Relative_Effectiveness.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import networkx as nx\n",
    "from sklearn.linear_model import LinearRegression\n",
    "import statsmodels.api as sm\n",
    "import os\n",
    "import math\n",
    "import itertools\n",
    "import pickle\n",
    "import seaborn as sns\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read in Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['beta006_multiNPI_0vax_hicom_0.csv',\n",
       " 'beta006_multiNPI_0vax_hicom_1.csv',\n",
       " 'beta006_multiNPI_0vax_hicom_2.csv',\n",
       " 'beta006_multiNPI_0vax_hicom_3.csv',\n",
       " 'beta006_multiNPI_0vax_hicom_4.csv',\n",
       " 'beta006_multiNPI_20vax_hicom_0.csv',\n",
       " 'beta006_multiNPI_20vax_hicom_1.csv',\n",
       " 'beta006_multiNPI_20vax_hicom_2.csv',\n",
       " 'beta006_multiNPI_20vax_hicom_3.csv',\n",
       " 'beta006_multiNPI_20vax_hicom_4.csv',\n",
       " 'beta006_multiNPI_40vax_hicom_0.csv',\n",
       " 'beta006_multiNPI_40vax_hicom_1.csv',\n",
       " 'beta006_multiNPI_40vax_hicom_2.csv',\n",
       " 'beta006_multiNPI_40vax_hicom_3.csv',\n",
       " 'beta006_multiNPI_40vax_hicom_4.csv']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sims = ['beta006_multiNPI_0vax_hicom', 'beta006_multiNPI_20vax_hicom', 'beta006_multiNPI_40vax_hicom']\n",
    "partitions = np.arange(5)\n",
    "sim_names = [sim + '_' + str(part) + '.csv' for sim in sims for part in partitions]\n",
    "sim_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame()\n",
    "for file in sim_names:\n",
    "    tmpdf = pd.read_csv(file, header = 0, dtype = {'rep': int, 'p_vax': float, 'test_freq': int})\n",
    "    part = int(file[-5])\n",
    "    tmpdf['partition'] = np.repeat(part, repeats = tmpdf.shape[0])\n",
    "    df = df.append(tmpdf, ignore_index = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>rep</th>\n",
       "      <th>test_freq</th>\n",
       "      <th>p_sd</th>\n",
       "      <th>p_fm</th>\n",
       "      <th>p_vax</th>\n",
       "      <th>cum_uniq_inf</th>\n",
       "      <th>ext_inf_ct</th>\n",
       "      <th>partition</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>181495</th>\n",
       "      <td>96</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>85</td>\n",
       "      <td>89</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181496</th>\n",
       "      <td>97</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>82</td>\n",
       "      <td>82</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181497</th>\n",
       "      <td>98</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>112</td>\n",
       "      <td>111</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181498</th>\n",
       "      <td>99</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>91</td>\n",
       "      <td>92</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181499</th>\n",
       "      <td>100</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>97</td>\n",
       "      <td>95</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        rep  test_freq  p_sd  p_fm  p_vax  cum_uniq_inf  ext_inf_ct  partition\n",
       "181495   96         28   1.0   1.0    0.4            85          89          4\n",
       "181496   97         28   1.0   1.0    0.4            82          82          4\n",
       "181497   98         28   1.0   1.0    0.4           112         111          4\n",
       "181498   99         28   1.0   1.0    0.4            91          92          4\n",
       "181499  100         28   1.0   1.0    0.4            97          95          4"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Renumber reps to remove duplicates\n",
    "df['rep'] = df['rep'] + df['partition']*20\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fit OLS Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# If no testing is included, it has to be treated as categorical since 0 is worst followed by 28, but 3 is best\n",
    "testdf = df.loc[df.test_freq!=0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>test_freq</th>\n",
       "      <th>p_sd</th>\n",
       "      <th>p_fm</th>\n",
       "      <th>p_vax</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>181495</th>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181496</th>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181497</th>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181498</th>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181499</th>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        test_freq  p_sd  p_fm  p_vax\n",
       "181495         28   1.0   1.0    0.4\n",
       "181496         28   1.0   1.0    0.4\n",
       "181497         28   1.0   1.0    0.4\n",
       "181498         28   1.0   1.0    0.4\n",
       "181499         28   1.0   1.0    0.4"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = testdf[['test_freq', 'p_sd', 'p_fm', 'p_vax']]\n",
    "X.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = testdf['cum_uniq_inf']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:           cum_uniq_inf   R-squared:                       0.670\n",
      "Model:                            OLS   Adj. R-squared:                  0.670\n",
      "Method:                 Least Squares   F-statistic:                 7.369e+04\n",
      "Date:                Wed, 26 May 2021   Prob (F-statistic):               0.00\n",
      "Time:                        09:01:31   Log-Likelihood:            -6.7770e+05\n",
      "No. Observations:              145200   AIC:                         1.355e+06\n",
      "Df Residuals:                  145195   BIC:                         1.355e+06\n",
      "Df Model:                           4                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "const        213.8188      0.207   1033.838      0.000     213.413     214.224\n",
      "test_freq      0.5646      0.007     79.480      0.000       0.551       0.578\n",
      "p_sd         -45.6743      0.214   -213.741      0.000     -46.093     -45.255\n",
      "p_fm         -57.5501      0.214   -269.316      0.000     -57.969     -57.131\n",
      "p_vax       -170.7266      0.414   -412.574      0.000    -171.538    -169.916\n",
      "==============================================================================\n",
      "Omnibus:                    69956.954   Durbin-Watson:                   0.623\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):           677577.570\n",
      "Skew:                           2.099   Prob(JB):                         0.00\n",
      "Kurtosis:                      12.714   Cond. No.                         101.\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "X_train_Sm = sm.add_constant(X)\n",
    "ls = sm.OLS(y,X_train_Sm).fit()\n",
    "print(ls.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* For every 10% increase in social distancing, cumulative incidence declines by 45.6743/10 = 4.56743\n",
    "* For every 10% increase in mask wearing, cumulative incidence declines on average by 57.5501/10 = 5.75501\n",
    "* For every 10% increase in immunity, cumulative incidence declines on average by 170.7266/10 = 17.07266\n",
    "* After controlling for these other factors, for every one day increase in testing frequency, cumulative incidence increases by 0.5646 on average. Conversely, for every one day decrease in testing frequency, cumulative incidence decreases by 0.5646 on average."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7.912557562876374"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "4.46743/0.5646"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.193074743181015"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "5.75501/0.5646"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "30.238505136379736"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "17.07266/0.5646"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* A 10% increase in social distancing is equivalent to testing 7.9 days more frequently.\n",
    "* A 10% increase in masking is equivalent to testing 10.2 days more frequently.\n",
    "* A 10% increase in immunity is equivalent to testing 30.0 days more frequently."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.11303653661251963"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(4.46743/(7*0.5646))*0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.1456153534740145"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(5.75501/(7*0.5646))*0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.43197864480542486"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(17.07266/(7*0.5646))*0.1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Moving from testing once every 14 days to once every 7 days is equivalent to:\n",
    "* 11.4% increase in social distancing\n",
    "* 14.3% increase in mask wearing\n",
    "* 42.7% increase in immunity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "redux = 10\n",
    "tcoeff = 0.5646\n",
    "sdcoeff = 4.46743\n",
    "fmcoeff = 5.75501\n",
    "vxcoeff = 17.07266"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "17.711654268508678\n",
      "0.22384234336072417\n"
     ]
    }
   ],
   "source": [
    "print(redux/tcoeff)\n",
    "print((redux/sdcoeff)*0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.1737616442021821\n",
      "0.05857318074629262\n"
     ]
    }
   ],
   "source": [
    "print((redux/fmcoeff)*0.1)\n",
    "print((redux/vxcoeff)*0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To reduce cumulative incidence by 10 people (1.5% of student population), you need to:\n",
    "* increase testing frequency by 17.7 days\n",
    "* increase social distancing by 22.4%\n",
    "* increase masking by 17.4%\n",
    "* increase immunity by 5.9%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, axs = plt.subplots(1, 4, figsize = (20, 10))\n",
    "for (i,col) in enumerate(X):\n",
    "    axs[i].scatter(testdf[col], testdf['cum_uniq_inf'])\n",
    "    axs[i].set_xlabel(col)\n",
    "    axs[i].set_ylabel('cumulative incidence')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}