking\n",
       ":   0.585479432787212elevation\n",
       ":   -0.354751971504714no_lung\n",
       ":   0.121633937567613black\n",
       ":   0.148946139270269education\n",
       ":   -0.303981290117252\n",
       "\n"
      ],
      "text/plain": [
       "   smoking  elevation    no_lung      black  education \n",
       " 0.5854794 -0.3547520  0.1216339  0.1489461 -0.3039813 "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sd_y = weighted.sd(lung_df$lung, lung_df$weight)\n",
    "sd_x = apply(lung_df[coef_df$term[-1]], 2, weighted.sd, w = lung_df$weight)\n",
    "coef_df$estimate[-1] * sd_x / sd_y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Lasso model\n",
    "\n",
    "Since the best-subset approach is prone to overfitting, we also fit a lasso model. The lasso model uses cross-validation to find an optimal penalty of complexity, which in effect prevents overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "x_matrix = lung_df %>%\n",
    "  dplyr::select(elevation, one_of(covariates)) %>%\n",
    "  as.matrix()\n",
    "y = lung_df$lung\n",
    "w = lung_df$weight\n",
    "\n",
    "# Fit the lasso using cross validation\n",
    "lasso = glmnet::cv.glmnet(x = x_matrix, y = y, weights = w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13 x 1 sparse Matrix of class \"dgCMatrix\"\n",
       "                      1\n",
       "(Intercept)  0.72552620\n",
       "elevation   -6.68174482\n",
       "metro        .         \n",
       "white        .         \n",
       "black        0.24769996\n",
       "education   -0.30998886\n",
       "income       .         \n",
       "obesity      .         \n",
       "no_lung      0.02252882\n",
       "male         .         \n",
       "smoking      1.29573258\n",
       "radon       -0.19956167\n",
       "particulate  .         "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Coefficients from the lasso model\n",
    "coef(lasso, s = 'lambda.1se')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate lasso R-squared"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "calculate_R2 <- function(y, y_pred, w) {\n",
    "  # Calculate R2 using the Sum of Squares method\n",
    "  # https://en.wikipedia.org/wiki/Coefficient_of_determination#Definitions\n",
    "  ss_residual = sum(w * (y - y_pred) ^ 2)\n",
    "  ss_total = sum(w * (y - weighted.mean(y, w)) ^ 2)\n",
    "  return(1 - ss_residual / ss_total)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "0.680556515726018"
      ],
      "text/latex": [
       "0.680556515726018"
      ],
      "text/markdown": [
       "0.680556515726018"
      ],
      "text/plain": [
       "[1] 0.6805565"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = predict(lasso, s = 'lambda.1se', newx = x_matrix)[, 1]\n",
    "calculate_R2(y, y_pred, w)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected the lasso R-squared is lower than the best-subset R-squared. However, the difference is minimal suggesting overfitting was not a major problem."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Session info\n",
    "\n",
    "Added for reproducibility."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "R version 3.2.2 (2015-08-14)\n",
       "Platform: x86_64-apple-darwin11.4.2 (64-bit)\n",
       "Running under: OS X 10.11.3 (El Capitan)\n",
       "\n",
       "locale:\n",
       "[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8\n",
       "\n",
       "attached base packages:\n",
       "[1] stats     graphics  grDevices utils     datasets  methods   base     \n",
       "\n",
       "other attached packages:\n",
       "[1] dplyr_0.4.3\n",
       "\n",
       "loaded via a namespace (and not attached):\n",
       " [1] Rcpp_0.12.2      plyr_1.8.3       iterators_1.0.8  base64enc_0.1-3 \n",
       " [5] tools_3.2.2      digest_0.6.8     uuid_0.1-2       jsonlite_0.9.17 \n",
       " [9] evaluate_0.8     gtable_0.1.2     nlme_3.1-122     lattice_0.20-33 \n",
       "[13] Matrix_1.2-2     psych_1.5.8      foreach_1.4.3    IRdisplay_0.3   \n",
       "[17] DBI_0.3.1        curl_0.9.5       IRkernel_0.5     parallel_3.2.2  \n",
       "[21] proto_0.3-10     rzmq_0.7.7       repr_0.3         stringr_1.0.0   \n",
       "[25] grid_3.2.2       glmnet_2.0-2     R6_2.1.1         ggplot2_1.0.1   \n",
       "[29] reshape2_1.4.1   readr_0.2.2      tidyr_0.3.1      magrittr_1.5    \n",
       "[33] codetools_0.2-14 scales_0.3.0     MASS_7.3-45      assertthat_0.1  \n",
       "[37] mnormt_1.5-3     colorspace_1.2-6 labeling_0.3     stringi_1.0-1   \n",
       "[41] lazyeval_0.1.10  munsell_0.4.2    broom_0.4.0     "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sessionInfo()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "R",
   "language": "R",
   "name": "ir"
  },
  "language_info": {
   "codemirror_mode": "r",
   "file_extension": ".r",
   "mimetype": "text/x-r-source",
   "name": "R",
   "pygments_lexer": "r",
   "version": "3.2.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}