patient-zero.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import pickle as pkl\n",
    "from itertools import product\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error\n",
    "from tqdm import tqdm\n",
    "from matplotlib import pyplot as plt\n",
    "from datetime import timedelta\n",
    "from joblib import delayed, Parallel\n",
    "from copy import copy\n",
    "\n",
    "from core import  do_simulation\n",
    "from helpers import Params, T, get_T1_and_T2, R0, plot_total, save_bundle, save_to_json, makedir_if_not_there\n",
    "\n",
    "from const import STATE, COLORS, NUM_STATES,  STATES\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.999881671966743"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tran_coef_before = np.array([4.28, 5.28, 5.33, 4.38, 4.13, 5.27, 5.28, 5.32, 5.15, 5.22, 4.46, 4.20, 5.17, 5.11, 4.91, 4.92, 4.87, 4.04, 4.32, 4.22 , 3.60, 2.88])\n",
    "tran_coef_after = np.array([0.66 ,0.67 ,0.68 ,0.66 ,0.69 ,0.69 ,0.66 ,0.69 ,0.68 ,0.67 ,0.60 ,0.62 ,0.61 ,0.62 ,0.62 ,0.61 ,0.60 ,0.62 ,0.67 ,0.60 ,0.58 ,0.59 ,0.58 ,0.57 ,0.59 ,0.60 ,0.57 ])\n",
    "infection_factor = np.mean(tran_coef_before) / np.mean(tran_coef_after)\n",
    "infection_factor = np.log(infection_factor)\n",
    "infection_factor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "params_jan27 = pkl.load(\n",
    "    open('output/params_after_lockdown.pkl', 'rb')\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "total_population: 9000000\n",
       "initial_num_E: 5000.0\n",
       "initial_num_I: 5000\n",
       "initial_num_M: 4000.0\n",
       "\n",
       "alpha: 1.5999999999999998e-08\n",
       "beta:  1.8e-09\n",
       "\n",
       "mu_ei: 6\n",
       "mu_mo: 10\n",
       "\n",
       "x0_pt: 12000\n",
       "k_pt:  0.00014931328910233792\n",
       "mean_IM: 7\n",
       "\n",
       "k_days: 14\n",
       "        "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "params_jan27"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "datetime.datetime(2020, 1, 23, 0, 0)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lockdown_date = T('23/01/2020')\n",
    "lockdown_date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([5000.]), array([5000]), array([4000.]))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target_E = params_jan27.initial_num_E\n",
    "target_I = params_jan27.initial_num_I\n",
    "target_M = params_jan27.initial_num_M\n",
    "\n",
    "actual_I = np.array([target_I])\n",
    "actual_E = np.array([target_E])\n",
    "actual_M = np.array([target_M])\n",
    "\n",
    "actual_E, actual_I, actual_M"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepare_params(t):\n",
    "    p = Params(\n",
    "        alpha=[(0, infection_factor * params_jan27.alpha), (t, params_jan27.alpha)],\n",
    "        beta=[(0, infection_factor * params_jan27.beta), (t, params_jan27.beta)],\n",
    "        stages=[t],\n",
    "        initial_num_E=1,\n",
    "        initial_num_I=0,\n",
    "        initial_num_M=0,\n",
    "        mu_ei=params_jan27.mu_ei,\n",
    "        mu_mo=params_jan27.mu_mo,\n",
    "        x0_pt=params_jan27.x0_pt,\n",
    "        k_days=params_jan27.k_days\n",
    "    )\n",
    "    return p\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "bed_info = [(0, 3000)]\n",
    "def one_run(t):\n",
    "    params = prepare_params(t)\n",
    "    # t days **before** lockdown (Jan 23)\n",
    "    # simulation finishes at Jan 27 (after lockdown for 5 days)\n",
    "    # t=1, 2, 3, ... means patient zero  appeared in Jan 22, 21, 20\n",
    "    p0_time = lockdown_date - timedelta(days=t)\n",
    "    \n",
    "    total, _, _, _, stats = do_simulation(\n",
    "        t+5, bed_info, params, p0_time=p0_time,\n",
    "        verbose=0\n",
    "    )\n",
    "    \n",
    "    pred_I = np.array([total[-1, STATE.I]])\n",
    "    pred_E = np.array([total[-1, STATE.E]])\n",
    "    pred_M = np.array([total[-1, STATE.M]])\n",
    "    \n",
    "    mse_I = mean_absolute_error(actual_I,  pred_I)\n",
    "    mse_E = mean_absolute_error(actual_E, pred_E)\n",
    "    mse_M = mean_absolute_error(actual_M, pred_M)\n",
    "    mse_IM = mean_absolute_error(actual_M + actual_I, pred_M + pred_I)\n",
    "    mse_IEM = mean_absolute_error(actual_I + actual_E + actual_M,  pred_I + pred_E + pred_M)\n",
    "    return (t, actual_I[0], pred_I[0], mse_I, mse_E, mse_M, mse_IM, mse_IEM, stats)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 31/31 [00:04<00:00,  6.25it/s]\n"
     ]
    }
   ],
   "source": [
    "# t is the number of days back\n",
    "rows = Parallel(n_jobs=-1)(delayed(one_run)(t) for t in tqdm(range(30, 61)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(rows, columns=('t', 'actual_I', 'pred_I', 'mse_I', 'mse_E', 'mse_M', 'mse_IM', 'mse_IEM', 'r0_info'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>t</th>\n",
       "      <th>actual_I</th>\n",
       "      <th>pred_I</th>\n",
       "      <th>mse_I</th>\n",
       "      <th>mse_E</th>\n",
       "      <th>mse_M</th>\n",
       "      <th>mse_IM</th>\n",
       "      <th>mse_IEM</th>\n",
       "      <th>r0_info</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>53</td>\n",
       "      <td>5000</td>\n",
       "      <td>5044.748507</td>\n",
       "      <td>44.748507</td>\n",
       "      <td>1552.612290</td>\n",
       "      <td>1000.000000</td>\n",
       "      <td>955.251493</td>\n",
       "      <td>597.360797</td>\n",
       "      <td>{'R0_by_stage': {0: (5.1215123171320585, 3.150...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>52</td>\n",
       "      <td>5000</td>\n",
       "      <td>3885.334894</td>\n",
       "      <td>1114.665106</td>\n",
       "      <td>558.921346</td>\n",
       "      <td>1062.156712</td>\n",
       "      <td>2176.821818</td>\n",
       "      <td>1617.900472</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121544198771868, 3.1398...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>54</td>\n",
       "      <td>5000</td>\n",
       "      <td>6687.138370</td>\n",
       "      <td>1687.138370</td>\n",
       "      <td>2731.283735</td>\n",
       "      <td>1201.733980</td>\n",
       "      <td>485.404390</td>\n",
       "      <td>3216.688125</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121466025584044, 3.1624...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>51</td>\n",
       "      <td>5000</td>\n",
       "      <td>3243.199885</td>\n",
       "      <td>1756.800115</td>\n",
       "      <td>283.687985</td>\n",
       "      <td>1455.399054</td>\n",
       "      <td>3212.199168</td>\n",
       "      <td>3495.887153</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121562154320443, 3.1307...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>50</td>\n",
       "      <td>5000</td>\n",
       "      <td>2712.188405</td>\n",
       "      <td>2287.811595</td>\n",
       "      <td>998.276604</td>\n",
       "      <td>1802.128329</td>\n",
       "      <td>4089.939924</td>\n",
       "      <td>5088.216527</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121566395321335, 3.1230...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>49</td>\n",
       "      <td>5000</td>\n",
       "      <td>2272.050067</td>\n",
       "      <td>2727.949933</td>\n",
       "      <td>1604.367105</td>\n",
       "      <td>2106.228703</td>\n",
       "      <td>4834.178636</td>\n",
       "      <td>6438.545741</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121556838562287, 3.1163...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>48</td>\n",
       "      <td>5000</td>\n",
       "      <td>1906.385945</td>\n",
       "      <td>3093.614055</td>\n",
       "      <td>2118.489290</td>\n",
       "      <td>2371.707795</td>\n",
       "      <td>5465.321850</td>\n",
       "      <td>7583.811140</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121533083188067, 3.1105...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>55</td>\n",
       "      <td>5000</td>\n",
       "      <td>8184.867009</td>\n",
       "      <td>3184.867009</td>\n",
       "      <td>4132.860163</td>\n",
       "      <td>1000.000000</td>\n",
       "      <td>2184.867009</td>\n",
       "      <td>6317.727172</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121404593397314, 3.1765...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>47</td>\n",
       "      <td>5000</td>\n",
       "      <td>1601.907769</td>\n",
       "      <td>3398.092231</td>\n",
       "      <td>2554.641722</td>\n",
       "      <td>2602.539226</td>\n",
       "      <td>6000.631457</td>\n",
       "      <td>8555.273178</td>\n",
       "      <td>{'R0_by_stage': {0: (5.121494384772303, 3.1055...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>46</td>\n",
       "      <td>5000</td>\n",
       "      <td>1347.832039</td>\n",
       "      <td>3652.167961</td>\n",
       "      <td>2924.681901</td>\n",
       "      <td>2802.547335</td>\n",
       "      <td>6454.715296</td>\n",
       "      <td>9379.397197</td>\n",
       "      <td>{'R0_by_stage': {0: (5.12143962412649, 3.10124...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t  actual_I       pred_I        mse_I        mse_E        mse_M  \\\n",
       "23  53      5000  5044.748507    44.748507  1552.612290  1000.000000   \n",
       "22  52      5000  3885.334894  1114.665106   558.921346  1062.156712   \n",
       "24  54      5000  6687.138370  1687.138370  2731.283735  1201.733980   \n",
       "21  51      5000  3243.199885  1756.800115   283.687985  1455.399054   \n",
       "20  50      5000  2712.188405  2287.811595   998.276604  1802.128329   \n",
       "19  49      5000  2272.050067  2727.949933  1604.367105  2106.228703   \n",
       "18  48      5000  1906.385945  3093.614055  2118.489290  2371.707795   \n",
       "25  55      5000  8184.867009  3184.867009  4132.860163  1000.000000   \n",
       "17  47      5000  1601.907769  3398.092231  2554.641722  2602.539226   \n",
       "16  46      5000  1347.832039  3652.167961  2924.681901  2802.547335   \n",
       "\n",
       "         mse_IM      mse_IEM  \\\n",
       "23   955.251493   597.360797   \n",
       "22  2176.821818  1617.900472   \n",
       "24   485.404390  3216.688125   \n",
       "21  3212.199168  3495.887153   \n",
       "20  4089.939924  5088.216527   \n",
       "19  4834.178636  6438.545741   \n",
       "18  5465.321850  7583.811140   \n",
       "25  2184.867009  6317.727172   \n",
       "17  6000.631457  8555.273178   \n",
       "16  6454.715296  9379.397197   \n",
       "\n",
       "                                              r0_info  \n",
       "23  {'R0_by_stage': {0: (5.1215123171320585, 3.150...  \n",
       "22  {'R0_by_stage': {0: (5.121544198771868, 3.1398...  \n",
       "24  {'R0_by_stage': {0: (5.121466025584044, 3.1624...  \n",
       "21  {'R0_by_stage': {0: (5.121562154320443, 3.1307...  \n",
       "20  {'R0_by_stage': {0: (5.121566395321335, 3.1230...  \n",
       "19  {'R0_by_stage': {0: (5.121556838562287, 3.1163...  \n",
       "18  {'R0_by_stage': {0: (5.121533083188067, 3.1105...  \n",
       "25  {'R0_by_stage': {0: (5.121404593397314, 3.1765...  \n",
       "17  {'R0_by_stage': {0: (5.121494384772303, 3.1055...  \n",
       "16  {'R0_by_stage': {0: (5.12143962412649, 3.10124...  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values(by='mse_I').head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2019-11-29 00:00:00\n"
     ]
    }
   ],
   "source": [
    "best_t = int(df.sort_values(by='mse_IM').iloc[0].t)\n",
    "p0_time = T('22/01/2020') - timedelta(days=best_t)\n",
    "print(p0_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "makedir_if_not_there('output/tbl/p0-time/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('output/tbl/p0-time/error.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "pkl.dump(p0_time, open('output/p0_time.pkl', 'wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = copy(params_jan27)\n",
    "params.alpha = infection_factor * params.alpha\n",
    "params.beta = infection_factor * params.beta\n",
    "params.initial_num_E = 1\n",
    "params.initial_num_I = 0\n",
    "params.initial_num_M = 0\n",
    "\n",
    "    \n",
    "total, delta, increase, trans, stats = do_simulation(best_t, bed_info, params, verbose=0,  p0_time=p0_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "makedir_if_not_there('output/tbl/before-lockdown/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_bundle([total, delta, increase, trans], p0_time, best_t, 'output/tbl/before-lockdown')\n",
    "save_to_json(stats, 'output/tbl/before-lockdown/stat.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "pkl.dump(\n",
    "    params,\n",
    "    open('output/params_before_lockdown.pkl', 'wb')\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/cloud-user/code/covid-19/helpers.py:287: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  subdf['index'] = df.index\n",
      "/home/cloud-user/code/covid-19/helpers.py:288: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  subdf['value'] = subdf[state].copy()\n",
      "/home/cloud-user/code/covid-19/helpers.py:290: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  subdf['state'] = state\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['29/11/19' '10/12/19' '21/12/19' '01/01/20' '12/01/20']\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plot_total(total, p0_time, best_t)\n",
    "fig.savefig('figs/before-lockdown.pdf')"
   ]
  }
 ],
 "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": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}