{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Basic Predictions\n",
"===============\n",
"\n",
"This notebook tries to predict number of \"confirmed cases\" from patientdb available at http://api.covid19india.org/. \n",
"\n",
"Growth rate is computed as rate of change of confirmed cases and smoothened using a 7d rolling average.\n",
"It is extrapolated using the following methods:\n",
"* Quadartic curve fitting - fails to fit the curve correctly\n",
"* Hyperbolic curve fitting using a neural net - fits well and asymptotes the growth rate at ~7%.\n",
"\n",
"The model is rudimentary and points towards a scenario where we fail to contain the epidemic and it eventually moves to stage 3/4."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"from matplotlib import pyplot as plt\n",
"from matplotlib.ticker import FuncFormatter\n",
"%matplotlib inline\n",
"import matplotlib as mpl\n",
"from matplotlib.dates import date2num, AutoDateLocator, ConciseDateFormatter, DayLocator, AutoDateFormatter\n",
"\n",
"import datetime as dt\n",
"import pandas as pd\n",
"import numpy as np\n",
"import requests\n",
"\n",
"from sklearn.neural_network import MLPRegressor"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Records: 81\n"
]
}
],
"source": [
"resp = requests.get(\"https://api.covid19india.org/data.json\")\n",
"ts = resp.json()['cases_time_series']\n",
"\n",
"r = {\n",
" \"date\": [],\n",
" \"confirmed\": [],\n",
" \"deceased\": [],\n",
" \"recovered\": []\n",
"}\n",
"\n",
"for d in ts:\n",
" try:\n",
" r['date'].append(dt.datetime.strptime(d['date'] + \"2020\", '%d %B %Y'))\n",
" r['confirmed'].append(int(d['dailyconfirmed']))\n",
" r['deceased'].append(int(d['dailydeceased']))\n",
" r['recovered'].append(int(d['dailyrecovered']))\n",
" except Exception as e:\n",
" print(d['date'])\n",
" tb.print_exc()\n",
"\n",
"df = pd.DataFrame(data=r, index=r['date'], columns=['confirmed', 'deceased', 'recovered'])\n",
"df.sort_index(inplace=True)\n",
"df = df[31:][:] #drop Feb data\n",
"print(\"Records:\", len(df))"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" confirmed | \n",
" deceased | \n",
" recovered | \n",
" total | \n",
" growth | \n",
" sma_growth | \n",
"
\n",
" \n",
" \n",
" \n",
" | 2020-05-16 | \n",
" 4794 | \n",
" 120 | \n",
" 4012 | \n",
" 90646 | \n",
" 5.584028 | \n",
" 5.370063 | \n",
"
\n",
" \n",
" | 2020-05-17 | \n",
" 5049 | \n",
" 152 | \n",
" 2538 | \n",
" 95695 | \n",
" 5.570020 | \n",
" 5.186083 | \n",
"
\n",
" \n",
" | 2020-05-18 | \n",
" 4628 | \n",
" 131 | \n",
" 2482 | \n",
" 100323 | \n",
" 4.836198 | \n",
" 5.113057 | \n",
"
\n",
" \n",
" | 2020-05-19 | \n",
" 6154 | \n",
" 146 | \n",
" 3032 | \n",
" 106477 | \n",
" 6.134187 | \n",
" 5.270289 | \n",
"
\n",
" \n",
" | 2020-05-20 | \n",
" 5716 | \n",
" 134 | \n",
" 3113 | \n",
" 112193 | \n",
" 5.368296 | \n",
" 5.321048 | \n",
"
\n",
" \n",
"
\n",
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# cum sum\n",
"df['total'] = df['confirmed'].cumsum()\n",
"\n",
"# calc growth rate & its moving avg\n",
"df['growth'] = 100 * df['confirmed'] / (df['total'] - df['confirmed'])\n",
"sma_gr = df['growth'].rolling(7).mean()\n",
"df['sma_growth'] = sma_gr\n",
"\n",
"# daily plot\n",
"ax = df.plot(y='confirmed', kind='line', title=\"Daily cases\", grid=True)\n",
"df.tail()"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" confirmed | \n",
" total | \n",
" sma_growth | \n",
" day_of_yr | \n",
"
\n",
" \n",
" \n",
" \n",
" | 2020-05-16 | \n",
" 4794.0 | \n",
" 90646.0 | \n",
" 5.370063 | \n",
" 137 | \n",
"
\n",
" \n",
" | 2020-05-17 | \n",
" 5049.0 | \n",
" 95695.0 | \n",
" 5.186083 | \n",
" 138 | \n",
"
\n",
" \n",
" | 2020-05-18 | \n",
" 4628.0 | \n",
" 100323.0 | \n",
" 5.113057 | \n",
" 139 | \n",
"
\n",
" \n",
" | 2020-05-19 | \n",
" 6154.0 | \n",
" 106477.0 | \n",
" 5.270289 | \n",
" 140 | \n",
"
\n",
" \n",
" | 2020-05-20 | \n",
" 5716.0 | \n",
" 112193.0 | \n",
" 5.321048 | \n",
" 141 | \n",
"
\n",
" \n",
"
\n",
"