Analise_de_Dados_Spotify.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Análise de Dados - Spotify",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a4hKy-bJKAls",
        "colab_type": "text"
      },
      "source": [
        "# **Mini projeto - Turing Academy**\n",
        "Análise de dados: Spotify\n",
        "\n",
        "Autor: Enzo Bustos Da Silva\n",
        "\n",
        "Agradecimento especial aos meus mentores: Camila, Mura e William"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XepIwXzJhJKr",
        "colab_type": "text"
      },
      "source": [
        "# **Informações sobre o Dataset:**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zbCnnKlyhOoC",
        "colab_type": "text"
      },
      "source": [
        "## **DataSet:**\n",
        "Contém as mais diversas informações sobre cerca de 19000 músicas distribuídas entre 15 features.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zyGc0X1iWx7O",
        "colab_type": "text"
      },
      "source": [
        "## **Features:**\n",
        "\n",
        "**song_popularity** (popularidade): Classificação das músicas pela audiência do spotify. (*Target*)\n",
        "\n",
        "**song_duration_ms** (duração): A duração da música em milisegundos.\n",
        "\n",
        "**acousticness** (acústica): É uma medida de confiança de 0.0 até 1.0 que mostra o quão acústica é uma faixa. Sendo 1.0 uma faixa que apresenta alta confiabilidade de ser acústica.\n",
        "\n",
        "**danceability** (danceabilidade): Descreve o quão conveniente é dançar para esta faixa, baseado em uma combinação de elementos musicais como tempo (marcação de cada unidade de compasso, dita o ritmo), estabilidade do ritmo, força das bastidas e a regularidade de uma forma geral. Sendo 0.0 um valor para uma música pouco dançável e 1.0 é mais dançável.\n",
        "\n",
        "**energy** (energia): É uma medição de 0.0 até 1.0 que representa uma medida de intensidade e atividade. Em termos mais comuns, uma música energética parece mais rápida, alta e barulhenta. Por exemplo, death metal é bem energético, enquanto o prelúdio de Bach pontua baixo nessa escala.\n",
        "\n",
        "**instrumentalness** (instrumentalidade): Prevê se uma faixa tem ou não a presença de vocais. Sons como \"Ooh\" e \"Aah\" são tratados como instrumentais nesse contexto. Música rap ou palavras faladas são claramente \"vocais\". Quanto mais perto de 1.0 o valor de instrumentalidade, maior a probabilidade dessa faixa não conter conteúdo vocal. Valores acima de 0.5 tem a intenção de representar faixas instrumentais, porém a confiança é maior quando o valor se aproxima de 1.0.\n",
        "\n",
        "**key** (nota musical): Uma estimativa na nota musical da faixa. Números inteiros são mapeados às notas usando a notação padrão de Classe de Notas. Por Exemplo 0 = C (Dó), 1 = C♯/D♭ (Dó sustenido/Ré bemol), 2 = D (Ré), e assim por diante. Se nenhuma nota musical for detectada, é atribuido o valor de -1.\n",
        "\n",
        "**liveness** (ao vivo): É uma medição que detecta a presença de uma audiência na gravação. Valores mais altos indicam uma maior probabilidade da música ter sido gravada ao vivo. Um valor acima de 0.8 indica uma probabilidade bem alta da música ser ao vivo.\n",
        "\n",
        "**loudness** (volume): O de forma geral o volume da música em debicéis (dB). Esses valores são uma média do volume ao longo de toda a faixa e são úteis para comparar o volume relativo entre outras faixas de música. O volume é a primeira qualidade de um som que é psicológicamente correlacionada com a força física (amplitude). Valores tipicamente estão entre -60 e 0 dB.\n",
        "\n",
        "**audio_mode** (modo de áudio): Indica a modalidade (maior ou menor) de uma música, mostra o tipo da escala a qual o seu conteúdo melódico é derivada. Maior é representado por 1 e menor por 0.\n",
        "\n",
        "**speechiness** (fala/discurso): Esse fator detecta a presença de palavras faladas dentro da musica. Quanto mais voltada para o discurso é a gravação (por exemplo: entrevistas, audiolivro, poesia), mais próximo de 1.0 vale esse atributo. Valores acima de 0.66 descrevem faixas que provavelmente são feitas apedas de discursos. Valores entre 0.33 e 0.66 descrevem faixam que provavelmente contém tanto música e fala, em seções ou camadas, incuindo casos como música rap. Valores abaixo de 0.33 são mais prováveis de representarem música e outras faixas que não são como discursos.\n",
        "\n",
        "**tempo** (ritmo): De forma geral mostra qual o ritmo estimado da faixa em batidas por minuto (BPM). Em terminologia musical, tempo é a velocidade ou duração de cada compasso, o que deriva diretamente da média de duração da batida.\n",
        "\n",
        "**time_signature** (Assinatura de tempo): Uma estimativa geral da assinatura de tempo de uma faixa. A assinatura de tempo (medida) é uma convenção notacional para especificar quantas batidas (pulsos) estão em cada barra (segmento de tempo dividido regularmente).\n",
        "\n",
        "**audio_valence** (valência de áudio): Uma medida de 0.0 até 1.0 que descreve a positividade musical transmitida por uma faixa de música. As faixas com alta valência coam mais positivas (por exemplo: felizes, alegra, eufóricas), enquanto faixas com baixa valência soam mais negativas (por exemplo: tristes, deprimidas, zangadas)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EUz2gbn2LPRz",
        "colab_type": "text"
      },
      "source": [
        "# **Importando bibliotecas e dados**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PIIBrKfeLLM8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import numpy as np"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jEES83T0nWmp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df = pd.read_csv('spotify.csv')\n",
        "df.drop('Unnamed: 0', axis=1, inplace=True) #Tira segunda coluna de index numérico (inútil)\n",
        "df.set_index('song_name', inplace=True) #Tira index padrão do pandas e usa o nome"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JkFlQdqYcsTe",
        "colab_type": "text"
      },
      "source": [
        "# **Observações iniciais:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M4pkKEdZLqt5",
        "colab_type": "code",
        "outputId": "8777e588-4565-4563-b10c-c43f6278e9de",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 820
        }
      },
      "source": [
        "df.sort_index(inplace=True) #Deixa os indexes em ordem alfabética\n",
        "df.head(15) #Mostra primeiras linhas do dataframe"
      ],
      "execution_count": 860,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>song_popularity</th>\n",
              "      <th>song_duration_ms</th>\n",
              "      <th>acousticness</th>\n",
              "      <th>danceability</th>\n",
              "      <th>energy</th>\n",
              "      <th>instrumentalness</th>\n",
              "      <th>key</th>\n",
              "      <th>liveness</th>\n",
              "      <th>loudness</th>\n",
              "      <th>audio_mode</th>\n",
              "      <th>speechiness</th>\n",
              "      <th>tempo</th>\n",
              "      <th>time_signature</th>\n",
              "      <th>audio_valence</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_name</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>\"40\" - Live Version 1983</th>\n",
              "      <td>1</td>\n",
              "      <td>225440</td>\n",
              "      <td>0.109kg</td>\n",
              "      <td>0.257mol/L</td>\n",
              "      <td>0.472</td>\n",
              "      <td>0.000945</td>\n",
              "      <td>8.000</td>\n",
              "      <td>0.8290000000000001</td>\n",
              "      <td>-16.259</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0399</td>\n",
              "      <td>143.517</td>\n",
              "      <td>4</td>\n",
              "      <td>0.286</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>\"45\"</th>\n",
              "      <td>48</td>\n",
              "      <td>202493</td>\n",
              "      <td>0.000696kg</td>\n",
              "      <td>0.315mol/L</td>\n",
              "      <td>0.97</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.000</td>\n",
              "      <td>0.277</td>\n",
              "      <td>-4.709</td>\n",
              "      <td>1</td>\n",
              "      <td>0.102</td>\n",
              "      <td>178.06799999999996</td>\n",
              "      <td>4</td>\n",
              "      <td>0.423</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>\"99\"</th>\n",
              "      <td>67</td>\n",
              "      <td>199413</td>\n",
              "      <td>0.00598kg</td>\n",
              "      <td>0.552mol/L</td>\n",
              "      <td>0.804</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.000</td>\n",
              "      <td>0.111</td>\n",
              "      <td>-4.299</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0303</td>\n",
              "      <td>95.98</td>\n",
              "      <td>4</td>\n",
              "      <td>0.714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>\"Heroes\" - Single Version; 2014 Remastered Version</th>\n",
              "      <td>58</td>\n",
              "      <td>213813</td>\n",
              "      <td>0.016kg</td>\n",
              "      <td>0.502mol/L</td>\n",
              "      <td>0.6809999999999999</td>\n",
              "      <td>0.107</td>\n",
              "      <td>7.000</td>\n",
              "      <td>0.102</td>\n",
              "      <td>-8.263</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0276</td>\n",
              "      <td>112.16</td>\n",
              "      <td>4</td>\n",
              "      <td>0.584</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>#CaixaGang</th>\n",
              "      <td>10</td>\n",
              "      <td>208484</td>\n",
              "      <td>0.612kg</td>\n",
              "      <td>0.6970000000000001mol/L</td>\n",
              "      <td>0.907</td>\n",
              "      <td>0.282</td>\n",
              "      <td>3.000</td>\n",
              "      <td>0.0624</td>\n",
              "      <td>-5.13</td>\n",
              "      <td>1</td>\n",
              "      <td>0.092</td>\n",
              "      <td>98.953</td>\n",
              "      <td>4</td>\n",
              "      <td>0.542</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>#SELFIE</th>\n",
              "      <td>27</td>\n",
              "      <td>183750</td>\n",
              "      <td>0.0136kg</td>\n",
              "      <td>0.8079999999999999mol/L</td>\n",
              "      <td>0.915</td>\n",
              "      <td>9.03e-06</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.0788</td>\n",
              "      <td>-3.277</td>\n",
              "      <td>1</td>\n",
              "      <td>0.233</td>\n",
              "      <td>127.984</td>\n",
              "      <td>4</td>\n",
              "      <td>0.647</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>$$$ - with Matt Ox</th>\n",
              "      <td>78</td>\n",
              "      <td>130403</td>\n",
              "      <td>0.0636kg</td>\n",
              "      <td>0.774mol/L</td>\n",
              "      <td>0.507</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.138</td>\n",
              "      <td>-6.952000000000001</td>\n",
              "      <td>0</td>\n",
              "      <td>0.065</td>\n",
              "      <td>140.067</td>\n",
              "      <td>4</td>\n",
              "      <td>0.508</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>$orries</th>\n",
              "      <td>62</td>\n",
              "      <td>214857</td>\n",
              "      <td>0.7609999999999999kg</td>\n",
              "      <td>0.526mol/L</td>\n",
              "      <td>0.236</td>\n",
              "      <td>0.00387</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.126</td>\n",
              "      <td>-13.59</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0955</td>\n",
              "      <td>103.401</td>\n",
              "      <td>4</td>\n",
              "      <td>0.299</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>&amp; Down</th>\n",
              "      <td>0</td>\n",
              "      <td>258133</td>\n",
              "      <td>0.0473kg</td>\n",
              "      <td>0.619mol/L</td>\n",
              "      <td>0.629</td>\n",
              "      <td>0.227</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.314</td>\n",
              "      <td>-4.109</td>\n",
              "      <td>0</td>\n",
              "      <td>0.33</td>\n",
              "      <td>123.003</td>\n",
              "      <td>4</td>\n",
              "      <td>0.262</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>&amp;burn (with Vince Staples)</th>\n",
              "      <td>28</td>\n",
              "      <td>179015</td>\n",
              "      <td>0.194kg</td>\n",
              "      <td>0.787mol/L</td>\n",
              "      <td>0.379</td>\n",
              "      <td>0.00131</td>\n",
              "      <td>9.000</td>\n",
              "      <td>0.099</td>\n",
              "      <td>-9.364</td>\n",
              "      <td>0</td>\n",
              "      <td>0.245</td>\n",
              "      <td>82.52</td>\n",
              "      <td>4</td>\n",
              "      <td>0.344</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>'Til I Get It Right</th>\n",
              "      <td>35</td>\n",
              "      <td>157933</td>\n",
              "      <td>0.843kg</td>\n",
              "      <td>0.528mol/L</td>\n",
              "      <td>0.146</td>\n",
              "      <td>0.000356</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.101</td>\n",
              "      <td>-13.271</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0288</td>\n",
              "      <td>133.894</td>\n",
              "      <td>4</td>\n",
              "      <td>0.336</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>'Till I Collapse</th>\n",
              "      <td>85</td>\n",
              "      <td>297893</td>\n",
              "      <td>0.0757kg</td>\n",
              "      <td>0.5720000000000001mol/L</td>\n",
              "      <td>0.853</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.0798</td>\n",
              "      <td>-3.2030000000000003</td>\n",
              "      <td>1</td>\n",
              "      <td>0.217</td>\n",
              "      <td>171.297</td>\n",
              "      <td>4</td>\n",
              "      <td>0.102</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>'Till I Collapse</th>\n",
              "      <td>85</td>\n",
              "      <td>297893</td>\n",
              "      <td>0.0757kg</td>\n",
              "      <td>0.5720000000000001mol/L</td>\n",
              "      <td>0.853</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.0798</td>\n",
              "      <td>-3.2030000000000003</td>\n",
              "      <td>1</td>\n",
              "      <td>0.217</td>\n",
              "      <td>171.297</td>\n",
              "      <td>4</td>\n",
              "      <td>0.102</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>'Till I Collapse</th>\n",
              "      <td>85</td>\n",
              "      <td>297893</td>\n",
              "      <td>0.0757kg</td>\n",
              "      <td>0.5720000000000001mol/L</td>\n",
              "      <td>0.853</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.0798</td>\n",
              "      <td>-3.2030000000000003</td>\n",
              "      <td>1</td>\n",
              "      <td>0.217</td>\n",
              "      <td>171.297</td>\n",
              "      <td>4</td>\n",
              "      <td>0.102</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>(Don't Fear) The Reaper</th>\n",
              "      <td>76</td>\n",
              "      <td>308120</td>\n",
              "      <td>0.0029kg</td>\n",
              "      <td>0.333mol/L</td>\n",
              "      <td>0.927</td>\n",
              "      <td>0.000211</td>\n",
              "      <td>9.000</td>\n",
              "      <td>0.297</td>\n",
              "      <td>-8.55</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0733</td>\n",
              "      <td>141.47</td>\n",
              "      <td>4</td>\n",
              "      <td>0.382</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                   song_popularity  ... audio_valence\n",
              "song_name                                                           ...              \n",
              "\"40\" - Live Version 1983                                         1  ...         0.286\n",
              "\"45\"                                                            48  ...         0.423\n",
              "\"99\"                                                            67  ...         0.714\n",
              "\"Heroes\" - Single Version; 2014 Remastered Version              58  ...         0.584\n",
              "#CaixaGang                                                      10  ...         0.542\n",
              "#SELFIE                                                         27  ...         0.647\n",
              "$$$ - with Matt Ox                                              78  ...         0.508\n",
              "$orries                                                         62  ...         0.299\n",
              "& Down                                                           0  ...         0.262\n",
              "&burn (with Vince Staples)                                      28  ...         0.344\n",
              "'Til I Get It Right                                             35  ...         0.336\n",
              "'Till I Collapse                                                85  ...         0.102\n",
              "'Till I Collapse                                                85  ...         0.102\n",
              "'Till I Collapse                                                85  ...         0.102\n",
              "(Don't Fear) The Reaper                                         76  ...         0.382\n",
              "\n",
              "[15 rows x 14 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 860
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oVeN5mwtrppL",
        "colab_type": "code",
        "outputId": "2f3d7308-edaf-4ac9-80a8-e392a40e3582",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "df.sort_values('song_popularity', ascending=False, inplace=True) #Deixa as músicas em ordem de popularidade\n",
        "df.head(15)"
      ],
      "execution_count": 861,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>song_popularity</th>\n",
              "      <th>song_duration_ms</th>\n",
              "      <th>acousticness</th>\n",
              "      <th>danceability</th>\n",
              "      <th>energy</th>\n",
              "      <th>instrumentalness</th>\n",
              "      <th>key</th>\n",
              "      <th>liveness</th>\n",
              "      <th>loudness</th>\n",
              "      <th>audio_mode</th>\n",
              "      <th>speechiness</th>\n",
              "      <th>tempo</th>\n",
              "      <th>time_signature</th>\n",
              "      <th>audio_valence</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_name</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>Party In The U.S.A.</th>\n",
              "      <td>nao_sei</td>\n",
              "      <td>0.8220000000000001kg</td>\n",
              "      <td>0.519mol/L</td>\n",
              "      <td>0.36</td>\n",
              "      <td>0.0</td>\n",
              "      <td>10</td>\n",
              "      <td>0.177</td>\n",
              "      <td>-8.575</td>\n",
              "      <td>0</td>\n",
              "      <td>0.105</td>\n",
              "      <td>97.42</td>\n",
              "      <td>4</td>\n",
              "      <td>0.7</td>\n",
              "      <td>nan</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946</td>\n",
              "      <td>0.0114kg</td>\n",
              "      <td>0.901mol/L</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.33</td>\n",
              "      <td>104.053</td>\n",
              "      <td>4</td>\n",
              "      <td>0.329</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                       song_popularity  ... audio_valence\n",
              "song_name                               ...              \n",
              "Party In The U.S.A.            nao_sei  ...           nan\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "I Love It (& Lil Pump)              99  ...         0.329\n",
              "\n",
              "[15 rows x 14 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 861
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mn21o_75Lq2h",
        "colab_type": "code",
        "outputId": "26f69f0f-aabf-404e-face-10bcf4ffa125",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "source": [
        "df.info()"
      ],
      "execution_count": 862,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Index: 18835 entries, Party In The U.S.A. to Coupe (feat. Rich The Kid)\n",
            "Data columns (total 14 columns):\n",
            " #   Column            Non-Null Count  Dtype  \n",
            "---  ------            --------------  -----  \n",
            " 0   song_popularity   18835 non-null  object \n",
            " 1   song_duration_ms  18835 non-null  object \n",
            " 2   acousticness      18835 non-null  object \n",
            " 3   danceability      18835 non-null  object \n",
            " 4   energy            18835 non-null  object \n",
            " 5   instrumentalness  18835 non-null  object \n",
            " 6   key               18835 non-null  float64\n",
            " 7   liveness          18835 non-null  object \n",
            " 8   loudness          18835 non-null  object \n",
            " 9   audio_mode        18835 non-null  object \n",
            " 10  speechiness       18835 non-null  object \n",
            " 11  tempo             18835 non-null  object \n",
            " 12  time_signature    18835 non-null  object \n",
            " 13  audio_valence     18834 non-null  float64\n",
            "dtypes: float64(2), object(12)\n",
            "memory usage: 2.2+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RTi_ayHSLq-S",
        "colab_type": "code",
        "outputId": "e5cda6bd-0f66-48b0-df8f-406dec718ef8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "df.describe()"
      ],
      "execution_count": 863,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>key</th>\n",
              "      <th>audio_valence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>18835.000</td>\n",
              "      <td>18834.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>5.289</td>\n",
              "      <td>0.528</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3.615</td>\n",
              "      <td>0.245</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>2.000</td>\n",
              "      <td>0.335</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>5.000</td>\n",
              "      <td>0.526</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>8.000</td>\n",
              "      <td>0.725</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>11.000</td>\n",
              "      <td>0.984</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            key  audio_valence\n",
              "count 18835.000      18834.000\n",
              "mean      5.289          0.528\n",
              "std       3.615          0.245\n",
              "min       0.000          0.000\n",
              "25%       2.000          0.335\n",
              "50%       5.000          0.526\n",
              "75%       8.000          0.725\n",
              "max      11.000          0.984"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 863
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F7BmNzPlhchL",
        "colab_type": "text"
      },
      "source": [
        "# **Limpeza dos dados:**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uA4oy4R2sWiI",
        "colab_type": "text"
      },
      "source": [
        "Apenas com as primeiras impressões podemos perceber certos problemas que podem agravar uma análise plena desse DataSet, como por exemplo: linhas repetidas, valores de tempo negativo, dados faltantes, diferentes formatações, valores fora dos limites cabíveis e variáveis numéricas armazenadas como strings.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fE3dVYcAtQXP",
        "colab_type": "code",
        "outputId": "8e941861-b475-4a28-fa0d-4cfdc7daa23f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        }
      },
      "source": [
        "duplicados = df[df.duplicated(keep='first')]\n",
        "print(duplicados)"
      ],
      "execution_count": 864,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                              song_popularity  ... audio_valence\n",
            "song_name                                      ...              \n",
            "I Love It (& Lil Pump)                     99  ...         0.329\n",
            "I Love It (& Lil Pump)                     99  ...         0.329\n",
            "I Love It (& Lil Pump)                     99  ...         0.329\n",
            "I Love It (& Lil Pump)                     99  ...         0.329\n",
            "I Love It (& Lil Pump)                     99  ...         0.329\n",
            "...                                       ...  ...           ...\n",
            "Kamakaze                                    0  ...         0.596\n",
            "Coupe (feat. Rich The Kid)                  0  ...         0.851\n",
            "Jet Lag (feat. Young Scooter)               0  ...         0.493\n",
            "John Madden 2                               0  ...         0.409\n",
            "Coupe (feat. Rich The Kid)                  0  ...         0.851\n",
            "\n",
            "[3905 rows x 14 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ynkIgozvlej",
        "colab_type": "text"
      },
      "source": [
        "Existem muitos dados repetidos, vamos tirá-los para não comprometer o nosso DataFrame"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LTOsMvf1vr93",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.drop_duplicates(keep='first', inplace=True) #Joga fora repetições (exceto a primeira ocorrência)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zIWZv1bFieYQ",
        "colab_type": "text"
      },
      "source": [
        "Vamos trocar o 'nao_sei', por NaN (Not a Number), valor ester usado para representar um valor numérico indefinido ou irrepresentável. (é útil para tratar de dados faltantes)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hw1ozRHklMVi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df = df.replace(['nao_sei'], np.nan) #Troca 'nao_sei' por NaN"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ht3DQ31qlWQu",
        "colab_type": "text"
      },
      "source": [
        "Vamos corrigir um pouco essa formatação"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5eHMS7HulW9T",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pd.set_option('display.float_format', lambda x: '%.3f' % x) #floats ficam com 3 casas depois da vírgula"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zo4M6TNFicZL",
        "colab_type": "text"
      },
      "source": [
        "Essas linhas de código abaixo servem para remover algumas inconsistências do DataSet, como por exemplo: as unidades estranhas, números fora da formatação e alguns números impossíveis dentro do intervalo."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oBk3e5DFBoqn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df['acousticness']       = df['acousticness'].str.strip('kg')\n",
        "df['danceability']       = df['danceability'].str.strip('mol/L')\n",
        "df['song_duration_ms']   = df['song_duration_ms'].str.strip('kg')\n",
        "df['acousticness']       = df['acousticness'].str.strip('mol/L')\n",
        "df['key']                = df['key'].replace([0.177], np.nan)\n",
        "df['audio_mode']         = df['audio_mode'].replace(['0.105'], np.nan)\n",
        "df['speechiness']        = df['speechiness'].replace(['0.nao_sei'], np.nan)\n",
        "df['time_signature']     = df['time_signature'].replace(['0.7', '2800000000'], np.nan)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IGOFHtzajKmm",
        "colab_type": "text"
      },
      "source": [
        "Essa sequência abaixo serve para converter nossas colunas para o verdadeiro tipo, dado que nas primeiras impressões vimos classes numéricas e categóricas sendo tratadas da forma indevida. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ux9e29OR-TFs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df['song_popularity']    = df['song_popularity'].astype('category')\n",
        "df['song_duration_ms']   = df['song_duration_ms'].astype('float')\n",
        "df['acousticness']       = df['acousticness'].astype('float')\n",
        "df['danceability']       = df['danceability'].astype('float')\n",
        "df['energy']             = df['energy'].astype('float')\n",
        "df['instrumentalness']   = df['instrumentalness'].astype('float')\n",
        "df['key']                = df['key'].astype('category')\n",
        "df['liveness']           = df['liveness'].astype('float')\n",
        "df['loudness']           = df['loudness'].astype('float')\n",
        "df['audio_mode']         = df['audio_mode'].astype('category')\n",
        "df['speechiness']        = df['speechiness'].astype('float')\n",
        "df['tempo']              = df['tempo'].astype('float')\n",
        "df['time_signature']     = df['time_signature'].astype('category')\n",
        "df['audio_valence']      = df['audio_valence'].astype('float')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pRwt7pfyjxct",
        "colab_type": "text"
      },
      "source": [
        "# **Removendo outliers**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "snTmNs3olzYY",
        "colab_type": "text"
      },
      "source": [
        "Outliers são dados que fogem muito o escopo do nosso trabalho e senso de plausível para uma análise geral, drenamos esses dados para modelos de inteligência artificial poderem de uma forma mais geral ter uma melhor acurácia na hora de fazer predições. Outliers entram como eventos impossíveis ou muito fora de proporções para análises consistentes."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B4Giwkc6mMB9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.sort_index(inplace=True) #Arrumando a ordem dos indexes"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RIZENZT_KWby",
        "colab_type": "code",
        "outputId": "84d86b11-eead-4ef9-e67f-48d1c7a212b1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 317
        }
      },
      "source": [
        "df.describe()"
      ],
      "execution_count": 871,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>song_duration_ms</th>\n",
              "      <th>acousticness</th>\n",
              "      <th>danceability</th>\n",
              "      <th>energy</th>\n",
              "      <th>instrumentalness</th>\n",
              "      <th>liveness</th>\n",
              "      <th>loudness</th>\n",
              "      <th>speechiness</th>\n",
              "      <th>tempo</th>\n",
              "      <th>audio_valence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>14930.000</td>\n",
              "      <td>14930.000</td>\n",
              "      <td>14930.000</td>\n",
              "      <td>14929.000</td>\n",
              "      <td>14928.000</td>\n",
              "      <td>14926.000</td>\n",
              "      <td>14929.000</td>\n",
              "      <td>14929.000</td>\n",
              "      <td>14929.000</td>\n",
              "      <td>14929.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-1240833227.180</td>\n",
              "      <td>0.270</td>\n",
              "      <td>0.625</td>\n",
              "      <td>0.640</td>\n",
              "      <td>0.093</td>\n",
              "      <td>0.180</td>\n",
              "      <td>-7.678</td>\n",
              "      <td>0.106</td>\n",
              "      <td>121.097</td>\n",
              "      <td>0.527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>151642153992.694</td>\n",
              "      <td>0.298</td>\n",
              "      <td>0.158</td>\n",
              "      <td>0.222</td>\n",
              "      <td>0.254</td>\n",
              "      <td>0.162</td>\n",
              "      <td>4.020</td>\n",
              "      <td>0.803</td>\n",
              "      <td>29.059</td>\n",
              "      <td>0.248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-18528908788662.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>-8.575</td>\n",
              "      <td>-38.768</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>183932.000</td>\n",
              "      <td>0.024</td>\n",
              "      <td>0.524</td>\n",
              "      <td>0.496</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.093</td>\n",
              "      <td>-9.390</td>\n",
              "      <td>0.037</td>\n",
              "      <td>98.117</td>\n",
              "      <td>0.332</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>211839.500</td>\n",
              "      <td>0.139</td>\n",
              "      <td>0.636</td>\n",
              "      <td>0.672</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.122</td>\n",
              "      <td>-6.751</td>\n",
              "      <td>0.054</td>\n",
              "      <td>120.016</td>\n",
              "      <td>0.527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>244724.500</td>\n",
              "      <td>0.458</td>\n",
              "      <td>0.740</td>\n",
              "      <td>0.818</td>\n",
              "      <td>0.005</td>\n",
              "      <td>0.224</td>\n",
              "      <td>-4.991</td>\n",
              "      <td>0.113</td>\n",
              "      <td>139.941</td>\n",
              "      <td>0.728</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>1799346.000</td>\n",
              "      <td>0.996</td>\n",
              "      <td>0.987</td>\n",
              "      <td>0.999</td>\n",
              "      <td>10.000</td>\n",
              "      <td>0.986</td>\n",
              "      <td>1.585</td>\n",
              "      <td>97.420</td>\n",
              "      <td>242.318</td>\n",
              "      <td>0.984</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         song_duration_ms  acousticness  ...     tempo  audio_valence\n",
              "count           14930.000     14930.000  ... 14929.000      14929.000\n",
              "mean      -1240833227.180         0.270  ...   121.097          0.527\n",
              "std      151642153992.694         0.298  ...    29.059          0.248\n",
              "min   -18528908788662.000         0.000  ...     0.000          0.000\n",
              "25%            183932.000         0.024  ...    98.117          0.332\n",
              "50%            211839.500         0.139  ...   120.016          0.527\n",
              "75%            244724.500         0.458  ...   139.941          0.728\n",
              "max           1799346.000         0.996  ...   242.318          0.984\n",
              "\n",
              "[8 rows x 10 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 871
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rVKwbFEEI1yk",
        "colab_type": "text"
      },
      "source": [
        "Podemos perceber alguns problemas ainda:\n",
        "*   Tempo negativo\n",
        "*   Músicas com mais de 30 minutos\n",
        "*   Músicas menores de 1 minuto\n",
        "*   Valores fora do range de 0.0 até 1.0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tvqm1pHXJ2Vi",
        "colab_type": "text"
      },
      "source": [
        "Vamos remover esses dados que não fazem o menor sentido."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9uSlu0qXJ3SL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.loc[df['song_duration_ms'] < 0, 'song_duration_ms'] = np.nan\n",
        "df.loc[df['instrumentalness'] > 1, 'instrumentalness'] = np.nan\n",
        "df.loc[df['liveness'] < 0, 'liveness'] = np.nan\n",
        "df.loc[df['loudness'] > 0, 'loudness'] = np.nan\n",
        "df.loc[df['speechiness'] > 1, 'speechiness'] = np.nan"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UsMj8PeIKtss",
        "colab_type": "text"
      },
      "source": [
        "Para lidarmos com os outliers, vamos primeiramente abordar uma forma de identificá-los e depois vamos removê-los."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "erqNBgsaI7Xz",
        "colab_type": "text"
      },
      "source": [
        "Para identificarmos podemos criar um intervalo de confiança desta forma:\n",
        "\n",
        "*   Limite_inferior = média_coluna - 2.7 * (desvio_padrao_coluna)\n",
        "*   Limite_superior = média_coluna + 2.7 * (desvio_padrao_coluna)\n",
        "\n",
        "Obs.: Vou fazer isso com todas as nossas variáveis numéricas\n",
        "\n",
        "Isso fará com que foquemos apenas em dados entre o Limite_inferior e Limite_superior e, valores fora deste intervalo serão tratados como NaN.\n",
        "\n",
        "A teoria por trás disso pode ser visualizada nessa figura da Wikipedia sobre distribuição normal: [imagem](https://pt.wikipedia.org/wiki/Distribui%C3%A7%C3%A3o_normal#/media/Ficheiro:Boxplot_vs_PDF.svg)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iFG3Pbihysom",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def exclui_outliers(DataFrame, col_name):\n",
        "  intervalo = 2.7*DataFrame[col_name].std()\n",
        "  media = DataFrame[col_name].mean()\n",
        "  DataFrame.loc[df[col_name] < (media - intervalo), col_name] = np.nan\n",
        "  DataFrame.loc[df[col_name] > (media + intervalo), col_name] = np.nan"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V31SBZ3q0c9t",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "float_cols = ['song_duration_ms', 'acousticness', 'danceability',\n",
        "              'energy', 'instrumentalness', 'liveness', \n",
        "              'loudness', 'speechiness', 'tempo','audio_valence']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WZ4VFggLzzpt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for col in float_cols:\n",
        "  exclui_outliers(df, col)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o5fCizKK5rvF",
        "colab_type": "text"
      },
      "source": [
        "# Trabalhando com valores faltantes"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FXFJsWB-LS_e",
        "colab_type": "text"
      },
      "source": [
        "Por mais que dê muita vontade de retirar todos os dados faltantes, fazer isso acarretaria na perda de cerca de 2300 linhas, o que para um DataSet de aproximadamente 15000 entradas (já excluindo as repetidas) é um pouco alto demais (~15% dos dados).\n",
        "\n",
        "Portanto vamos tentar analisar um pouco os nossos dados para saber o que fazer com os dados faltantes em cada caso."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8T_FK2iV5tSs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#df.dropna(inplace=True) #Remove NaN's\n",
        "#não vamos utilizar você hoje"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eUJUto_DOeOs",
        "colab_type": "text"
      },
      "source": [
        "Vamos dar mais uma avaliada no DataSet até o momento e plotar alguns gráficos para vermos o que está acontecendo."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fD-Jk173kok0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sns.set(style=\"ticks\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TNao5cOyb916",
        "colab_type": "code",
        "outputId": "617f7436-7a2a-4d83-c4f5-03ac7119f2fa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "source": [
        "df.info()"
      ],
      "execution_count": 878,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Index: 14930 entries, \"40\" - Live Version 1983 to 피 땀 눈물\n",
            "Data columns (total 14 columns):\n",
            " #   Column            Non-Null Count  Dtype   \n",
            "---  ------            --------------  -----   \n",
            " 0   song_popularity   14929 non-null  category\n",
            " 1   song_duration_ms  14681 non-null  float64 \n",
            " 2   acousticness      14930 non-null  float64 \n",
            " 3   danceability      14824 non-null  float64 \n",
            " 4   energy            14805 non-null  float64 \n",
            " 5   instrumentalness  13926 non-null  float64 \n",
            " 6   key               14929 non-null  category\n",
            " 7   liveness          14420 non-null  float64 \n",
            " 8   loudness          14583 non-null  float64 \n",
            " 9   audio_mode        14928 non-null  category\n",
            " 10  speechiness       14485 non-null  float64 \n",
            " 11  tempo             14814 non-null  float64 \n",
            " 12  time_signature    14927 non-null  category\n",
            " 13  audio_valence     14929 non-null  float64 \n",
            "dtypes: category(4), float64(10)\n",
            "memory usage: 1.3+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QELE1VjNOrlh",
        "colab_type": "text"
      },
      "source": [
        "Podemos perceber que existem, poucos dados faltantes nas maioria das colunas, mas vamos olhar um pouco mais de perto."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o2x68kzdZPjx",
        "colab_type": "code",
        "outputId": "2d324467-fac6-4cdf-b5f4-b785382f3334",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 317
        }
      },
      "source": [
        "df.describe()"
      ],
      "execution_count": 879,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>song_duration_ms</th>\n",
              "      <th>acousticness</th>\n",
              "      <th>danceability</th>\n",
              "      <th>energy</th>\n",
              "      <th>instrumentalness</th>\n",
              "      <th>liveness</th>\n",
              "      <th>loudness</th>\n",
              "      <th>speechiness</th>\n",
              "      <th>tempo</th>\n",
              "      <th>audio_valence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>14681.000</td>\n",
              "      <td>14930.000</td>\n",
              "      <td>14824.000</td>\n",
              "      <td>14805.000</td>\n",
              "      <td>13926.000</td>\n",
              "      <td>14420.000</td>\n",
              "      <td>14583.000</td>\n",
              "      <td>14485.000</td>\n",
              "      <td>14814.000</td>\n",
              "      <td>14929.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>214852.429</td>\n",
              "      <td>0.270</td>\n",
              "      <td>0.628</td>\n",
              "      <td>0.645</td>\n",
              "      <td>0.036</td>\n",
              "      <td>0.162</td>\n",
              "      <td>-7.319</td>\n",
              "      <td>0.088</td>\n",
              "      <td>120.491</td>\n",
              "      <td>0.527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>49794.983</td>\n",
              "      <td>0.298</td>\n",
              "      <td>0.153</td>\n",
              "      <td>0.215</td>\n",
              "      <td>0.120</td>\n",
              "      <td>0.106</td>\n",
              "      <td>3.217</td>\n",
              "      <td>0.080</td>\n",
              "      <td>28.193</td>\n",
              "      <td>0.248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>53066.000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.198</td>\n",
              "      <td>0.041</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.011</td>\n",
              "      <td>-18.515</td>\n",
              "      <td>0.000</td>\n",
              "      <td>46.591</td>\n",
              "      <td>0.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>183453.000</td>\n",
              "      <td>0.024</td>\n",
              "      <td>0.527</td>\n",
              "      <td>0.500</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.092</td>\n",
              "      <td>-9.151</td>\n",
              "      <td>0.037</td>\n",
              "      <td>98.040</td>\n",
              "      <td>0.332</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>211193.000</td>\n",
              "      <td>0.139</td>\n",
              "      <td>0.637</td>\n",
              "      <td>0.674</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.119</td>\n",
              "      <td>-6.663</td>\n",
              "      <td>0.053</td>\n",
              "      <td>120.002</td>\n",
              "      <td>0.527</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>242333.000</td>\n",
              "      <td>0.458</td>\n",
              "      <td>0.741</td>\n",
              "      <td>0.819</td>\n",
              "      <td>0.001</td>\n",
              "      <td>0.205</td>\n",
              "      <td>-4.957</td>\n",
              "      <td>0.103</td>\n",
              "      <td>139.512</td>\n",
              "      <td>0.728</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>386946.000</td>\n",
              "      <td>0.996</td>\n",
              "      <td>0.987</td>\n",
              "      <td>0.999</td>\n",
              "      <td>0.741</td>\n",
              "      <td>0.572</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.378</td>\n",
              "      <td>199.402</td>\n",
              "      <td>0.984</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       song_duration_ms  acousticness  ...     tempo  audio_valence\n",
              "count         14681.000     14930.000  ... 14814.000      14929.000\n",
              "mean         214852.429         0.270  ...   120.491          0.527\n",
              "std           49794.983         0.298  ...    28.193          0.248\n",
              "min           53066.000         0.000  ...    46.591          0.000\n",
              "25%          183453.000         0.024  ...    98.040          0.332\n",
              "50%          211193.000         0.139  ...   120.002          0.527\n",
              "75%          242333.000         0.458  ...   139.512          0.728\n",
              "max          386946.000         0.996  ...   199.402          0.984\n",
              "\n",
              "[8 rows x 10 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 879
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0CMCObLuPEsc",
        "colab_type": "text"
      },
      "source": [
        "Podemos notar algumas coisas estranhas em acousticness, instrumentalness, liveness e speechiness. Principalmente no fato das distribuições destas variáveis.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xcU0xHPanP82",
        "colab_type": "code",
        "outputId": "9cd5fbe8-bf1f-42a8-fa1c-68c970b5ad53",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 581
        }
      },
      "source": [
        "plt.figure(figsize=(10,8))\n",
        "cols = ['acousticness', 'instrumentalness', 'liveness', 'speechiness']\n",
        "for i in range(len(cols)):\n",
        "  plt.subplot(2, 2, i+1)\n",
        "  sns.distplot(df[cols[i]])\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 880,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hhh9puMHWZBr",
        "colab_type": "text"
      },
      "source": [
        "Podemos perceber algumas coisas interessantes, realmente elas tem uma distribuição estranha e mais concentrada no começo dos dados, mas especificamente para intrumentalness quase não existem valores que sejam muito longe de 0. Ou seja, essa coluna está mais atrapalhando do que ajudando, uma vez que são muito poucas as ocorrências de valores fora diferentes de 0, para comparar se isso tem ou não impacto direto com a popularidade da música.\n",
        "\n",
        "Dito isso:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nd8GXi_NZJNO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.drop('instrumentalness', axis = 1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xPA5lk4ki6wn",
        "colab_type": "text"
      },
      "source": [
        "Certo, tendo em vista a distribuição dessas colunas plotadas, vamos tratar os nossos dados faltantes dessas colunas seguindo a tendência \"deslocada\" desse gráfico e colocar valores faltantes como membros da moda."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bhgrzGWUjelm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for column in ['acousticness', 'liveness', 'speechiness']:\n",
        "    df[column].fillna(df[column].mode()[0], inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xm2F2vOTo5ip",
        "colab_type": "text"
      },
      "source": [
        "Vamos agora dar uma olhada nas nossas outras variáveis numéricas e ver como podemos tratar os dados faltantes."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qyNT3UcEpD9a",
        "colab_type": "code",
        "outputId": "c9496dd8-41bc-4bb4-fc2d-0878511afa78",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 581
        }
      },
      "source": [
        "plt.figure(figsize=(10,8))\n",
        "cols = ['song_duration_ms',  'danceability', 'energy', \n",
        "                'loudness', 'audio_valence', 'tempo']\n",
        "for i in range(len(cols)):\n",
        "  plt.subplot(2, 3, i+1)\n",
        "  sns.distplot(df[cols[i]])\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 883,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 6 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DuSSxlPZsqHT",
        "colab_type": "text"
      },
      "source": [
        "Como essas outras variáveis apresentam uma distribuição mais equilibrada dos seus dados, vou distribuir os dados faltantes como membros da média.\n",
        "\n",
        "Como 'tempo' apresenta um comportamento estranho com dois picos, vou deixar para tratá-lo depois de analisar as variáveis categóricas."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WfE77Z9xs_Qk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for column in ['song_duration_ms',  'danceability', 'energy', \n",
        "                'loudness'       , 'audio_valence']:\n",
        "    df[column].fillna(df[column].mean(), inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ifBxYhQwwTmE",
        "colab_type": "text"
      },
      "source": [
        "Agora vamos olhar as nossas variáveis categóricas."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dg5-6RrswYRj",
        "colab_type": "code",
        "outputId": "64e33714-e8e9-42f8-8422-33b44c1ac4d7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 581
        }
      },
      "source": [
        "plt.figure(figsize=(10,8))\n",
        "cols = ['audio_mode', 'time_signature', 'key']\n",
        "for i in range(len(cols)):\n",
        "  plt.subplot(3, 1, i+1)\n",
        "  sns.countplot(df[cols[i]])\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 885,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8U2fyfoKyyG7",
        "colab_type": "text"
      },
      "source": [
        "Da mesma forma que em 'instrumentalness', a variável time_signature é extremamente deslocada para valores do tipo 4, ou seja, não utilizaremos essa coluna para nossa análise.\n",
        "\n",
        "Dessa forma:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SEfNSkIKxGUk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.drop('time_signature', axis = 1, inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MyDe93I8zvco",
        "colab_type": "text"
      },
      "source": [
        "Vamos dar mais uma olhada nos nossos dados, para ver como eles se encontram."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NCpGRD8luhxj",
        "colab_type": "code",
        "outputId": "6fe6eb00-640f-42cb-8f97-5e7259476016",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 340
        }
      },
      "source": [
        "df.info()"
      ],
      "execution_count": 887,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Index: 14930 entries, \"40\" - Live Version 1983 to 피 땀 눈물\n",
            "Data columns (total 12 columns):\n",
            " #   Column            Non-Null Count  Dtype   \n",
            "---  ------            --------------  -----   \n",
            " 0   song_popularity   14929 non-null  category\n",
            " 1   song_duration_ms  14930 non-null  float64 \n",
            " 2   acousticness      14930 non-null  float64 \n",
            " 3   danceability      14930 non-null  float64 \n",
            " 4   energy            14930 non-null  float64 \n",
            " 5   key               14929 non-null  category\n",
            " 6   liveness          14930 non-null  float64 \n",
            " 7   loudness          14930 non-null  float64 \n",
            " 8   audio_mode        14928 non-null  category\n",
            " 9   speechiness       14930 non-null  float64 \n",
            " 10  tempo             14814 non-null  float64 \n",
            " 11  audio_valence     14930 non-null  float64 \n",
            "dtypes: category(3), float64(9)\n",
            "memory usage: 1.8+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2RqBdHpmz2O1",
        "colab_type": "text"
      },
      "source": [
        "Ótimo, a maior parte do trabalho está feita.\n",
        "\n",
        "Vamos tratar agora da nossa variável 'tempo', ela apresentava uma comportamento um tanto irregular, então vou só excluir as linhas faltantes, uma vez que nesse caso estaríamos deletando cerca de 120 linhas (~0.7% dos dados [muito mais aceitável])."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ISEi7WZ4uMry",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df.dropna(subset=['tempo'], inplace=True) #tiro valores NaN de 'tempo'\n",
        "df.dropna(inplace=True) #removo demais linhas com dados faltantes\n",
        "#Acabamos por usar realmente esse comando, mas de uma forma\n",
        "#muito mais consiente do que se tivéssemos usado lá em cima"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dXT4B5K21fxs",
        "colab_type": "text"
      },
      "source": [
        "Com isso nossos dados agora estão limpos de uma forma eficiente, vamos para o que interessa: as relações entre nossas features e quais características geram uma música com alta popularidade"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tPuMP0Arc4p8",
        "colab_type": "text"
      },
      "source": [
        "# Análise do target"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7jLF6jZchtmk",
        "colab_type": "text"
      },
      "source": [
        "Como eu optei por tratar a coluna target ('song_popularity') como uma variável categórica vamos fazer um processo de clusterização."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kuTsHCmLh8f7",
        "colab_type": "text"
      },
      "source": [
        "Obs.: Clustering é um conjunto de técnicas de agrupamento de dados com base em sua semelhança."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gyCKl_UFiKHA",
        "colab_type": "text"
      },
      "source": [
        "Dessa forma, podemos trabalhar com nossos dados de uma forma mais geral, ao invés de cada popularidade ter uma categoria correnpondente, podemos agrupar um conjunto de popularidades com mesmas características em um cluster."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WnuE9nvhkNCb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.cluster import KMeans"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fYgzuhPIkTwK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "clustering = KMeans(n_clusters=5) #define o número de grupos que eu quero, nesse caso, 5\n",
        "df['song_popularity_cluster'] = clustering.fit_predict(df['song_popularity'].astype(int).values.reshape(-1, 1))\n",
        "df['song_popularity'] = df['song_popularity'].astype(int)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-6eOxSzxNYf7",
        "colab_type": "text"
      },
      "source": [
        "Essa primeira parte serve para adicionarmos uma coluna 'song_popularity_cluster' que é nossa classificação já clusterizada, agora vamos analisar umas tabelas, da relação da popularidade com as variáveis numéricas."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C2Ax5D21aXI8",
        "colab_type": "text"
      },
      "source": [
        "**Clusterização:**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PMNjfiZrmX4M",
        "colab_type": "text"
      },
      "source": [
        "Primeiramente gostaria de mostrar esse gráfico da distribuição da popularidade, pois ele ~é muito lindo~ é bom para visualizarmos a distribuição, com ele conseguimos perceber que ter uma popularidade próxima de 100 é algo melhor, uma vez que menos músicas conseguem ter uma alta popularidade."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S7_yMSFAox1g",
        "colab_type": "text"
      },
      "source": [
        "Isso é mais para não acontecer confusão em relação à outra métrica de músicas, o ranking. Neste geralmente ter um menor número significa uma melhor classificação (famoso top 1 ou top 10 que escutamos)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gT01VFw-lVAv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 305
        },
        "outputId": "1e4042fb-f3b1-4917-a12b-ef894db0170a"
      },
      "source": [
        "plt.figure(figsize=(16,5))\n",
        "sns.countplot(x='song_popularity', data=df)\n",
        "plt.xticks(rotation=90, horizontalalignment='right', fontweight='bold', fontsize='small')\n",
        "plt.yticks(fontweight='bold', horizontalalignment='right')\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 891,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Hv7zJn-5pZ8k",
        "colab_type": "text"
      },
      "source": [
        "Agora vamos para as tabelas, primeiro vamos ver as relações com as variáveis numéricas e depois com as categóricas."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j43SvUNoPj1p",
        "colab_type": "code",
        "outputId": "c3ffbdb4-c663-479b-8c82-b3a6d216ee36",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        }
      },
      "source": [
        "df[['song_popularity','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 892,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">song_popularity</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>45.435</td>\n",
              "      <td>4.527</td>\n",
              "      <td>37.000</td>\n",
              "      <td>42.000</td>\n",
              "      <td>46.000</td>\n",
              "      <td>49.000</td>\n",
              "      <td>52.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>74.878</td>\n",
              "      <td>5.985</td>\n",
              "      <td>68.000</td>\n",
              "      <td>70.000</td>\n",
              "      <td>73.000</td>\n",
              "      <td>78.000</td>\n",
              "      <td>100.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>7.170</td>\n",
              "      <td>5.682</td>\n",
              "      <td>0.000</td>\n",
              "      <td>2.000</td>\n",
              "      <td>6.000</td>\n",
              "      <td>12.000</td>\n",
              "      <td>17.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>59.731</td>\n",
              "      <td>4.265</td>\n",
              "      <td>53.000</td>\n",
              "      <td>56.000</td>\n",
              "      <td>60.000</td>\n",
              "      <td>63.000</td>\n",
              "      <td>67.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>28.172</td>\n",
              "      <td>5.426</td>\n",
              "      <td>18.000</td>\n",
              "      <td>24.000</td>\n",
              "      <td>29.000</td>\n",
              "      <td>33.000</td>\n",
              "      <td>36.000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        song_popularity               ...                      \n",
              "                                  count   mean   std  ...    50%    75%     max\n",
              "song_popularity_cluster                               ...                      \n",
              "0                              4098.000 45.435 4.527  ... 46.000 49.000  52.000\n",
              "1                              2620.000 74.878 5.985  ... 73.000 78.000 100.000\n",
              "2                              1449.000  7.170 5.682  ...  6.000 12.000  17.000\n",
              "3                              4527.000 59.731 4.265  ... 60.000 63.000  67.000\n",
              "4                              2119.000 28.172 5.426  ... 29.000 33.000  36.000\n",
              "\n",
              "[5 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 892
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "01ZwQINoafWL",
        "colab_type": "text"
      },
      "source": [
        "Essa primeira tabela é justamente para analisarmos onde foi feito esse corte, eu decidi trabalhar com 5 clusters. Segundo a análise de proximidade do KMeans, esses dados foram divididos dessa forma:\n",
        "\n",
        "0.   Popularidade entre: 37 e 52.....(médias)\n",
        "1.   Popularidade entre: 67 e 100...(melhores)\n",
        "2.   Popularidade entre:  0 e 17.......(piores)\n",
        "3.   Popularidade entre: 53 e 67.....(médias-altas)\n",
        "4.   Popularidade entre: 18 e 37.....(médias-baixas)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AlGlFqjlqb2u",
        "colab_type": "text"
      },
      "source": [
        "**Duração:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P3U7q9JYq-3d",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "0bedd632-579e-417f-e144-0baaaed484f9"
      },
      "source": [
        "df[['song_duration_ms','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 893,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">song_duration_ms</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>211752.208</td>\n",
              "      <td>52238.136</td>\n",
              "      <td>54539.000</td>\n",
              "      <td>179561.250</td>\n",
              "      <td>209284.000</td>\n",
              "      <td>239516.250</td>\n",
              "      <td>386026.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>219732.798</td>\n",
              "      <td>44681.578</td>\n",
              "      <td>67000.000</td>\n",
              "      <td>191706.500</td>\n",
              "      <td>214852.429</td>\n",
              "      <td>243843.000</td>\n",
              "      <td>386906.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>218187.203</td>\n",
              "      <td>50822.465</td>\n",
              "      <td>53066.000</td>\n",
              "      <td>187040.000</td>\n",
              "      <td>213800.000</td>\n",
              "      <td>245195.000</td>\n",
              "      <td>384850.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>215479.882</td>\n",
              "      <td>47466.786</td>\n",
              "      <td>55213.000</td>\n",
              "      <td>185605.500</td>\n",
              "      <td>212480.000</td>\n",
              "      <td>242380.000</td>\n",
              "      <td>386373.000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>211759.185</td>\n",
              "      <td>51509.312</td>\n",
              "      <td>55720.000</td>\n",
              "      <td>177613.000</td>\n",
              "      <td>209440.000</td>\n",
              "      <td>238266.500</td>\n",
              "      <td>386946.000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        song_duration_ms             ...                      \n",
              "                                   count       mean  ...        75%        max\n",
              "song_popularity_cluster                              ...                      \n",
              "0                               4098.000 211752.208  ... 239516.250 386026.000\n",
              "1                               2620.000 219732.798  ... 243843.000 386906.000\n",
              "2                               1449.000 218187.203  ... 245195.000 384850.000\n",
              "3                               4527.000 215479.882  ... 242380.000 386373.000\n",
              "4                               2119.000 211759.185  ... 238266.500 386946.000\n",
              "\n",
              "[5 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 893
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gfDRmGVEsga0",
        "colab_type": "text"
      },
      "source": [
        "Percebemos aqui que a duração não tem muita relação com a popularidade da música, observe que entre as piores (2) e melhores (1) não há uma diferença significativa de tempo para tirarmos alguma conclusão útil."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lS0bQq22tsDV",
        "colab_type": "text"
      },
      "source": [
        "**Acústica:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lg1rrbT_t00v",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "6a074e94-2c5c-456c-91d7-b0f1ec2141fb"
      },
      "source": [
        "df[['acousticness','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 894,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">acousticness</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.288</td>\n",
              "      <td>0.304</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.026</td>\n",
              "      <td>0.159</td>\n",
              "      <td>0.512</td>\n",
              "      <td>0.996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.219</td>\n",
              "      <td>0.254</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.028</td>\n",
              "      <td>0.114</td>\n",
              "      <td>0.318</td>\n",
              "      <td>0.995</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.228</td>\n",
              "      <td>0.287</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.009</td>\n",
              "      <td>0.083</td>\n",
              "      <td>0.372</td>\n",
              "      <td>0.996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.277</td>\n",
              "      <td>0.307</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.022</td>\n",
              "      <td>0.142</td>\n",
              "      <td>0.463</td>\n",
              "      <td>0.996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.310</td>\n",
              "      <td>0.313</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.030</td>\n",
              "      <td>0.191</td>\n",
              "      <td>0.571</td>\n",
              "      <td>0.995</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        acousticness                                          \n",
              "                               count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                       \n",
              "0                           4098.000 0.288 0.304 0.000 0.026 0.159 0.512 0.996\n",
              "1                           2620.000 0.219 0.254 0.000 0.028 0.114 0.318 0.995\n",
              "2                           1449.000 0.228 0.287 0.000 0.009 0.083 0.372 0.996\n",
              "3                           4527.000 0.277 0.307 0.000 0.022 0.142 0.463 0.996\n",
              "4                           2119.000 0.310 0.313 0.000 0.030 0.191 0.571 0.995"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 894
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Miou67dQvLhQ",
        "colab_type": "text"
      },
      "source": [
        "A acústica também não parece influenciar muito no nosso problema, podemos perceber que as músicas gostam de estar entre uma faixa de 0.2 e 0.3 de acústica (olhando a média de todas), mas se olharmos o minínimo e máximo, todas variam de 0 até algo muito próximo de 1. De qualquer forma, as melhores apresentam uma acústica um pouco menor que as outras na média."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "efM3gD6PwEGi",
        "colab_type": "text"
      },
      "source": [
        "**Danceabilidade:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oUuduXKlvvYX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "811646df-b06c-4fca-bc4a-95556d4b3181"
      },
      "source": [
        "df[['danceability','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 895,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">danceability</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.629</td>\n",
              "      <td>0.155</td>\n",
              "      <td>0.199</td>\n",
              "      <td>0.529</td>\n",
              "      <td>0.639</td>\n",
              "      <td>0.741</td>\n",
              "      <td>0.981</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.649</td>\n",
              "      <td>0.147</td>\n",
              "      <td>0.202</td>\n",
              "      <td>0.555</td>\n",
              "      <td>0.658</td>\n",
              "      <td>0.754</td>\n",
              "      <td>0.978</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.611</td>\n",
              "      <td>0.156</td>\n",
              "      <td>0.208</td>\n",
              "      <td>0.507</td>\n",
              "      <td>0.616</td>\n",
              "      <td>0.723</td>\n",
              "      <td>0.970</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.623</td>\n",
              "      <td>0.152</td>\n",
              "      <td>0.199</td>\n",
              "      <td>0.520</td>\n",
              "      <td>0.628</td>\n",
              "      <td>0.736</td>\n",
              "      <td>0.980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.628</td>\n",
              "      <td>0.150</td>\n",
              "      <td>0.198</td>\n",
              "      <td>0.532</td>\n",
              "      <td>0.634</td>\n",
              "      <td>0.740</td>\n",
              "      <td>0.987</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        danceability                                          \n",
              "                               count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                       \n",
              "0                           4098.000 0.629 0.155 0.199 0.529 0.639 0.741 0.981\n",
              "1                           2620.000 0.649 0.147 0.202 0.555 0.658 0.754 0.978\n",
              "2                           1449.000 0.611 0.156 0.208 0.507 0.616 0.723 0.970\n",
              "3                           4527.000 0.623 0.152 0.199 0.520 0.628 0.736 0.980\n",
              "4                           2119.000 0.628 0.150 0.198 0.532 0.634 0.740 0.987"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 895
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QQdQvwKWwBrr",
        "colab_type": "text"
      },
      "source": [
        "Eu confesso que fiquei um pouco surpreso, achei que danceabilidade fosse algo que tivesse uma influência maior sobre o quanto uma música é ou não popular, mas pelo que vemos isso não é algo muito constante, até aparenta existir uma certa crescente, mas é algo muito leve para podermos deter como algo impactante. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MPRQ43GfwhAU",
        "colab_type": "text"
      },
      "source": [
        "O máximo que podemos concluir é que músicas mais populares tendem a ter uma danceabilidade levemente maior, mas isso não parece ser um fator marcante, pois as médias-altas (3) tem um valor médio abaixo que a das médias-baixas (4), o que quebra essa linha de pensamento. De qualquer forma, as melhores apresentam danceabilidade levemente maiores que as outras na média."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lfvghdVyxEV0",
        "colab_type": "text"
      },
      "source": [
        "**Energia:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8v2bEwucxG6o",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "3d990724-fb10-4c9e-f559-4e18c5a69afb"
      },
      "source": [
        "df[['energy','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 896,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">energy</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.633</td>\n",
              "      <td>0.221</td>\n",
              "      <td>0.041</td>\n",
              "      <td>0.485</td>\n",
              "      <td>0.654</td>\n",
              "      <td>0.811</td>\n",
              "      <td>0.996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.659</td>\n",
              "      <td>0.190</td>\n",
              "      <td>0.043</td>\n",
              "      <td>0.538</td>\n",
              "      <td>0.683</td>\n",
              "      <td>0.807</td>\n",
              "      <td>0.997</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.679</td>\n",
              "      <td>0.210</td>\n",
              "      <td>0.067</td>\n",
              "      <td>0.534</td>\n",
              "      <td>0.711</td>\n",
              "      <td>0.851</td>\n",
              "      <td>0.996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.643</td>\n",
              "      <td>0.220</td>\n",
              "      <td>0.041</td>\n",
              "      <td>0.501</td>\n",
              "      <td>0.673</td>\n",
              "      <td>0.820</td>\n",
              "      <td>0.997</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.634</td>\n",
              "      <td>0.216</td>\n",
              "      <td>0.043</td>\n",
              "      <td>0.479</td>\n",
              "      <td>0.661</td>\n",
              "      <td>0.810</td>\n",
              "      <td>0.999</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                          energy                                          \n",
              "                           count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                   \n",
              "0                       4098.000 0.633 0.221 0.041 0.485 0.654 0.811 0.996\n",
              "1                       2620.000 0.659 0.190 0.043 0.538 0.683 0.807 0.997\n",
              "2                       1449.000 0.679 0.210 0.067 0.534 0.711 0.851 0.996\n",
              "3                       4527.000 0.643 0.220 0.041 0.501 0.673 0.820 0.997\n",
              "4                       2119.000 0.634 0.216 0.043 0.479 0.661 0.810 0.999"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 896
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eOQ5iEWr3YVB",
        "colab_type": "text"
      },
      "source": [
        "Uma linha de pensamento de que músicas mais populares também são mais energéticas até que começa a nascer quando, vemos as melhores no topo tendo uma apresentação decrescente até chegar nas piores. De qualquer forma, as piores apresentam um valor de energia mais alto na média."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KKcAymm84CpE",
        "colab_type": "text"
      },
      "source": [
        "**Fator ao vivo:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zwouZL2l4KRk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "87edd087-ffa0-4dbc-fbcb-d4cf0198268a"
      },
      "source": [
        "df[['liveness','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 899,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">liveness</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.164</td>\n",
              "      <td>0.108</td>\n",
              "      <td>0.015</td>\n",
              "      <td>0.094</td>\n",
              "      <td>0.117</td>\n",
              "      <td>0.205</td>\n",
              "      <td>0.572</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.156</td>\n",
              "      <td>0.102</td>\n",
              "      <td>0.021</td>\n",
              "      <td>0.091</td>\n",
              "      <td>0.115</td>\n",
              "      <td>0.189</td>\n",
              "      <td>0.571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.165</td>\n",
              "      <td>0.107</td>\n",
              "      <td>0.021</td>\n",
              "      <td>0.093</td>\n",
              "      <td>0.119</td>\n",
              "      <td>0.216</td>\n",
              "      <td>0.569</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.156</td>\n",
              "      <td>0.102</td>\n",
              "      <td>0.020</td>\n",
              "      <td>0.093</td>\n",
              "      <td>0.114</td>\n",
              "      <td>0.190</td>\n",
              "      <td>0.572</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.164</td>\n",
              "      <td>0.102</td>\n",
              "      <td>0.011</td>\n",
              "      <td>0.095</td>\n",
              "      <td>0.122</td>\n",
              "      <td>0.214</td>\n",
              "      <td>0.568</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        liveness                                          \n",
              "                           count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                   \n",
              "0                       4098.000 0.164 0.108 0.015 0.094 0.117 0.205 0.572\n",
              "1                       2620.000 0.156 0.102 0.021 0.091 0.115 0.189 0.571\n",
              "2                       1449.000 0.165 0.107 0.021 0.093 0.119 0.216 0.569\n",
              "3                       4527.000 0.156 0.102 0.020 0.093 0.114 0.190 0.572\n",
              "4                       2119.000 0.164 0.102 0.011 0.095 0.122 0.214 0.568"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 899
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kV8qALLZ4fcT",
        "colab_type": "text"
      },
      "source": [
        "Mesmo que o fator ao vivo não seja algo muito impactante para nosso target, podemos ver um padrão interessante, a maioria das músicas converge quanto a lançar músicas que não tem uma plateia ou algo do tipo, o que de fato é esperado. E podemos checar isso com os valores máximos sendo ainda menores que 0.6 para qualquer caso."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cR-U5D9v7ozT",
        "colab_type": "text"
      },
      "source": [
        "Interessante notar que as melhores músicas (1 e 3) ficam na faixa de 0.156, enquanto as piores (0, 2 e 4) tendem a ficar na faixa de 0.164 o que é curioso, pois mesmo existindo uma preferência geral por faixas gravadas sem uma platéia, as músicas com piores popularidades tendem a ter uma tolerância a mais por músicas gravadas com plateia, que talvez agrave sua popularidade."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uQHY4st78opl",
        "colab_type": "text"
      },
      "source": [
        "**Sonoridade:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7QXv_7s04-Oe",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "41848dbb-afeb-4f18-ab0b-dcc1bd6b73a2"
      },
      "source": [
        "df[['loudness','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 900,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">loudness</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>-7.652</td>\n",
              "      <td>3.296</td>\n",
              "      <td>-18.478</td>\n",
              "      <td>-9.557</td>\n",
              "      <td>-7.086</td>\n",
              "      <td>-5.228</td>\n",
              "      <td>-0.790</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>-6.558</td>\n",
              "      <td>2.736</td>\n",
              "      <td>-18.515</td>\n",
              "      <td>-7.913</td>\n",
              "      <td>-6.013</td>\n",
              "      <td>-4.640</td>\n",
              "      <td>-0.739</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>-7.150</td>\n",
              "      <td>3.111</td>\n",
              "      <td>-18.354</td>\n",
              "      <td>-8.843</td>\n",
              "      <td>-6.698</td>\n",
              "      <td>-4.930</td>\n",
              "      <td>-0.757</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>-7.257</td>\n",
              "      <td>3.170</td>\n",
              "      <td>-18.504</td>\n",
              "      <td>-8.952</td>\n",
              "      <td>-6.719</td>\n",
              "      <td>-4.970</td>\n",
              "      <td>-0.257</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>-7.867</td>\n",
              "      <td>3.323</td>\n",
              "      <td>-18.512</td>\n",
              "      <td>-9.974</td>\n",
              "      <td>-7.319</td>\n",
              "      <td>-5.416</td>\n",
              "      <td>-0.398</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        loudness               ...                     \n",
              "                           count   mean   std  ...    50%    75%    max\n",
              "song_popularity_cluster                        ...                     \n",
              "0                       4098.000 -7.652 3.296  ... -7.086 -5.228 -0.790\n",
              "1                       2620.000 -6.558 2.736  ... -6.013 -4.640 -0.739\n",
              "2                       1449.000 -7.150 3.111  ... -6.698 -4.930 -0.757\n",
              "3                       4527.000 -7.257 3.170  ... -6.719 -4.970 -0.257\n",
              "4                       2119.000 -7.867 3.323  ... -7.319 -5.416 -0.398\n",
              "\n",
              "[5 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 900
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qW4IA9_Q8S23",
        "colab_type": "text"
      },
      "source": [
        "Mesmo não podendo obervar nenhuma tendência ou relação clara entre popularidade e sonoridade, percebemos que as mais populares apresentam músicas levemente mais altas, isso por conta no salto da média e também pelo desvio padrão significamente menor que o das outras, o que indica maior concentração em torno da média."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tydNsbtW9DXd",
        "colab_type": "text"
      },
      "source": [
        "**Discurso:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BEDjVXgp9KRC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "21c80d29-a024-4679-e430-75b16c27bc65"
      },
      "source": [
        "df[['speechiness','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 901,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">speechiness</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.087</td>\n",
              "      <td>0.079</td>\n",
              "      <td>0.023</td>\n",
              "      <td>0.036</td>\n",
              "      <td>0.052</td>\n",
              "      <td>0.102</td>\n",
              "      <td>0.377</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.091</td>\n",
              "      <td>0.082</td>\n",
              "      <td>0.022</td>\n",
              "      <td>0.037</td>\n",
              "      <td>0.054</td>\n",
              "      <td>0.111</td>\n",
              "      <td>0.378</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.085</td>\n",
              "      <td>0.077</td>\n",
              "      <td>0.024</td>\n",
              "      <td>0.036</td>\n",
              "      <td>0.052</td>\n",
              "      <td>0.097</td>\n",
              "      <td>0.378</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.083</td>\n",
              "      <td>0.076</td>\n",
              "      <td>0.023</td>\n",
              "      <td>0.035</td>\n",
              "      <td>0.049</td>\n",
              "      <td>0.091</td>\n",
              "      <td>0.378</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.087</td>\n",
              "      <td>0.081</td>\n",
              "      <td>0.023</td>\n",
              "      <td>0.035</td>\n",
              "      <td>0.051</td>\n",
              "      <td>0.101</td>\n",
              "      <td>0.375</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        speechiness                                          \n",
              "                              count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                      \n",
              "0                          4098.000 0.087 0.079 0.023 0.036 0.052 0.102 0.377\n",
              "1                          2620.000 0.091 0.082 0.022 0.037 0.054 0.111 0.378\n",
              "2                          1449.000 0.085 0.077 0.024 0.036 0.052 0.097 0.378\n",
              "3                          4527.000 0.083 0.076 0.023 0.035 0.049 0.091 0.378\n",
              "4                          2119.000 0.087 0.081 0.023 0.035 0.051 0.101 0.375"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 901
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CkZCIIad9d6L",
        "colab_type": "text"
      },
      "source": [
        "Como o esperado nn existe muita variação nessa métrica, pois estamos trabalhando com músicas e não muito com discursos ou algo do gênero."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xcjkP54w9oLi",
        "colab_type": "text"
      },
      "source": [
        "**Ritmo:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EvOI1fgS9u1y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "503c651e-4013-48f2-ddc4-51b8e9f2be8c"
      },
      "source": [
        "df[['tempo','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 902,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">tempo</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>120.437</td>\n",
              "      <td>28.775</td>\n",
              "      <td>47.953</td>\n",
              "      <td>97.058</td>\n",
              "      <td>119.997</td>\n",
              "      <td>139.928</td>\n",
              "      <td>199.402</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>120.127</td>\n",
              "      <td>27.638</td>\n",
              "      <td>57.178</td>\n",
              "      <td>98.072</td>\n",
              "      <td>118.941</td>\n",
              "      <td>138.193</td>\n",
              "      <td>199.373</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>122.454</td>\n",
              "      <td>27.702</td>\n",
              "      <td>52.181</td>\n",
              "      <td>101.015</td>\n",
              "      <td>122.311</td>\n",
              "      <td>140.024</td>\n",
              "      <td>198.870</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>119.827</td>\n",
              "      <td>28.234</td>\n",
              "      <td>46.591</td>\n",
              "      <td>97.799</td>\n",
              "      <td>119.990</td>\n",
              "      <td>137.395</td>\n",
              "      <td>198.670</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>121.111</td>\n",
              "      <td>27.941</td>\n",
              "      <td>59.906</td>\n",
              "      <td>99.916</td>\n",
              "      <td>120.031</td>\n",
              "      <td>139.889</td>\n",
              "      <td>198.372</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                           tempo                 ...                        \n",
              "                           count    mean    std  ...     50%     75%     max\n",
              "song_popularity_cluster                          ...                        \n",
              "0                       4098.000 120.437 28.775  ... 119.997 139.928 199.402\n",
              "1                       2620.000 120.127 27.638  ... 118.941 138.193 199.373\n",
              "2                       1449.000 122.454 27.702  ... 122.311 140.024 198.870\n",
              "3                       4527.000 119.827 28.234  ... 119.990 137.395 198.670\n",
              "4                       2119.000 121.111 27.941  ... 120.031 139.889 198.372\n",
              "\n",
              "[5 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 902
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FGWUPb5_94xA",
        "colab_type": "text"
      },
      "source": [
        "Não parece haver uma relação muito clara entre o ritmo e popularidade, o que é esperado, pois diferentes estilos musicais tem seu ritmo próprio. Os valores para as 5 categorias são distribuídos de forma muito parecida, é legal notar apenas essa faixa entre 45 e 200 BPM que as músicas estão, em sua maioria, contidas."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0m0xecxU-xgw",
        "colab_type": "text"
      },
      "source": [
        "**Valência:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "is-Dazop-w90",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "8efac39f-e917-4a05-94f5-e22fa4fc296d"
      },
      "source": [
        "df[['audio_valence','song_popularity_cluster']].groupby('song_popularity_cluster').describe()"
      ],
      "execution_count": 903,
      "outputs": [
        {
          "output_type": "execute_result",
          "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",
              "\n",
              "    .dataframe thead tr:last-of-type th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th colspan=\"8\" halign=\"left\">audio_valence</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "      <th>mean</th>\n",
              "      <th>std</th>\n",
              "      <th>min</th>\n",
              "      <th>25%</th>\n",
              "      <th>50%</th>\n",
              "      <th>75%</th>\n",
              "      <th>max</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_popularity_cluster</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4098.000</td>\n",
              "      <td>0.527</td>\n",
              "      <td>0.257</td>\n",
              "      <td>0.023</td>\n",
              "      <td>0.325</td>\n",
              "      <td>0.532</td>\n",
              "      <td>0.740</td>\n",
              "      <td>0.982</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2620.000</td>\n",
              "      <td>0.524</td>\n",
              "      <td>0.233</td>\n",
              "      <td>0.035</td>\n",
              "      <td>0.341</td>\n",
              "      <td>0.522</td>\n",
              "      <td>0.708</td>\n",
              "      <td>0.980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1449.000</td>\n",
              "      <td>0.534</td>\n",
              "      <td>0.244</td>\n",
              "      <td>0.037</td>\n",
              "      <td>0.337</td>\n",
              "      <td>0.532</td>\n",
              "      <td>0.736</td>\n",
              "      <td>0.978</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4527.000</td>\n",
              "      <td>0.509</td>\n",
              "      <td>0.246</td>\n",
              "      <td>0.028</td>\n",
              "      <td>0.316</td>\n",
              "      <td>0.505</td>\n",
              "      <td>0.699</td>\n",
              "      <td>0.984</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2119.000</td>\n",
              "      <td>0.561</td>\n",
              "      <td>0.249</td>\n",
              "      <td>0.033</td>\n",
              "      <td>0.369</td>\n",
              "      <td>0.572</td>\n",
              "      <td>0.769</td>\n",
              "      <td>0.981</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                        audio_valence                                          \n",
              "                                count  mean   std   min   25%   50%   75%   max\n",
              "song_popularity_cluster                                                        \n",
              "0                            4098.000 0.527 0.257 0.023 0.325 0.532 0.740 0.982\n",
              "1                            2620.000 0.524 0.233 0.035 0.341 0.522 0.708 0.980\n",
              "2                            1449.000 0.534 0.244 0.037 0.337 0.532 0.736 0.978\n",
              "3                            4527.000 0.509 0.246 0.028 0.316 0.505 0.699 0.984\n",
              "4                            2119.000 0.561 0.249 0.033 0.369 0.572 0.769 0.981"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 903
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cYXpnkk7AYlh",
        "colab_type": "text"
      },
      "source": [
        "Valência de áudio não parece ser algo muito discrepante para podermos fazer uma conclusão."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1hVTMd1cCo20",
        "colab_type": "text"
      },
      "source": [
        "**Nota musical:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P9CmW4PxAzH4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "outputId": "21b4d768-7c7c-4499-bd3f-4b0f814949c6"
      },
      "source": [
        "plt.figure(figsize=(16,8))\n",
        "sns.countplot(x='key', data=df, hue='song_popularity_cluster')"
      ],
      "execution_count": 908,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3f17e62630>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 908
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lZ9F28L_CFCo",
        "colab_type": "text"
      },
      "source": [
        "Parece que a nota musical predominante não é um fator que afete a popularidade diretamente, vemos que o desenho das barras basicamente se repete ao longo das diferentes notas."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ishTM59C8Nn",
        "colab_type": "text"
      },
      "source": [
        "**Modo de áudio:**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z-i6tR11Cv_Z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 510
        },
        "outputId": "216c75c4-2010-4754-ee23-00e6c74f9573"
      },
      "source": [
        "plt.figure(figsize=(16,8))\n",
        "sns.countplot(x='audio_mode', data=df, hue='song_popularity_cluster')"
      ],
      "execution_count": 909,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f3f18e73d68>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 909
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HuNBWiK0C_ex",
        "colab_type": "text"
      },
      "source": [
        "Novamente parece que não há uma relação direta, entretanto é possível perceber que há mais músicas que preferem o modo Maior."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GSfjOiTkDaIU",
        "colab_type": "text"
      },
      "source": [
        "# Conclusão"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZFPCT0tqDc3p",
        "colab_type": "text"
      },
      "source": [
        "Com base com os dados que temos não é possível perceber uma relação muito bem definida sobre as nossas variáveis, parece que estes detalhes técnicos não importam muito para definir se uma música vai ser ou não popular."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ovrl8vjODvNU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "ee1f6d3f-2547-47a9-c0e7-5e14a7a8ecb0"
      },
      "source": [
        "df[df['song_popularity'] >= 95]"
      ],
      "execution_count": 911,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>song_popularity</th>\n",
              "      <th>song_duration_ms</th>\n",
              "      <th>acousticness</th>\n",
              "      <th>danceability</th>\n",
              "      <th>energy</th>\n",
              "      <th>key</th>\n",
              "      <th>liveness</th>\n",
              "      <th>loudness</th>\n",
              "      <th>audio_mode</th>\n",
              "      <th>speechiness</th>\n",
              "      <th>tempo</th>\n",
              "      <th>audio_valence</th>\n",
              "      <th>song_popularity_cluster</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>song_name</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>BEBE</th>\n",
              "      <td>95</td>\n",
              "      <td>217822.000</td>\n",
              "      <td>0.172</td>\n",
              "      <td>0.635</td>\n",
              "      <td>0.902</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.125</td>\n",
              "      <td>-3.559</td>\n",
              "      <td>1</td>\n",
              "      <td>0.109</td>\n",
              "      <td>102.063</td>\n",
              "      <td>0.286</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Be Alright</th>\n",
              "      <td>96</td>\n",
              "      <td>196373.000</td>\n",
              "      <td>0.697</td>\n",
              "      <td>0.553</td>\n",
              "      <td>0.586</td>\n",
              "      <td>11.000</td>\n",
              "      <td>0.081</td>\n",
              "      <td>-6.319</td>\n",
              "      <td>1</td>\n",
              "      <td>0.036</td>\n",
              "      <td>126.684</td>\n",
              "      <td>0.443</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Better</th>\n",
              "      <td>95</td>\n",
              "      <td>229412.000</td>\n",
              "      <td>0.098</td>\n",
              "      <td>0.442</td>\n",
              "      <td>0.585</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.140</td>\n",
              "      <td>-10.332</td>\n",
              "      <td>0</td>\n",
              "      <td>0.096</td>\n",
              "      <td>97.565</td>\n",
              "      <td>0.116</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Better Now</th>\n",
              "      <td>95</td>\n",
              "      <td>231266.000</td>\n",
              "      <td>0.354</td>\n",
              "      <td>0.680</td>\n",
              "      <td>0.563</td>\n",
              "      <td>10.000</td>\n",
              "      <td>0.136</td>\n",
              "      <td>-5.843</td>\n",
              "      <td>1</td>\n",
              "      <td>0.045</td>\n",
              "      <td>145.028</td>\n",
              "      <td>0.374</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Eastside (with Halsey &amp; Khalid)</th>\n",
              "      <td>98</td>\n",
              "      <td>173799.000</td>\n",
              "      <td>0.555</td>\n",
              "      <td>0.560</td>\n",
              "      <td>0.680</td>\n",
              "      <td>6.000</td>\n",
              "      <td>0.116</td>\n",
              "      <td>-7.648</td>\n",
              "      <td>0</td>\n",
              "      <td>0.321</td>\n",
              "      <td>89.391</td>\n",
              "      <td>0.319</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>FEFE (feat. Nicki Minaj &amp; Murda Beatz)</th>\n",
              "      <td>96</td>\n",
              "      <td>179404.000</td>\n",
              "      <td>0.088</td>\n",
              "      <td>0.931</td>\n",
              "      <td>0.387</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.136</td>\n",
              "      <td>-9.127</td>\n",
              "      <td>1</td>\n",
              "      <td>0.031</td>\n",
              "      <td>125.978</td>\n",
              "      <td>0.376</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Falling Down</th>\n",
              "      <td>97</td>\n",
              "      <td>196400.000</td>\n",
              "      <td>0.018</td>\n",
              "      <td>0.669</td>\n",
              "      <td>0.574</td>\n",
              "      <td>4.000</td>\n",
              "      <td>0.146</td>\n",
              "      <td>-6.442</td>\n",
              "      <td>0</td>\n",
              "      <td>0.029</td>\n",
              "      <td>120.013</td>\n",
              "      <td>0.273</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Happier</th>\n",
              "      <td>100</td>\n",
              "      <td>214289.000</td>\n",
              "      <td>0.191</td>\n",
              "      <td>0.687</td>\n",
              "      <td>0.792</td>\n",
              "      <td>5.000</td>\n",
              "      <td>0.167</td>\n",
              "      <td>-2.749</td>\n",
              "      <td>1</td>\n",
              "      <td>0.045</td>\n",
              "      <td>100.015</td>\n",
              "      <td>0.671</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Like It</th>\n",
              "      <td>95</td>\n",
              "      <td>253390.000</td>\n",
              "      <td>0.099</td>\n",
              "      <td>0.816</td>\n",
              "      <td>0.726</td>\n",
              "      <td>5.000</td>\n",
              "      <td>0.372</td>\n",
              "      <td>-3.998</td>\n",
              "      <td>0</td>\n",
              "      <td>0.129</td>\n",
              "      <td>136.048</td>\n",
              "      <td>0.650</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>I Love It (&amp; Lil Pump)</th>\n",
              "      <td>99</td>\n",
              "      <td>127946.000</td>\n",
              "      <td>0.011</td>\n",
              "      <td>0.901</td>\n",
              "      <td>0.522</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.259</td>\n",
              "      <td>-8.304</td>\n",
              "      <td>1</td>\n",
              "      <td>0.330</td>\n",
              "      <td>104.053</td>\n",
              "      <td>0.329</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>In My Feelings</th>\n",
              "      <td>98</td>\n",
              "      <td>217925.000</td>\n",
              "      <td>0.059</td>\n",
              "      <td>0.835</td>\n",
              "      <td>0.626</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.396</td>\n",
              "      <td>-5.833</td>\n",
              "      <td>1</td>\n",
              "      <td>0.125</td>\n",
              "      <td>91.030</td>\n",
              "      <td>0.350</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>In My Mind</th>\n",
              "      <td>97</td>\n",
              "      <td>184560.000</td>\n",
              "      <td>0.176</td>\n",
              "      <td>0.694</td>\n",
              "      <td>0.770</td>\n",
              "      <td>6.000</td>\n",
              "      <td>0.118</td>\n",
              "      <td>-5.335</td>\n",
              "      <td>1</td>\n",
              "      <td>0.149</td>\n",
              "      <td>125.905</td>\n",
              "      <td>0.163</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Lucid Dreams</th>\n",
              "      <td>97</td>\n",
              "      <td>239835.000</td>\n",
              "      <td>0.349</td>\n",
              "      <td>0.511</td>\n",
              "      <td>0.566</td>\n",
              "      <td>6.000</td>\n",
              "      <td>0.340</td>\n",
              "      <td>-7.230</td>\n",
              "      <td>0</td>\n",
              "      <td>0.200</td>\n",
              "      <td>83.903</td>\n",
              "      <td>0.218</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Lucky You (feat. Joyner Lucas)</th>\n",
              "      <td>96</td>\n",
              "      <td>244679.000</td>\n",
              "      <td>0.063</td>\n",
              "      <td>0.876</td>\n",
              "      <td>0.786</td>\n",
              "      <td>10.000</td>\n",
              "      <td>0.127</td>\n",
              "      <td>-4.884</td>\n",
              "      <td>0</td>\n",
              "      <td>0.306</td>\n",
              "      <td>153.068</td>\n",
              "      <td>0.575</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Moonlight</th>\n",
              "      <td>96</td>\n",
              "      <td>135090.000</td>\n",
              "      <td>0.556</td>\n",
              "      <td>0.921</td>\n",
              "      <td>0.537</td>\n",
              "      <td>9.000</td>\n",
              "      <td>0.102</td>\n",
              "      <td>-5.723</td>\n",
              "      <td>0</td>\n",
              "      <td>0.080</td>\n",
              "      <td>128.009</td>\n",
              "      <td>0.711</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Natural</th>\n",
              "      <td>95</td>\n",
              "      <td>189466.000</td>\n",
              "      <td>0.216</td>\n",
              "      <td>0.448</td>\n",
              "      <td>0.612</td>\n",
              "      <td>2.000</td>\n",
              "      <td>0.080</td>\n",
              "      <td>-6.106</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071</td>\n",
              "      <td>97.773</td>\n",
              "      <td>0.213</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Promises (with Sam Smith)</th>\n",
              "      <td>98</td>\n",
              "      <td>213309.000</td>\n",
              "      <td>0.012</td>\n",
              "      <td>0.781</td>\n",
              "      <td>0.768</td>\n",
              "      <td>11.000</td>\n",
              "      <td>0.325</td>\n",
              "      <td>-5.991</td>\n",
              "      <td>1</td>\n",
              "      <td>0.039</td>\n",
              "      <td>123.070</td>\n",
              "      <td>0.486</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SAD!</th>\n",
              "      <td>96</td>\n",
              "      <td>166605.000</td>\n",
              "      <td>0.258</td>\n",
              "      <td>0.740</td>\n",
              "      <td>0.613</td>\n",
              "      <td>8.000</td>\n",
              "      <td>0.123</td>\n",
              "      <td>-4.880</td>\n",
              "      <td>1</td>\n",
              "      <td>0.145</td>\n",
              "      <td>75.023</td>\n",
              "      <td>0.473</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SICKO MODE</th>\n",
              "      <td>97</td>\n",
              "      <td>312820.000</td>\n",
              "      <td>0.005</td>\n",
              "      <td>0.834</td>\n",
              "      <td>0.730</td>\n",
              "      <td>8.000</td>\n",
              "      <td>0.124</td>\n",
              "      <td>-3.714</td>\n",
              "      <td>1</td>\n",
              "      <td>0.222</td>\n",
              "      <td>155.008</td>\n",
              "      <td>0.446</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Taki Taki (with Selena Gomez, Ozuna &amp; Cardi B)</th>\n",
              "      <td>98</td>\n",
              "      <td>212500.000</td>\n",
              "      <td>0.153</td>\n",
              "      <td>0.841</td>\n",
              "      <td>0.798</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.062</td>\n",
              "      <td>-4.206</td>\n",
              "      <td>0</td>\n",
              "      <td>0.229</td>\n",
              "      <td>95.948</td>\n",
              "      <td>0.591</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Taste (feat. Offset)</th>\n",
              "      <td>95</td>\n",
              "      <td>232959.000</td>\n",
              "      <td>0.024</td>\n",
              "      <td>0.884</td>\n",
              "      <td>0.559</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.101</td>\n",
              "      <td>-7.442</td>\n",
              "      <td>1</td>\n",
              "      <td>0.120</td>\n",
              "      <td>97.994</td>\n",
              "      <td>0.342</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                song_popularity  ...  song_popularity_cluster\n",
              "song_name                                                        ...                         \n",
              "BEBE                                                         95  ...                        1\n",
              "Be Alright                                                   96  ...                        1\n",
              "Better                                                       95  ...                        1\n",
              "Better Now                                                   95  ...                        1\n",
              "Eastside (with Halsey & Khalid)                              98  ...                        1\n",
              "FEFE (feat. Nicki Minaj & Murda Beatz)                       96  ...                        1\n",
              "Falling Down                                                 97  ...                        1\n",
              "Happier                                                     100  ...                        1\n",
              "I Like It                                                    95  ...                        1\n",
              "I Love It (& Lil Pump)                                       99  ...                        1\n",
              "In My Feelings                                               98  ...                        1\n",
              "In My Mind                                                   97  ...                        1\n",
              "Lucid Dreams                                                 97  ...                        1\n",
              "Lucky You (feat. Joyner Lucas)                               96  ...                        1\n",
              "Moonlight                                                    96  ...                        1\n",
              "Natural                                                      95  ...                        1\n",
              "Promises (with Sam Smith)                                    98  ...                        1\n",
              "SAD!                                                         96  ...                        1\n",
              "SICKO MODE                                                   97  ...                        1\n",
              "Taki Taki (with Selena Gomez, Ozuna & Cardi B)               98  ...                        1\n",
              "Taste (feat. Offset)                                         95  ...                        1\n",
              "\n",
              "[21 rows x 13 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 911
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ymVBGrESELrv",
        "colab_type": "text"
      },
      "source": [
        "Essa parte do dataset mostra músicas com uma popularidade igual ou maior que 95, o que seriam excelentes neste quesito. Dando uma olhada podemos pensar em fatores que talvez sejam mais impactantes nessa análise, por exemplo: Estilo da música, cantor, pública alvo, idade média das pessoas que ouvem tal música, entre outros, mas aqui estamos entrando um pouco em especulação minha."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ph02WlSCE1F6",
        "colab_type": "text"
      },
      "source": [
        "Talvez não seja em si os detalhes técnicos, mas um misto disso com também fatores muito mais subjetivos, como gosto musical da maioria ou influências artísticas daquela da população."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f0UIxWrEFmmL",
        "colab_type": "text"
      },
      "source": [
        "# Relações entre variáveis"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q_tP27CaFpIz",
        "colab_type": "text"
      },
      "source": [
        "Mesmo que uma possível conclusão já tenha sido feita, não custa nada ver relações entre as nossas variáveis e ver se descobrimos alguma coisa que pode ser interessante."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SQuXTxR6FmJl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "outputId": "41fcbd73-27f4-47dd-df28-fd26ffe32ffd"
      },
      "source": [
        "corr = df.corr()\n",
        "sns.heatmap(corr, square=True, cmap=\"coolwarm\", center=0)\n",
        "plt.show()"
      ],
      "execution_count": 922,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ZQTIVVhJfG2",
        "colab_type": "text"
      },
      "source": [
        "Nossas variáveis não apresentam uma relação tão forte, porém podemos destacar a relação entre algumas delas:\n",
        "\n",
        "*   Energia e sonoridade\n",
        "*   Acústica e sonoridade\n",
        "*   Acústica e energia\n",
        "\n"
      ]
    }
  ]
}