estudando_atmosfera_python.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Estudando nossa atmosfera com Python\n",
    "\n",
    "Incolor, de constituintes invisíveis e inodoro (espera-se!). Por vezes nos esquecemos que estamos rodeados de gases que constituem nossa atmosfera. Nesse artigo, vamos ver como podemos facilmente, com a linguagem Python, conseguir uma listagem dos principais constituintes do ar com web scrapping. E, também, fazer algumas contas como, por exemplo, de densidade e de massa molar média e, claro, apresentar tudo em gráficos espetaculares.\n",
    "\n",
    "Vamos fazer uma pequena lista de ferramentas que utilizaremos aqui:\n",
    "\n",
    "- Pint\n",
    "- Chempy\n",
    "- SciPy\n",
    "- Pandas\n",
    "- Web scraping com Pandas\n",
    "- Expressões regulares\n",
    "- Matplotlib\n",
    "    - estilos prontos\n",
    "    - escala de cores\n",
    "    - scatter plot\n",
    "    - personalização avançada de gráficos\n",
    "- Python padrão\n",
    "    - map\n",
    "    - funções anônimas (lambda)\n",
    "    - list comprehensions\n",
    "\n",
    "Muita coisa útil, então vamos lá.\n",
    "\n",
    "Em física, química ou matérias ligadas ao estudo das engenharias é comum aparecerem situações onde precisamos fazer contas que dependem de características do ar. Pode ser, por exemplo, uma reação química de combustão acontecendo a ar aberto onde precisamos considerar uma determinada porcentagem de oxigênio no ar ou a situação de um compressor onde precisamos saber qual a massa de ar que ele comporta numa determinada condição. Independente da situação, usualmente precisamos da densidade do ar ou da massa molar média do mesmo. Na realidade, tais quantidades estão intimamente relacionadas, como veremos.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup inicial"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para começar, vamos importar algumas bibliotecas que usaremos. No decorrer do artigo será detalhado o papel de cada uma. A `pint` já apareceu em outros artigos aqui do site, você pode [ler mais sobre ela aqui](https://cienciaprogramada.com.br/2020/09/python-unidades-cerveja-pint/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:35.854690Z",
     "iopub.status.busy": "2020-12-14T20:14:35.854518Z",
     "iopub.status.idle": "2020-12-14T20:14:36.471898Z",
     "shell.execute_reply": "2020-12-14T20:14:36.471391Z",
     "shell.execute_reply.started": "2020-12-14T20:14:35.854656Z"
    }
   },
   "outputs": [],
   "source": [
    "from scipy.constants import gas_constant\n",
    "import matplotlib.pyplot as plt\n",
    "import pint\n",
    "import chempy\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos criar um registro de unidades e um criador de quantidades do `pint` e, também, configurar o pacote para trabalhar com gráficos criados via `matplotlib`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:36.472893Z",
     "iopub.status.busy": "2020-12-14T20:14:36.472657Z",
     "iopub.status.idle": "2020-12-14T20:14:36.613281Z",
     "shell.execute_reply": "2020-12-14T20:14:36.612784Z",
     "shell.execute_reply.started": "2020-12-14T20:14:36.472801Z"
    }
   },
   "outputs": [],
   "source": [
    "ureg = pint.UnitRegistry()\n",
    "ureg.setup_matplotlib(True)\n",
    "Q_ = ureg.Quantity"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Embasamento teórico e obtendo dados químicos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos conhecer um pouco o pacote `chempy`, cujo [site do projeto é este](https://github.com/bjodah/chempy). Com o `chempy` podemos facilmente obter informações de um composto a partir de sua fórmula. Vamos começar utilizando o gás hidrogênio como exemplo:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:36.614501Z",
     "iopub.status.busy": "2020-12-14T20:14:36.614369Z",
     "iopub.status.idle": "2020-12-14T20:14:36.837972Z",
     "shell.execute_reply": "2020-12-14T20:14:36.837477Z",
     "shell.execute_reply.started": "2020-12-14T20:14:36.614476Z"
    }
   },
   "outputs": [],
   "source": [
    "gas_hidrogenio = chempy.Substance.from_formula('H2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "O código acima, bem autoexplicativo, instancia uma substância `Substance` a partir da fórmula na forma de string."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos olhar os atributos e métodos disponíveis publicamente no objeto `gas_hidrogenio` que criamos:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:36.838908Z",
     "iopub.status.busy": "2020-12-14T20:14:36.838726Z",
     "iopub.status.idle": "2020-12-14T20:14:37.218117Z",
     "shell.execute_reply": "2020-12-14T20:14:37.215682Z",
     "shell.execute_reply.started": "2020-12-14T20:14:36.838833Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['attrs',\n",
       " 'charge',\n",
       " 'composition',\n",
       " 'composition_keys',\n",
       " 'data',\n",
       " 'from_formula',\n",
       " 'html_name',\n",
       " 'latex_name',\n",
       " 'mass',\n",
       " 'molar_mass',\n",
       " 'name',\n",
       " 'unicode_name']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[x for x in dir(gas_hidrogenio) if not x.startswith('_')]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "É interessante termos a opção de exportar a fórmula em formatos como LaTeX e unicode:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:37.222524Z",
     "iopub.status.busy": "2020-12-14T20:14:37.221623Z",
     "iopub.status.idle": "2020-12-14T20:14:37.752515Z",
     "shell.execute_reply": "2020-12-14T20:14:37.752081Z",
     "shell.execute_reply.started": "2020-12-14T20:14:37.222360Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'H_{2}'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.latex_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:37.753419Z",
     "iopub.status.busy": "2020-12-14T20:14:37.753216Z",
     "iopub.status.idle": "2020-12-14T20:14:38.158813Z",
     "shell.execute_reply": "2020-12-14T20:14:38.157013Z",
     "shell.execute_reply.started": "2020-12-14T20:14:37.753337Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'H₂'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.unicode_name"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mas, para nossa finalidade aqui, o importante é termos o valor de massa molar. Podemos acessar de duas formas, através do atributo `mass` e do método `molar_mass`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:38.161921Z",
     "iopub.status.busy": "2020-12-14T20:14:38.161253Z",
     "iopub.status.idle": "2020-12-14T20:14:38.781102Z",
     "shell.execute_reply": "2020-12-14T20:14:38.779336Z",
     "shell.execute_reply.started": "2020-12-14T20:14:38.161736Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2.016"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.mass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:38.787072Z",
     "iopub.status.busy": "2020-12-14T20:14:38.786464Z",
     "iopub.status.idle": "2020-12-14T20:14:39.412732Z",
     "shell.execute_reply": "2020-12-14T20:14:39.410384Z",
     "shell.execute_reply.started": "2020-12-14T20:14:38.786962Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(2.016) * g/mol"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.molar_mass()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Como podemos perceber, o `mass` retorna um *float* adimensional, enquanto que `molar_mass` retorna com unidades. Podemos separar a magnitude da unidade caso desejado:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:39.418126Z",
     "iopub.status.busy": "2020-12-14T20:14:39.417491Z",
     "iopub.status.idle": "2020-12-14T20:14:39.670007Z",
     "shell.execute_reply": "2020-12-14T20:14:39.667549Z",
     "shell.execute_reply.started": "2020-12-14T20:14:39.417957Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(2.016)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.molar_mass().magnitude"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:39.672114Z",
     "iopub.status.busy": "2020-12-14T20:14:39.671742Z",
     "iopub.status.idle": "2020-12-14T20:14:39.944522Z",
     "shell.execute_reply": "2020-12-14T20:14:39.941957Z",
     "shell.execute_reply.started": "2020-12-14T20:14:39.672037Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(1.) * g/mol"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_hidrogenio.molar_mass().units"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Trabalhar com unidades é sempre desejável. No entanto, o `chempy` não utiliza o `pint` para gerenciar unidades e, sim, o [projeto quantities](https://github.com/python-quantities/python-quantities) que não é tão flexível quanto o `pint` e não integra muito bem com outros pacotes. Assim, neste artigo vamos trabalhar com o atributo `mass` e associar a ele uma unidade `pint`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quando trabalhamos com um gás, precisamos decidir qual modelo que será utilizado para descrever esse gás. A escolha do modelo depende do contexto. Se for um contexto onde uma grande exatidão é necessária ou próximo de condições pouco usuais (altíssima pressão, por exemplo), há [diversos modelos que podem ser utilizados](https://en.wikipedia.org/wiki/Real_gas), sendo o de [Van der Walls](https://en.wikipedia.org/wiki/Van_der_Waals_equation) provavelmente o mais conhecido. No nosso caso, o modelo simples de [gás ideal](https://en.wikipedia.org/wiki/Ideal_gas_law), que é ensinado no ensino médio em física e/ou química, é suficiente."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A equação dos gases ideais é:\n",
    "\n",
    "\n",
    "$$\n",
    "PV = nRT\n",
    "$$\n",
    "\n",
    "onde *P*, *V* e *T* são pressão, volume e temperatura; *n* é a quantidade de matéria; e *R* é a chamada constante dos gases ideais. Como quantidade de matéria de uma determinada substância é a massa da substância *m* dividida por sua massa molar *M*, podemos reescrever a equação como:\n",
    "\n",
    "$$\n",
    "PV = \\frac{m}{M} RT\n",
    "$$\n",
    "\n",
    "Lembrando que densidade de um dado material é sua massa dividida por seu volume, podemos rearranjar a equação:\n",
    "\n",
    "$$\n",
    "\\frac{m}{V} = \\frac{PM}{RT} \\therefore d = \\frac{PM}{RT}\n",
    "$$\n",
    "\n",
    "Assim, vemos que há uma ligação entre a massa molar da substância com sua densidade."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A biblioteca `SciPy` fornece um grande conjunto de constantes. No início fizemos a importação da constante dos gases com a linha\n",
    "\n",
    "```python\n",
    "from scipy.constants import gas_constant\n",
    "```\n",
    "\n",
    "Vamos verificar o valor:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:39.948067Z",
     "iopub.status.busy": "2020-12-14T20:14:39.947210Z",
     "iopub.status.idle": "2020-12-14T20:14:40.357321Z",
     "shell.execute_reply": "2020-12-14T20:14:40.354646Z",
     "shell.execute_reply.started": "2020-12-14T20:14:39.947706Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8.314462618"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gas_constant"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Conforme a [documentação da biblioteca](https://docs.scipy.org/doc/scipy/reference/constants.html), o valor da constante se encontra no Sistema Internacional de unidades (SI). Podemos associar a unidade ao valor com o criador de quantidades do `pint`, que instanciamos no início do artigo:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:40.360637Z",
     "iopub.status.busy": "2020-12-14T20:14:40.360035Z",
     "iopub.status.idle": "2020-12-14T20:14:40.851415Z",
     "shell.execute_reply": "2020-12-14T20:14:40.849673Z",
     "shell.execute_reply.started": "2020-12-14T20:14:40.360527Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "8.314462618 joule/(kelvin mole)"
      ],
      "text/latex": [
       "$8.314462618\\ \\frac{\\mathrm{joule}}{\\left(\\mathrm{kelvin} \\cdot \\mathrm{mole}\\right)}$"
      ],
      "text/plain": [
       "8.314462618 <Unit('joule / kelvin / mole')>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Q_(gas_constant, 'J/(mol K)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Obviamente que você não precisaria importar o SciPy apenas para ter o valor dessa constante, poderia simplesmente colocar o valor numa variável no código e associar a unidade. Mas fiz isso aqui no artigo por dois motivos. O primeiro é informativo, mostrar que o [SciPy](https://scipy.org/), uma das bibliotecas científicas mais sensacionais, possui esse acervo de constantes. O segundo é mostrar a quantidade de significativos do valor que a biblioteca apresenta. Sobre isso vou escrever um pouco mais.\n",
    "\n",
    "É uma (má) prática corriqueira em nosso sistema de ensino arredondar valores sem uma discussão mais detalhada das consequências disso. Assim, grandes erros acabam surgindo nas contas fruto dos arredondamentos feitos. Você que está lendo esse artigo até aqui certamente tem interesse na área científica então coloque uma coisa na cabeça: **o que tem que limitar a precisão de seus cálculos é a precisão de suas medições**. Toda medida tem uma incerteza associada. Mas essas constantes comuns em ciência, como a dos gases ideais ou a velocidade da luz, costumam passar por diversos métodos de determinação com elevada precisão, existindo inclusive o *Committee on Data of the International Science Council* (CODATA) que periodicamente publica [valores atualizados das principais constantes físicas](https://codata.org/initiatives/strategic-programme/fundamental-physical-constants/) das quais outras constantes podem ser derivadas. \n",
    "\n",
    "Caso tenha clicado no link, perceba a quantidade de algarismos significativos das constantes. É altamente provável que não será alguma dessas constantes que irá limitar a precisão de seus cálculos. Não adianta nada você, por exemplo, se esforçar para ter um valor de pressão que possa ser expresso com 4 significativos como, por exemplo, `1.200 Pa` mas arredondar a magnitude da constante dos gases para `8.31`. **Você jogou fora todo seu esforço experimental**. Utilize as constantes físicas com o máximo de significativos disponíveis. Cada arredondamento mal feito faz você perder um ano de vida (exagero para você lembrar da mensagem). \n",
    "\n",
    "Ah, e você pode estar achando estranho o valor apresentado para a constante dos gases. Talvez tenha pensado \"*ué, não era 0,082?*\" Sim, em outro sistema de unidades. O valor do SciPy está conforme o Sistema Internacional de Unidades (SI). No ensino médio e até mesmo em situações práticas usualmente utilizamos o valor mais adequado para trabalhar com as unidades atmosfera e litro. Veja a conversão:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:40.854879Z",
     "iopub.status.busy": "2020-12-14T20:14:40.854258Z",
     "iopub.status.idle": "2020-12-14T20:14:41.123063Z",
     "shell.execute_reply": "2020-12-14T20:14:41.121082Z",
     "shell.execute_reply.started": "2020-12-14T20:14:40.854768Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "0.08205736607944732 liter standard_atmosphere/(kelvin mole)"
      ],
      "text/latex": [
       "$0.08205736607944732\\ \\frac{\\mathrm{liter} \\cdot \\mathrm{standard\\_atmosphere}}{\\left(\\mathrm{kelvin} \\cdot \\mathrm{mole}\\right)}$"
      ],
      "text/plain": [
       "0.08205736607944732 <Unit('liter * standard_atmosphere / kelvin / mole')>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Q_(gas_constant, 'J/(mol K)').to('atm l / (mol K)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Novamente, atenção para os significativos do valor."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A função que irá sustentar todo o trabalho"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bom, vamos voltar ao nosso gás ideal. Podemos criar uma função que calcule a densidade de um gás. De acordo com a equação que obtivemos acima, precisamos apenas da pressão, da massa molar e da temperatura. Mas, com a ajuda do `chempy` podemos obter a massa molar a partir da fórmula do gás. Logo, nossa função pode receber uma fórmula, a pressão e a temperatura. Podemos considerar valores padrão para a pressão e a temperatura, para condições ambiente, por exemplo:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:41.127863Z",
     "iopub.status.busy": "2020-12-14T20:14:41.126900Z",
     "iopub.status.idle": "2020-12-14T20:14:41.415002Z",
     "shell.execute_reply": "2020-12-14T20:14:41.412955Z",
     "shell.execute_reply.started": "2020-12-14T20:14:41.127681Z"
    }
   },
   "outputs": [],
   "source": [
    "def densidade_gas(formula, pressao=Q_('1 atm'), temperatura=Q_(25, 'degC')):\n",
    "    \"\"\"Calcula a densidade de um gás ideal\"\"\"\n",
    "    \n",
    "    R = Q_(gas_constant, 'J/(mol K)')\n",
    "    M = chempy.Substance.from_formula(formula).mass\n",
    "    M = Q_(M, 'g/mol')\n",
    "    \n",
    "    return (pressao * M/(R * temperatura.to('kelvin'))).to_base_units()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Os detalhes de uso do pacote `pint` podem ser [lidos aqui](https://cienciaprogramada.com.br/2020/09/python-unidades-cerveja-pint/)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora podemos passar uma fórmula e teremos a densidade para as condições de 1 atm e 25 °C."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:41.418003Z",
     "iopub.status.busy": "2020-12-14T20:14:41.417434Z",
     "iopub.status.idle": "2020-12-14T20:14:41.824999Z",
     "shell.execute_reply": "2020-12-14T20:14:41.824391Z",
     "shell.execute_reply.started": "2020-12-14T20:14:41.417895Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "1.7988258254712144 kilogram/meter<sup>3</sup>"
      ],
      "text/latex": [
       "$1.7988258254712144\\ \\frac{\\mathrm{kilogram}}{\\mathrm{meter}^{3}}$"
      ],
      "text/plain": [
       "1.7988258254712144 <Unit('kilogram / meter ** 3')>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "densidade_gas('CO2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se quisermos outras condições, basta passar os parâmetros `pressao` e `temperatura`. Inclusive em qualquer unidade, já que o `pint` cuida da conversão internamente."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:41.826046Z",
     "iopub.status.busy": "2020-12-14T20:14:41.825824Z",
     "iopub.status.idle": "2020-12-14T20:14:42.080715Z",
     "shell.execute_reply": "2020-12-14T20:14:42.078707Z",
     "shell.execute_reply.started": "2020-12-14T20:14:41.826003Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "0.1937787323145848 kilogram/meter<sup>3</sup>"
      ],
      "text/latex": [
       "$0.1937787323145848\\ \\frac{\\mathrm{kilogram}}{\\mathrm{meter}^{3}}$"
      ],
      "text/plain": [
       "0.1937787323145848 <Unit('kilogram / meter ** 3')>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "densidade_gas('CO2', pressao=Q_('1 * 10^4 Pa'), temperatura=Q_(0, 'degC'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ótimo, já temos como calcular a densidade de um gás. Mas e no caso da nossa atmosfera, que é um conjunto de gases?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Web scraping com Pandas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Primeiro precisamos de uma fonte que nos dê os gases e suas quantidades relativas. [Esse link da Wikipedia](https://en.wikipedia.org/wiki/Atmosphere_of_Earth) possui uma tabela com essas informações. Vamos associar esse link a uma variável:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:42.084993Z",
     "iopub.status.busy": "2020-12-14T20:14:42.083940Z",
     "iopub.status.idle": "2020-12-14T20:14:42.356772Z",
     "shell.execute_reply": "2020-12-14T20:14:42.354925Z",
     "shell.execute_reply.started": "2020-12-14T20:14:42.084563Z"
    }
   },
   "outputs": [],
   "source": [
    "url = 'https://en.wikipedia.org/wiki/Atmosphere_of_Earth'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos entender uma das importações que fizemos no início do artigo, a da biblioteca [Pandas](https://pandas.pydata.org/). Esta biblioteca é muito utilizada para análise e manipulação de dados e é excelente para dados que podem ser organizados em tabelas. Dentro do contexto do Pandas, um conjunto de dados tabelados é chamado de **DataFrame**.\n",
    "\n",
    "O método `read_html` oferece uma lista de tabelas de uma URL na forma de objeto DataFrame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:42.360280Z",
     "iopub.status.busy": "2020-12-14T20:14:42.359283Z",
     "iopub.status.idle": "2020-12-14T20:14:43.691265Z",
     "shell.execute_reply": "2020-12-14T20:14:43.690771Z",
     "shell.execute_reply.started": "2020-12-14T20:14:42.359838Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(pd.read_html(url))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-11T23:17:05.888296Z",
     "iopub.status.busy": "2020-12-11T23:17:05.888130Z",
     "iopub.status.idle": "2020-12-11T23:17:06.875249Z",
     "shell.execute_reply": "2020-12-11T23:17:06.874756Z",
     "shell.execute_reply.started": "2020-12-11T23:17:05.888269Z"
    }
   },
   "source": [
    "Observe que foram encontrados 13 objetos que podem ser lidados como DataFrame no link. Você pode passar por cada um para ver qual é o de seu interesse. No nosso caso, é logo a primeira posição da lista, lembrando que estruturas de armazenamento em Python, como listas, começam em índice zero. Assim:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:43.692126Z",
     "iopub.status.busy": "2020-12-14T20:14:43.691929Z",
     "iopub.status.idle": "2020-12-14T20:14:44.726715Z",
     "shell.execute_reply": "2020-12-14T20:14:44.726288Z",
     "shell.execute_reply.started": "2020-12-14T20:14:43.692053Z"
    }
   },
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">Gas</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Volume(A)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Formula</th>\n",
       "      <th>in ppmv(B)</th>\n",
       "      <th>in %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Nitrogen</td>\n",
       "      <td>N2</td>\n",
       "      <td>780840</td>\n",
       "      <td>78.084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Oxygen</td>\n",
       "      <td>O2</td>\n",
       "      <td>209460</td>\n",
       "      <td>20.946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Argon</td>\n",
       "      <td>Ar</td>\n",
       "      <td>9340</td>\n",
       "      <td>0.9340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Carbon dioxide(April, 2020)(C)[13]</td>\n",
       "      <td>CO2</td>\n",
       "      <td>413.61</td>\n",
       "      <td>0.041361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Neon</td>\n",
       "      <td>Ne</td>\n",
       "      <td>18.18</td>\n",
       "      <td>0.001818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Helium</td>\n",
       "      <td>He</td>\n",
       "      <td>5.24</td>\n",
       "      <td>0.000524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Methane</td>\n",
       "      <td>CH4</td>\n",
       "      <td>1.87</td>\n",
       "      <td>0.000187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Krypton</td>\n",
       "      <td>Kr</td>\n",
       "      <td>1.14</td>\n",
       "      <td>0.000114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Not included in above dry atmosphere:</td>\n",
       "      <td>Not included in above dry atmosphere:</td>\n",
       "      <td>Not included in above dry atmosphere:</td>\n",
       "      <td>Not included in above dry atmosphere:</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Water vapor(D)</td>\n",
       "      <td>H2O</td>\n",
       "      <td>0–30,000(D)</td>\n",
       "      <td>0–3%(E)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>notes: (A) volume fraction is equal to mole fr...</td>\n",
       "      <td>notes: (A) volume fraction is equal to mole fr...</td>\n",
       "      <td>notes: (A) volume fraction is equal to mole fr...</td>\n",
       "      <td>notes: (A) volume fraction is equal to mole fr...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  Gas  \\\n",
       "                                                 Name   \n",
       "0                                            Nitrogen   \n",
       "1                                              Oxygen   \n",
       "2                                               Argon   \n",
       "3                  Carbon dioxide(April, 2020)(C)[13]   \n",
       "4                                                Neon   \n",
       "5                                              Helium   \n",
       "6                                             Methane   \n",
       "7                                             Krypton   \n",
       "8               Not included in above dry atmosphere:   \n",
       "9                                      Water vapor(D)   \n",
       "10  notes: (A) volume fraction is equal to mole fr...   \n",
       "\n",
       "                                                       \\\n",
       "                                              Formula   \n",
       "0                                                  N2   \n",
       "1                                                  O2   \n",
       "2                                                  Ar   \n",
       "3                                                 CO2   \n",
       "4                                                  Ne   \n",
       "5                                                  He   \n",
       "6                                                 CH4   \n",
       "7                                                  Kr   \n",
       "8               Not included in above dry atmosphere:   \n",
       "9                                                 H2O   \n",
       "10  notes: (A) volume fraction is equal to mole fr...   \n",
       "\n",
       "                                            Volume(A)  \\\n",
       "                                           in ppmv(B)   \n",
       "0                                              780840   \n",
       "1                                              209460   \n",
       "2                                                9340   \n",
       "3                                              413.61   \n",
       "4                                               18.18   \n",
       "5                                                5.24   \n",
       "6                                                1.87   \n",
       "7                                                1.14   \n",
       "8               Not included in above dry atmosphere:   \n",
       "9                                         0–30,000(D)   \n",
       "10  notes: (A) volume fraction is equal to mole fr...   \n",
       "\n",
       "                                                       \n",
       "                                                 in %  \n",
       "0                                              78.084  \n",
       "1                                              20.946  \n",
       "2                                              0.9340  \n",
       "3                                            0.041361  \n",
       "4                                            0.001818  \n",
       "5                                            0.000524  \n",
       "6                                            0.000187  \n",
       "7                                            0.000114  \n",
       "8               Not included in above dry atmosphere:  \n",
       "9                                             0–3%(E)  \n",
       "10  notes: (A) volume fraction is equal to mole fr...  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_html(url)[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A parte de notas sobre a tabela também foi reconhecida pelo Pandas como parte do objeto. Não é interesse nosso essa parte, então vamos retirar. Como essa parte começa a partir do índice 8 do DataFrame (o índice é a primeira coluna), vamos fazer um *slice* com `iloc`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:44.727643Z",
     "iopub.status.busy": "2020-12-14T20:14:44.727444Z",
     "iopub.status.idle": "2020-12-14T20:14:45.732345Z",
     "shell.execute_reply": "2020-12-14T20:14:45.731790Z",
     "shell.execute_reply.started": "2020-12-14T20:14:44.727570Z"
    }
   },
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">Gas</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Volume(A)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Formula</th>\n",
       "      <th>in ppmv(B)</th>\n",
       "      <th>in %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Nitrogen</td>\n",
       "      <td>N2</td>\n",
       "      <td>780840</td>\n",
       "      <td>78.084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Oxygen</td>\n",
       "      <td>O2</td>\n",
       "      <td>209460</td>\n",
       "      <td>20.946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Argon</td>\n",
       "      <td>Ar</td>\n",
       "      <td>9340</td>\n",
       "      <td>0.9340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Carbon dioxide(April, 2020)(C)[13]</td>\n",
       "      <td>CO2</td>\n",
       "      <td>413.61</td>\n",
       "      <td>0.041361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Neon</td>\n",
       "      <td>Ne</td>\n",
       "      <td>18.18</td>\n",
       "      <td>0.001818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Helium</td>\n",
       "      <td>He</td>\n",
       "      <td>5.24</td>\n",
       "      <td>0.000524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Methane</td>\n",
       "      <td>CH4</td>\n",
       "      <td>1.87</td>\n",
       "      <td>0.000187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Krypton</td>\n",
       "      <td>Kr</td>\n",
       "      <td>1.14</td>\n",
       "      <td>0.000114</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  Gas          Volume(A)          \n",
       "                                 Name Formula in ppmv(B)      in %\n",
       "0                            Nitrogen      N2     780840    78.084\n",
       "1                              Oxygen      O2     209460    20.946\n",
       "2                               Argon      Ar       9340    0.9340\n",
       "3  Carbon dioxide(April, 2020)(C)[13]     CO2     413.61  0.041361\n",
       "4                                Neon      Ne      18.18  0.001818\n",
       "5                              Helium      He       5.24  0.000524\n",
       "6                             Methane     CH4       1.87  0.000187\n",
       "7                             Krypton      Kr       1.14  0.000114"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_html(url)[0].iloc[:8]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Será esse nosso DataFrame, vamos associar a uma variável:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:45.733308Z",
     "iopub.status.busy": "2020-12-14T20:14:45.733084Z",
     "iopub.status.idle": "2020-12-14T20:14:45.735881Z",
     "shell.execute_reply": "2020-12-14T20:14:45.735305Z",
     "shell.execute_reply.started": "2020-12-14T20:14:45.733280Z"
    }
   },
   "outputs": [],
   "source": [
    "df = _"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "É sempre interessante dar uma olhada em como o Pandas reconheceu os dados. Para isso temos os métodos `info` e `describe`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:45.736636Z",
     "iopub.status.busy": "2020-12-14T20:14:45.736448Z",
     "iopub.status.idle": "2020-12-14T20:14:46.238385Z",
     "shell.execute_reply": "2020-12-14T20:14:46.237843Z",
     "shell.execute_reply.started": "2020-12-14T20:14:45.736594Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8 entries, 0 to 7\n",
      "Data columns (total 4 columns):\n",
      " #   Column                   Non-Null Count  Dtype \n",
      "---  ------                   --------------  ----- \n",
      " 0   (Gas, Name)              8 non-null      object\n",
      " 1   (Gas, Formula)           8 non-null      object\n",
      " 2   (Volume(A), in ppmv(B))  8 non-null      object\n",
      " 3   (Volume(A), in %)        8 non-null      object\n",
      "dtypes: object(4)\n",
      "memory usage: 384.0+ bytes\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:46.240614Z",
     "iopub.status.busy": "2020-12-14T20:14:46.240300Z",
     "iopub.status.idle": "2020-12-14T20:14:46.675249Z",
     "shell.execute_reply": "2020-12-14T20:14:46.674722Z",
     "shell.execute_reply.started": "2020-12-14T20:14:46.240432Z"
    }
   },
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">Gas</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Volume(A)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Formula</th>\n",
       "      <th>in ppmv(B)</th>\n",
       "      <th>in %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>Nitrogen</td>\n",
       "      <td>Ne</td>\n",
       "      <td>209460</td>\n",
       "      <td>0.041361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Gas          Volume(A)          \n",
       "            Name Formula in ppmv(B)      in %\n",
       "count          8       8          8         8\n",
       "unique         8       8          8         8\n",
       "top     Nitrogen      Ne     209460  0.041361\n",
       "freq           1       1          1         1"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Observe que o Pandas reconheceu todas as colunas como sendo de tipo `object`. Esse tipo é genérico no Pandas. Usualmente, quando o algoritmo não reconhece exatamente o tipo, enquadra como `object`. Para todos os efeitos, os dados dessas colunas se comportam como strings. Isso fica claro no resultado de `describe` pois, se as colunas com números fosse reconhecidas como tais, apareceriam métricas como média, valor máximo e valor mínimo. Logo, uma das coisas que precisaremos fazer é converter a coluna de volume em porcentagem para o tipo correto, *float*, para que possamos utilizar os valores em nossas contas."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Outro aspecto que esses métodos nos permitem visualizar de forma mais clara é que há subcolunas ou, numa linguagem mais próxima a usada no contexto de Pandas, há multi-índices. Dentro do índice *Gas* há *Name* e *Formula*, e também há índices dentro de *Volume(A)*. Vamos ver como visualizar os valores da coluna *Formula*:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:46.676841Z",
     "iopub.status.busy": "2020-12-14T20:14:46.676641Z",
     "iopub.status.idle": "2020-12-14T20:14:47.050774Z",
     "shell.execute_reply": "2020-12-14T20:14:47.050184Z",
     "shell.execute_reply.started": "2020-12-14T20:14:46.676796Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     N2\n",
       "1     O2\n",
       "2     Ar\n",
       "3    CO2\n",
       "4     Ne\n",
       "5     He\n",
       "6    CH4\n",
       "7     Kr\n",
       "Name: Formula, dtype: object"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Gas']['Formula']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Outra forma seria usando a notação do método `loc`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:47.051967Z",
     "iopub.status.busy": "2020-12-14T20:14:47.051709Z",
     "iopub.status.idle": "2020-12-14T20:14:47.436340Z",
     "shell.execute_reply": "2020-12-14T20:14:47.435950Z",
     "shell.execute_reply.started": "2020-12-14T20:14:47.051932Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     N2\n",
       "1     O2\n",
       "2     Ar\n",
       "3    CO2\n",
       "4     Ne\n",
       "5     He\n",
       "6    CH4\n",
       "7     Kr\n",
       "Name: (Gas, Formula), dtype: object"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[:, ('Gas', 'Formula')]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Podemos ver apenas os dados em *Volume(A)*, que são nossos dados numéricos:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:47.437104Z",
     "iopub.status.busy": "2020-12-14T20:14:47.436920Z",
     "iopub.status.idle": "2020-12-14T20:14:47.985402Z",
     "shell.execute_reply": "2020-12-14T20:14:47.984232Z",
     "shell.execute_reply.started": "2020-12-14T20:14:47.437042Z"
    }
   },
   "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>in ppmv(B)</th>\n",
       "      <th>in %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>780840</td>\n",
       "      <td>78.084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>209460</td>\n",
       "      <td>20.946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9340</td>\n",
       "      <td>0.9340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>413.61</td>\n",
       "      <td>0.041361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>18.18</td>\n",
       "      <td>0.001818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5.24</td>\n",
       "      <td>0.000524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.87</td>\n",
       "      <td>0.000187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.14</td>\n",
       "      <td>0.000114</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  in ppmv(B)      in %\n",
       "0     780840    78.084\n",
       "1     209460    20.946\n",
       "2       9340    0.9340\n",
       "3     413.61  0.041361\n",
       "4      18.18  0.001818\n",
       "5       5.24  0.000524\n",
       "6       1.87  0.000187\n",
       "7       1.14  0.000114"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Volume(A)']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos converter esses dados para o tipo `float`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:47.986277Z",
     "iopub.status.busy": "2020-12-14T20:14:47.986125Z",
     "iopub.status.idle": "2020-12-14T20:14:48.434815Z",
     "shell.execute_reply": "2020-12-14T20:14:48.434350Z",
     "shell.execute_reply.started": "2020-12-14T20:14:47.986249Z"
    }
   },
   "outputs": [],
   "source": [
    "df['Volume(A)'] = df['Volume(A)'].astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:48.435812Z",
     "iopub.status.busy": "2020-12-14T20:14:48.435613Z",
     "iopub.status.idle": "2020-12-14T20:14:48.740571Z",
     "shell.execute_reply": "2020-12-14T20:14:48.738128Z",
     "shell.execute_reply.started": "2020-12-14T20:14:48.435774Z"
    }
   },
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">Gas</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Volume(A)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Formula</th>\n",
       "      <th>in ppmv(B)</th>\n",
       "      <th>in %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Nitrogen</td>\n",
       "      <td>N2</td>\n",
       "      <td>780840.00</td>\n",
       "      <td>78.084000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Oxygen</td>\n",
       "      <td>O2</td>\n",
       "      <td>209460.00</td>\n",
       "      <td>20.946000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Argon</td>\n",
       "      <td>Ar</td>\n",
       "      <td>9340.00</td>\n",
       "      <td>0.934000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Carbon dioxide(April, 2020)(C)[13]</td>\n",
       "      <td>CO2</td>\n",
       "      <td>413.61</td>\n",
       "      <td>0.041361</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Neon</td>\n",
       "      <td>Ne</td>\n",
       "      <td>18.18</td>\n",
       "      <td>0.001818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Helium</td>\n",
       "      <td>He</td>\n",
       "      <td>5.24</td>\n",
       "      <td>0.000524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Methane</td>\n",
       "      <td>CH4</td>\n",
       "      <td>1.87</td>\n",
       "      <td>0.000187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Krypton</td>\n",
       "      <td>Kr</td>\n",
       "      <td>1.14</td>\n",
       "      <td>0.000114</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  Gas          Volume(A)           \n",
       "                                 Name Formula in ppmv(B)       in %\n",
       "0                            Nitrogen      N2  780840.00  78.084000\n",
       "1                              Oxygen      O2  209460.00  20.946000\n",
       "2                               Argon      Ar    9340.00   0.934000\n",
       "3  Carbon dioxide(April, 2020)(C)[13]     CO2     413.61   0.041361\n",
       "4                                Neon      Ne      18.18   0.001818\n",
       "5                              Helium      He       5.24   0.000524\n",
       "6                             Methane     CH4       1.87   0.000187\n",
       "7                             Krypton      Kr       1.14   0.000114"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos agora nos preparar para criar uma coluna para a densidade de cada gás listado na subcoluna *Formula*. Vamos ver como seria o resultado de aplicar a função `densidade_gas` em cada fórmula usando a função `map` do Python:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:48.744723Z",
     "iopub.status.busy": "2020-12-14T20:14:48.743855Z",
     "iopub.status.idle": "2020-12-14T20:14:49.081099Z",
     "shell.execute_reply": "2020-12-14T20:14:49.079090Z",
     "shell.execute_reply.started": "2020-12-14T20:14:48.744575Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1.145045483304565 <Unit('kilogram / meter ** 3')>,\n",
       " 1.3078876766894934 <Unit('kilogram / meter ** 3')>,\n",
       " 1.6328363306579126 <Unit('kilogram / meter ** 3')>,\n",
       " 1.7988258254712144 <Unit('kilogram / meter ** 3')>,\n",
       " 0.8248259562876108 <Unit('kilogram / meter ** 3')>,\n",
       " 0.1636025323611701 <Unit('kilogram / meter ** 3')>,\n",
       " 0.6557422963038172 <Unit('kilogram / meter ** 3')>,\n",
       " 3.4251631830497584 <Unit('kilogram / meter ** 3')>]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(map(densidade_gas, df['Gas']['Formula']))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Observe que as unidades aparecem, já que nossa função utiliza o `pint`. Seria interessante ter apenas o valor numérico, usando o atributo `magnitude` de uma quantidade do `pint`, já que não faz sentido aparecer unidade em todas as linhas de uma tabela. Além de também facilitar nossa vida em cálculos futuros ter apenas números. Podemos criar uma função anônima para isso e aplicá-la com `map` no resultado anterior:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:49.084449Z",
     "iopub.status.busy": "2020-12-14T20:14:49.083629Z",
     "iopub.status.idle": "2020-12-14T20:14:49.365766Z",
     "shell.execute_reply": "2020-12-14T20:14:49.363645Z",
     "shell.execute_reply.started": "2020-12-14T20:14:49.084088Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1.145045483304565,\n",
       " 1.3078876766894934,\n",
       " 1.6328363306579126,\n",
       " 1.7988258254712144,\n",
       " 0.8248259562876108,\n",
       " 0.1636025323611701,\n",
       " 0.6557422963038172,\n",
       " 3.4251631830497584]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list( map(lambda quantidade: quantidade.magnitude, map(densidade_gas, df['Gas']['Formula'])) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos armazenar o resultado anterior em uma coluna de nosso DataFrame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:49.369473Z",
     "iopub.status.busy": "2020-12-14T20:14:49.368481Z",
     "iopub.status.idle": "2020-12-14T20:14:49.698548Z",
     "shell.execute_reply": "2020-12-14T20:14:49.696467Z",
     "shell.execute_reply.started": "2020-12-14T20:14:49.368967Z"
    }
   },
   "outputs": [],
   "source": [
    "df['density'] = _"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:49.702405Z",
     "iopub.status.busy": "2020-12-14T20:14:49.701535Z",
     "iopub.status.idle": "2020-12-14T20:14:49.991388Z",
     "shell.execute_reply": "2020-12-14T20:14:49.988901Z",
     "shell.execute_reply.started": "2020-12-14T20:14:49.702239Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    1.145045\n",
       "1    1.307888\n",
       "2    1.632836\n",
       "3    1.798826\n",
       "4    0.824826\n",
       "5    0.163603\n",
       "6    0.655742\n",
       "7    3.425163\n",
       "Name: density, dtype: float64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['density']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Relembrando nossas premissas, esses valores são a densidade de cada gás separadamente considerando um modelo de gás ideal, a pressão de 1 atm e a temperatura de 25 °C."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Densidade e massa molar média do ar"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Conforme explicado no link da Wikipedia de onde tiramos a tabela, os valores de porcentagem apresentados consideram ar seco, já que o teor de umidade varia muito mesmo em uma dada localidade. Assim, quando falamos de densidade do ar usualmente estamos nos referindo a densidade do ar seco. Essa densidade nada mais é que a média ponderada das densidades de cada constituinte do ar. Os pesoas são as porcentagens de cada constituinte. Assim:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:49.995449Z",
     "iopub.status.busy": "2020-12-14T20:14:49.994560Z",
     "iopub.status.idle": "2020-12-14T20:14:50.287256Z",
     "shell.execute_reply": "2020-12-14T20:14:50.285290Z",
     "shell.execute_reply.started": "2020-12-14T20:14:49.995292Z"
    }
   },
   "outputs": [],
   "source": [
    "densidade_ar_seco = sum([densidade_gas * porcentagem_gas for densidade_gas, porcentagem_gas in \n",
    "                         zip(df['density'], df['Volume(A)']['in %'])]) / sum(df['Volume(A)']['in %'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:50.291258Z",
     "iopub.status.busy": "2020-12-14T20:14:50.290350Z",
     "iopub.status.idle": "2020-12-14T20:14:50.565152Z",
     "shell.execute_reply": "2020-12-14T20:14:50.563205Z",
     "shell.execute_reply.started": "2020-12-14T20:14:50.291082Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.1839683903282519"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "densidade_ar_seco"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Esse é o valor calculado da densidade do ar seco. Considerando o número de significativos dos dados apresentados e colocando a devida unidade, é mais adequado representar o resultado como:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:50.568712Z",
     "iopub.status.busy": "2020-12-14T20:14:50.567924Z",
     "iopub.status.idle": "2020-12-14T20:14:50.838901Z",
     "shell.execute_reply": "2020-12-14T20:14:50.836533Z",
     "shell.execute_reply.started": "2020-12-14T20:14:50.568585Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.2 kg/m³\n"
     ]
    }
   ],
   "source": [
    "print(f\"{Q_(densidade_ar_seco, 'kg/m^3'):.2~P}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Seguindo a mesma linha de raciocínio, podemos calcular a massa molar média:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:50.841074Z",
     "iopub.status.busy": "2020-12-14T20:14:50.840901Z",
     "iopub.status.idle": "2020-12-14T20:14:51.162159Z",
     "shell.execute_reply": "2020-12-14T20:14:51.160103Z",
     "shell.execute_reply.started": "2020-12-14T20:14:50.841044Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "28.97 g/mol\n"
     ]
    }
   ],
   "source": [
    "df['molar mass'] = [chempy.Substance.from_formula(formula).mass for formula in df['Gas']['Formula']]\n",
    "\n",
    "massa_molar_media_ar_seco = sum([massa_molar_gas * porcentagem_gas for massa_molar_gas, porcentagem_gas in zip(df['molar mass'], df['Volume(A)']['in %'])]) / sum(df['Volume(A)']['in %'])\n",
    "\n",
    "print(f\"{Q_(massa_molar_media_ar_seco, 'g/mol'):.4~P}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gráficos! Eu quero imagens"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos usar todo o poder do Matplotlib para criar gráficos que facilitem a visualização da nossa análise e mesmo expanda ainda mais nossa compreensão.\n",
    "\n",
    "O Pandas possui um método `plot` que é construído em cima do `plot` do Matplotlib. Podemos, por exemplo, criar facilmente um gráfico de barras com as densidades de cada gás do nosso DataFrame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:51.165283Z",
     "iopub.status.busy": "2020-12-14T20:14:51.164659Z",
     "iopub.status.idle": "2020-12-14T20:14:51.612793Z",
     "shell.execute_reply": "2020-12-14T20:14:51.612332Z",
     "shell.execute_reply.started": "2020-12-14T20:14:51.165167Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fed70f07b50>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.plot.bar(x=('Gas', 'Formula'), y='density')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As configurações padrão do Matplotlib são boas, mas há muito o que melhorar nesse gráfico. Afinal, falta mostrar a unidade no eixo vertical, títulos melhores nos eixos e os números nas fórmulas químicas deveriam estar em subscrito. Vamos resolver essas questões."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para as fórmulas químicas, podemos criar uma função que, com auxílio de expressões regulares, retorna o código LaTeX necessário para apresentar as fórmulas corretamente. Isso porque o Matplotlib interage muito bem com LaTeX, o que é ótimo para nós da área de ciência que utilizamos muito em nossos trabalhos."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:51.613922Z",
     "iopub.status.busy": "2020-12-14T20:14:51.613616Z",
     "iopub.status.idle": "2020-12-14T20:14:51.674461Z",
     "shell.execute_reply": "2020-12-14T20:14:51.673930Z",
     "shell.execute_reply.started": "2020-12-14T20:14:51.613790Z"
    }
   },
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "def format_formula(formula):\n",
    "        \"\"\" Retorna código LaTeX para apresentar fórmulas químicas corretamente\"\"\"\n",
    "        \n",
    "        label_formula = re.sub(\"([0-9])\", r\"_\\1\", formula)\n",
    "        \n",
    "        return ''.join(['$\\mathregular{', label_formula, '}$'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Basicamente a função acrescenta antes de qualquer dígito um `_` de forma que o código LaTeX passa a reconhecer o dígito como subscrito. Para entender mais sobre expressões regulares [leia esse excelente artigo aqui do site](https://cienciaprogramada.com.br/2020/09/regex-expressoes-regulares-python/)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "O esquema de cores padrão do Matplotlib é bom e apresentável, mas particularmente gosto bastante de gráficos com fundo escuro. Assim, vou usar um dos [estilos prontos](https://matplotlib.org/3.1.0/gallery/style_sheets/style_sheets_reference.html) que há na biblioteca:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:51.675390Z",
     "iopub.status.busy": "2020-12-14T20:14:51.675145Z",
     "iopub.status.idle": "2020-12-14T20:14:52.070555Z",
     "shell.execute_reply": "2020-12-14T20:14:52.069778Z",
     "shell.execute_reply.started": "2020-12-14T20:14:51.675309Z"
    }
   },
   "outputs": [],
   "source": [
    "plt.style.use('dark_background')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos criar uma figura, associar a ela um sistema de eixos e colocar as informações que queríamos de unidade, nomes nos eixos, fórmulas químicas corretas e, para aproveitar, uma linha indicando a densidade do ar seco que calculamos anteriormente:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:52.072169Z",
     "iopub.status.busy": "2020-12-14T20:14:52.071844Z",
     "iopub.status.idle": "2020-12-14T20:14:52.714857Z",
     "shell.execute_reply": "2020-12-14T20:14:52.714304Z",
     "shell.execute_reply.started": "2020-12-14T20:14:52.072108Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(8,6), constrained_layout=True)\n",
    "\n",
    "# Dados\n",
    "ax.bar(df['Gas']['Formula'], df['density'], color='blue', label='Densidade do gás')\n",
    "ax.axhline(densidade_ar_seco, color='red', label='Densidade do ar seco')\n",
    "\n",
    "# Fórmulas químicas certas\n",
    "ax.set_xticklabels(list(map(format_formula, df['Gas']['Formula'])))\n",
    "\n",
    "# Identificação\n",
    "ax.set_xlabel('Composto', size=14)\n",
    "ax.set_ylabel('Densidade / kg/m³', size=14)\n",
    "ax.set_title('Densidade dos gases componentes do ar seco - 1 atm e 25 °C', size=16)\n",
    "ax.legend(fontsize=12)\n",
    "\n",
    "# Ajuste da fonte das escalas\n",
    "ax.tick_params(labelsize=14)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Esse gráfico é interessante, mas ele não apresenta ao leitor informações o suficiente para que ele entenda porque a densidade do ar seco é muito mais próxima daquela do gás nitrogênio do que do gás criptônio. Obviamente que os mais atentos deduzirão que isso provavelmente ocorre por ter muito pouco criptônio na atmosfera, mas seria útil esse dado estar explícito no gráfico.\n",
    "\n",
    "Além disso, se o público for mais acadêmico como, por exemplo, em uma aula, poderia estar no gráfico também alguma indicação de que a densidade de um dado gás é proporcional à sua massa molar. Assim, mais discussões poderiam ser feitas sobre modelos de gases e mesmo sobre outras consequências a depender do objetivo do instrutor."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Então, agora vamos fazer um gráfico mais ousado para terminar com chave de ouro esse artigo. Podemos criar uma escala de intensidade de cores para indicar a concentração de gás na atmosfera. Por exemplo, uma cor amarela fraca indicaria baixa concentração e uma vermelho forte, alta concentração. Um gráfico de barras nos limita um pouco, vamos mudar para um gráfico de pontos do tipo *scatter* (dispersão em português). Nesse tipo de gráfico, o tamanho do ponto, ou melhor círculo, pode ser associado a uma quantidade. Assim, vamos associar ao valor da massa molar para termos a informação adicional que desejamos."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para a criação da escala de cores, precisaremos de dois métodos do Matplotlib que ainda não importamos:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:52.716081Z",
     "iopub.status.busy": "2020-12-14T20:14:52.715772Z",
     "iopub.status.idle": "2020-12-14T20:14:52.719150Z",
     "shell.execute_reply": "2020-12-14T20:14:52.718424Z",
     "shell.execute_reply.started": "2020-12-14T20:14:52.716008Z"
    }
   },
   "outputs": [],
   "source": [
    "import matplotlib.colors as mcolors\n",
    "import matplotlib.cm as cm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos, então, criar uma figura com sistema de eixos e fazer os ajustes necessários no nosso gráfico. Usaremos o mapa de cores `YlOrRd` do Matplotlib e normalizaremos as cores com base nos valores de porcentagem em volume. Assim, o gás que tiver a menor porcentagem terá a cor amarela mais clara e o que tiver a maior, cor vermelha escura. Com esses dados, utilizaremos métodos do Matplotlib para construir uma escala de cores. Usando o `chempy`, obteremos a massa molar de cada fórmula no DataFrame e associaremos os valores a área de cada círculo. O código a seguir faz todas essas etapas. Recomendo que rode o código em sua máquina e interaja com ele, modificando cada parte e vendo seus efeitos. Dá para aprender muita coisa de Matplotlib aqui."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-12-14T20:14:52.720474Z",
     "iopub.status.busy": "2020-12-14T20:14:52.720205Z",
     "iopub.status.idle": "2020-12-14T20:14:53.353206Z",
     "shell.execute_reply": "2020-12-14T20:14:53.352628Z",
     "shell.execute_reply.started": "2020-12-14T20:14:52.720436Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(10,6), constrained_layout=True)\n",
    "\n",
    "# tamanho dos círculos com base na massa molar\n",
    "raio_base = 15  # mude para ter círculos menores ou maiores\n",
    "area_circulos = [chempy.Substance.from_formula(i).mass * raio_base for i in df['Gas']['Formula']]\n",
    "\n",
    "# cores dos círculos com base nos valores de porcentagem em volume\n",
    "cores_circulos = df['Volume(A)']['in %']\n",
    "normalize = mcolors.Normalize(vmin=min(cores_circulos), vmax=max(cores_circulos))\n",
    "colormap = cm.YlOrRd\n",
    "\n",
    "# dados\n",
    "sc = ax.scatter(df['Gas']['Formula'], df['density'], label='Densidade do gás - Círculo proporcional à massa molar', \n",
    "                s=area_circulos, c=cores_circulos, cmap=colormap)\n",
    "ax.axhline(densidade_ar_seco, color='red', label='Densidade do ar seco')\n",
    "\n",
    "# escala de cores\n",
    "scalarmappaple = cm.ScalarMappable(norm=normalize, cmap=colormap)\n",
    "scalarmappaple.set_array(cores_circulos)\n",
    "colorbar = fig.colorbar(scalarmappaple)\n",
    "\n",
    "# Identificação\n",
    "ax.set_xticklabels(list(map(format_formula, df['Gas']['Formula'])))\n",
    "legenda = ax.legend(scatterpoints=3, fontsize=12)\n",
    "legenda.legendHandles[1]._sizes = [30]\n",
    "ax.set_xlabel('Composto', size=18)\n",
    "ax.set_ylabel('Densidade / kg/m³', size=18)\n",
    "colorbar.ax.set_ylabel('Composição do ar seco, em % de volume', rotation=270, labelpad=20, size=15)\n",
    "fig.suptitle('Densidade dos gases componentes do ar seco - 1 atm e 25 °C', size=20)\n",
    "\n",
    "# Ajuste da fonte das escalas\n",
    "colorbar.ax.tick_params(labelsize=14)\n",
    "ax.tick_params(labelsize=14)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora temos um gráfico muito rico em informações: densidade de cada gás isoladamente, densidade do ar seco, relação da densidade com a massa molar e composição percentual do ar. Tudo isso em apenas duas dimensões, o que é excelente para visualização estática. Gráficos em três dimensões são ótimos quando a pessoa pode interagir, rodando o mesmo, dando zoom. Mas em um material impresso ou online mas estático, pode ser um pouco confuso. Logo, colocar mais informações mantendo um gráfico bidimensional é uma melhor escolha."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusão. E um pedido..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "E aí curtiu? Vamos fazer uma pequena lista de ferramentas que utilizamos aqui:\n",
    "\n",
    "- Pint\n",
    "- Chempy\n",
    "- SciPy\n",
    "- Pandas\n",
    "- Web scraping com Pandas\n",
    "- Expressões regulares\n",
    "- Matplotlib\n",
    "    - estilos prontos\n",
    "    - escala de cores\n",
    "    - scatter plot\n",
    "    - personalização avançada de gráficos\n",
    "- Python padrão\n",
    "    - map\n",
    "    - funções anônimas (lambda)\n",
    "    - list comprehensions\n",
    "    \n",
    "Se você acha que esse artigo te ajudou, deixe um comentário e compartilhe em suas redes sociais. Isso ajuda muito esse projeto que está apenas começando a ser mais conhecido. É um pedido simples, mas que ajuda muito.\n",
    "\n",
    "\n",
    "Por fim, caso tenha interesse em estudar mais nossa atmosfera, fica a dica do excelente livro [Introdução à Química da Atmosfera](https://amzn.to/3oRNPVK):\n",
    "\n",
    "<a href=\"https://www.amazon.com.br/Introdu%C3%A7%C3%A3o-Qu%C3%ADmica-Atmosfera-Ci%C3%AAncia-Vida-Sobreviv%C3%AAncia/dp/8521634846/ref=as_li_ss_il?__mk_pt_BR=%C3%85M%C3%85%C5%BD%C3%95%C3%91&dchild=1&keywords=quimica+atmosfera&qid=1607965937&sr=8-1&linkCode=li3&tag=cienciaprogra-20&linkId=f694de05ca970f38a6302ec14c44cca4&language=pt_BR\" target=\"_blank\"><img border=\"0\" src=\"//ws-na.amazon-adsystem.com/widgets/q?_encoding=UTF8&ASIN=8521634846&Format=_SL250_&ID=AsinImage&MarketPlace=BR&ServiceVersion=20070822&WS=1&tag=cienciaprogra-20&language=pt_BR\" ></a><img src=\"https://ir-br.amazon-adsystem.com/e/ir?t=cienciaprogra-20&language=pt_BR&l=li3&o=33&a=8521634846\" width=\"1\" height=\"1\" border=\"0\" alt=\"\" style=\"border:none !important; margin:0px !important;\" />\n",
    "\n",
    "Recentemente lançado em sua segunda edição, possui uma linguagem rigorosa, mas acessível, sendo uma grande fonte de informações a respeito de nossa atmosfera.\n",
    "\n",
    "Até a próxima!"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:chempy]",
   "language": "python",
   "name": "conda-env-chempy-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}