analise_covid19.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fazendo análise dos dados da covid_2019\n",
    "\n",
    "### Digital inovation one(Dio.me)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importando os pacotes para a análise do dataset\n",
    "\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from datetime import datetime\n",
    "import plotly.express as px\n",
    "import plotly.graph_objects as go"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Carregando o dataFrame\n",
    "# Transformando as colunas de data para o valor date\n",
    "\n",
    "df = pd.read_csv('covid_19_data.csv', parse_dates=['ObservationDate', 'Last Update'])\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Analisando os tipos de dados do datFrame\n",
    "\n",
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nomes de colunas não devem ter letras maiúsculas e nem caracteres especiais.\n",
    "\n",
    "Será implementado uma função para fazer a limpeza dos nomes dessas colunas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re # para expressões regulares\n",
    "# função para corrigir os nomes da colunas\n",
    "\n",
    "def corrige_coluna(col_name):\n",
    "     return re.sub(r\"[/| ]\", \"\", col_name).lower()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Corrigindo o nome das colunas\n",
    "# transformando para minúsculo o nome de todas as colunas, e eliminando os espaços em branco e os carcteres especiais\n",
    "\n",
    "df.columns = [corrige_coluna(col) for col in df.columns]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Selecionando apenas os dados do brasil para fazer as análises"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# verificando os valores únicos do dataFrame\n",
    "\n",
    "df.countryregion.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Filtrando os dados usando o método loc\n",
    "# filtrando apenas por brasil(Brazil)\n",
    "\n",
    "df.loc[df.countryregion == 'Brazil']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criando um dataFrame somente para brasil\n",
    "# com casos confirmados maior que zero\n",
    "\n",
    "brasil = df.loc[\n",
    "    (df.countryregion == 'Brazil') &\n",
    "    (df.confirmed > 0)\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analisando os casos confirmados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criando um gráfico da evolução de casos confirmados\n",
    "\n",
    "px.line(brasil, 'observationdate', 'confirmed', title='Casos confirmados no Brasil')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Usando Lambda para identificar novos casos por dia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Usando programação funcional para novos casos a cada dia\n",
    "# criando uma nova coluna\n",
    "\n",
    "brasil['novoscasos'] = list(map(\n",
    "    lambda x: 0 if x==0 else brasil['confirmed'].iloc[x] - brasil['confirmed'].iloc[x-1],\n",
    "    np.arange(brasil.shape[0])\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Visualisando\n",
    "\n",
    "px.line(brasil, x='observationdate',y='novoscasos',title='Novos casoso por dia')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Mortes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Organizando a camada de dados\n",
    "# número de mortes por covid a cada dia no brasil\n",
    "\n",
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(\n",
    "    go.Scatter(x=brasil.observationdate, y=brasil.deaths, name='Mortes',\n",
    "              mode='lines+markers', line={'color':'red'})\n",
    ")\n",
    "\n",
    "# Organizando o layout\n",
    "\n",
    "fig.update_layout(title='Mortes por covid-19 no Brasil')\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Taxa de média de crescimento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criando uma funão para verificar a taxa média de crescimento diário\n",
    "\n",
    "def taxa_crescimento(data, variable, data_inicio=None, data_fim=None):\n",
    "    # se data inicio for None, define como a primeira data disponivel\n",
    "    if data_inicio == None:\n",
    "        data_inicio = data.observationdate.loc[data[variable] > 0].min()\n",
    "    else:\n",
    "        data_inicio = pd.to_datetime(data_inicio)\n",
    "\n",
    "    if data_fim == None:\n",
    "        data_fim = data.observationdate.iloc[-1]\n",
    "    else:\n",
    "        data_fim = pd.to_datetime(data_fim)\n",
    "\n",
    "    # Definir os valores de presente e passado\n",
    "    passado = data.loc[data.observationdate == data_inicio, variable].values[0]\n",
    "    presente = data.loc[data.observationdate == data_fim, variable].values[0]\n",
    "\n",
    "    # Define o número de pontos no tempo que vamos avaliar\n",
    "    n = (data_fim - data_inicio).days\n",
    "\n",
    "    # Calcular a taxa\n",
    "    taxa = (presente/passado)**(1/n) - 1\n",
    "\n",
    "    return taxa*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Taxa do crescimento médio em todo o brasil\n",
    "\n",
    "taxa_crescimento(brasil, 'confirmed')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Taxa de crescimento diário"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Criando uma função para taxa de crescimento diário\n",
    "\n",
    "def taxa_crescimento_diario(data, variable, data_inicio=None):\n",
    "    # se data inicio for None, define como a primeira data disponivel\n",
    "    if data_inicio == None:\n",
    "        data_inicio = data.observationdate.loc[data[variable] > 0].min()\n",
    "    else:\n",
    "        data_inicio = pd.to_datetime(data_inicio)\n",
    "\n",
    "    data_fim = data.observationdate.max()\n",
    "\n",
    "    # Define o número de pontos no tempo que vamos avaliar\n",
    "    n = (data_fim-data_inicio).days\n",
    "\n",
    "    # Taxa calculada de um dia para o outro\n",
    "    taxas = list(map(\n",
    "        lambda x: (data[variable].iloc[x] - data[variable].iloc[x-1]) / data[variable].iloc[x-1],\n",
    "        range(1, n+1)\n",
    "    ))\n",
    "    return np.array(taxas) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tx_dia = taxa_crescimento_diario(brasil, 'confirmed')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tx_dia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plotando no gráfico\n",
    "\n",
    "primeiro_dia = brasil.observationdate.loc[brasil.confirmed > 0].min()\n",
    "\n",
    "px.line(x=pd.date_range(primeiro_dia, brasil.observationdate.max())[1:], y=tx_dia, title='Taxa de crescimento de casos confirmados no Brasil')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Realisando cálculos de predição"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from statsmodels.tsa.seasonal import seasonal_decompose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "confirmados = brasil.confirmed\n",
    "confirmados.index = brasil.observationdate\n",
    "confirmados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "res = seasonal_decompose(confirmados)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, (ax1, ax2, ax3, ax4) = plt.subplots(4, 1, figsize=(10,8))\n",
    "\n",
    "ax1.plot(res.observed)\n",
    "ax2.plot(res.trend)\n",
    "ax3.plot(res.seasonal)\n",
    "ax4.plot(confirmados.index, res.resid)\n",
    "ax4.axhline(0, linestyle='dashed',c='black')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fazendo a modelagem com ARIMA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pmdarima.arima import auto_arima\n",
    "\n",
    "modelo = auto_arima(confirmados)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig = go.Figure(go.Scatter(\n",
    "    x= confirmados.index, y=confirmados, name='Observados'\n",
    "))\n",
    "fig.add_trace(go.Scatter(\n",
    "    x=confirmados.index, y=modelo.predict_in_sample(), name='Preditos'\n",
    "))\n",
    "fig.add_trace(go.Scatter(\n",
    "    x=pd.date_range('2020-05-20','2020-06-20'), y=modelo.predct(31), name='Forecast'\n",
    "))\n",
    "fig.update_layout(title='Previsão de casos confirmados no Brasil par os próximos 30 dias')\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modelo de crescimento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocessamentos\n",
    "\n",
    "from asyncio import futures\n",
    "\n",
    "\n",
    "train = confirmados.reset_index()[:5]\n",
    "test = confirmados.reset_index()[-5:]\n",
    "\n",
    "# Renomeando colunas\n",
    "train.rename(columns={'observationdate':'ds', 'confirmed':'y'}, inplace=True)\n",
    "test.rename(columns={'observationdate':'ds', 'confirmed':'y'}, inplace=True)\n",
    "\n",
    "# Defenir o modelo de crescimento\n",
    "profeta = Prophet(grow='logistic',changepoints=['2020-03-21', '2020-03-30','2020-04-25',\n",
    "                                                '2020-05-03','2020-05-10'])\n",
    "\n",
    "pop = 211463256\n",
    "train['cap'] = pop\n",
    "\n",
    "# Treina o modelo\n",
    "profeta.fit(train)\n",
    "\n",
    "# Construir previsões para o futuro\n",
    "future_dates = profeta.make_future_dataframe(periods=200)\n",
    "future_dates['cap'] = pop\n",
    "forecast = profeta.predict(future_dates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Gráfico para mostrar a predição\n",
    "\n",
    "fig = go.Figure()\n",
    "\n",
    "fig.add_trace(go.Scatter(x=forecast.ds, y=forecast.yhat, name='Predição'))\n",
    "fig.add_trace(go.Scatter(x=test.index, y=test, name='Observados - Teste'))\n",
    "fig.add_trace(go.Scatter(x=train.ds, y=train.y, name='Observados - Treino'))\n",
    "fig.update_layout(title='Predições de casos confirmados no Brasil')\n",
    "fig.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.8 64-bit",
   "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": "3.10.8"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "1e0a490dcc9734febb9d3e04be73961f07acdf47d6b405a440af439cedfe493c"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}