Análise Rap Letras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 0. IMPORT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:06:57.903322Z",
     "start_time": "2021-01-28T20:06:55.625910Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import nltk\n",
    "import matplotlib.pyplot as plt\n",
    "from wordcloud import WordCloud, ImageColorGenerator\n",
    "import unidecode\n",
    "import re\n",
    "import string\n",
    "import matplotlib.ticker as mtick\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:06:58.457860Z",
     "start_time": "2021-01-28T20:06:58.453874Z"
    }
   },
   "outputs": [],
   "source": [
    "#Usefull link\n",
    "#https://github.com/carlosfab/sigmoidal_ai/blob/master/Como_Criar_uma_WordCloud_em_Python.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:06:59.361186Z",
     "start_time": "2021-01-28T20:06:59.200508Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>song</th>\n",
       "      <th>artist</th>\n",
       "      <th>lyrics</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Deixe o Som Reviver</td>\n",
       "      <td>3030</td>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Lovesong (part. Shan Luango)</td>\n",
       "      <td>Baco Exu do Blues</td>\n",
       "      <td>[Baco],Sou um urso psicopata,Pele preta cheia ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Mensagem Pra Eternidade (Versão Hip Hop Machine)</td>\n",
       "      <td>3030</td>\n",
       "      <td>[Bruno Chelles],A minha voz ecoa pela eternida...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Lágrima (part. Gloria Groove, Baco Exu do Blue...</td>\n",
       "      <td>Baco Exu do Blues</td>\n",
       "      <td>Eu só de te ver (caiu uma lágrima),Vou confess...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                song             artist  \\\n",
       "0                                Deixe o Som Reviver               3030   \n",
       "1                       Lovesong (part. Shan Luango)  Baco Exu do Blues   \n",
       "2   Mensagem Pra Eternidade (Versão Hip Hop Machine)               3030   \n",
       "3  Lágrima (part. Gloria Groove, Baco Exu do Blue...  Baco Exu do Blues   \n",
       "\n",
       "                                              lyrics  \n",
       "0  Pra ascender e envolver toda gente,Deixa o som...  \n",
       "1  [Baco],Sou um urso psicopata,Pele preta cheia ...  \n",
       "2  [Bruno Chelles],A minha voz ecoa pela eternida...  \n",
       "3  Eu só de te ver (caiu uma lágrima),Vou confess...  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_raw = pd.read_csv('rap.csv')\n",
    "df_raw.head(4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Data Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:02.389163Z",
     "start_time": "2021-01-28T20:07:02.315368Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            song   artist lyrics\n",
      "1997    Silêncio  Emicida    NaN\n",
      "5583  Lembranças     Sant    NaN\n"
     ]
    }
   ],
   "source": [
    "print(df_raw[df_raw['lyrics'].isna() == True])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:07.585081Z",
     "start_time": "2021-01-28T20:07:07.571393Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "song      0\n",
      "artist    0\n",
      "lyrics    0\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "df_valid = df_raw.dropna()\n",
    "print(df_valid.isna().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:10.021697Z",
     "start_time": "2021-01-28T20:07:10.015688Z"
    }
   },
   "outputs": [],
   "source": [
    "def cleaning_text(text):\n",
    "    \n",
    "    regex = re.compile('[%s]' % re.escape(string.punctuation.replace('\\'','')))\n",
    "    text = text.lower().replace('’', '')\n",
    "    text = unidecode.unidecode(text) \n",
    "    text = re.sub('\\[(.+?)\\]', '', text) #excluir palavras entre []\n",
    "    text = re.sub('\\((.+?)\\)', '', text) #excluir palavras entre ()\n",
    "    text = re.sub(regex, ' ', text) #excluir pontuação\n",
    "    text = re.sub('\\s+',' ', text).strip() #substituir espaços maiores que 1 por 1.\n",
    "    \n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:16.206350Z",
     "start_time": "2021-01-28T20:07:12.508558Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>artist</th>\n",
       "      <th>lyrics</th>\n",
       "      <th>lyrics_clean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1Kilo</td>\n",
       "      <td>[Chris],Ei!,Para de me olhar como se eu te dev...</td>\n",
       "      <td>ei para de me olhar como se eu te devesse algo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3030</td>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "      <td>pra ascender e envolver toda gente deixa o som...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ADL MC's</td>\n",
       "      <td>eu to juntando dinheiro, por isso,ando a pé me...</td>\n",
       "      <td>eu to juntando dinheiro por isso ando a pe mes...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     artist                                             lyrics  \\\n",
       "0     1Kilo  [Chris],Ei!,Para de me olhar como se eu te dev...   \n",
       "1      3030  Pra ascender e envolver toda gente,Deixa o som...   \n",
       "2  ADL MC's  eu to juntando dinheiro, por isso,ando a pé me...   \n",
       "\n",
       "                                        lyrics_clean  \n",
       "0  ei para de me olhar como se eu te devesse algo...  \n",
       "1  pra ascender e envolver toda gente deixa o som...  \n",
       "2  eu to juntando dinheiro por isso ando a pe mes...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#3.7seg\n",
    "df = df_valid.groupby('artist')['lyrics'].agg(''.join).reset_index()\n",
    "df['lyrics_clean'] = df['lyrics'].apply(lambda x: cleaning_text(x))\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Data Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1. Tokenizing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:24.226315Z",
     "start_time": "2021-01-28T20:07:24.048909Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lyrics</th>\n",
       "      <th>lyrics_clean</th>\n",
       "      <th>tokens</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>artist</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1Kilo</th>\n",
       "      <td>[Chris],Ei!,Para de me olhar como se eu te dev...</td>\n",
       "      <td>ei para de me olhar como se eu te devesse algo...</td>\n",
       "      <td>[ei, para, de, me, olhar, como, se, eu, te, de...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3030</th>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "      <td>pra ascender e envolver toda gente deixa o som...</td>\n",
       "      <td>[pra, ascender, e, envolver, toda, gente, deix...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADL MC's</th>\n",
       "      <td>eu to juntando dinheiro, por isso,ando a pé me...</td>\n",
       "      <td>eu to juntando dinheiro por isso ando a pe mes...</td>\n",
       "      <td>[eu, to, juntando, dinheiro, por, isso, ando, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     lyrics  \\\n",
       "artist                                                        \n",
       "1Kilo     [Chris],Ei!,Para de me olhar como se eu te dev...   \n",
       "3030      Pra ascender e envolver toda gente,Deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro, por isso,ando a pé me...   \n",
       "\n",
       "                                               lyrics_clean  \\\n",
       "artist                                                        \n",
       "1Kilo     ei para de me olhar como se eu te devesse algo...   \n",
       "3030      pra ascender e envolver toda gente deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro por isso ando a pe mes...   \n",
       "\n",
       "                                                     tokens  \n",
       "artist                                                       \n",
       "1Kilo     [ei, para, de, me, olhar, como, se, eu, te, de...  \n",
       "3030      [pra, ascender, e, envolver, toda, gente, deix...  \n",
       "ADL MC's  [eu, to, juntando, dinheiro, por, isso, ando, ...  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"tokens\"] = df.lyrics_clean.str.split()\n",
    "\n",
    "df.set_index('artist', inplace = True)\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2. Filtering Stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:07:32.639678Z",
     "start_time": "2021-01-28T20:07:32.487537Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\breno\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "nltk.download('stopwords')\n",
    "stopwords_nltk = nltk.corpus.stopwords.words('portuguese')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:08:07.753004Z",
     "start_time": "2021-01-28T20:08:04.264007Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lyrics</th>\n",
       "      <th>lyrics_clean</th>\n",
       "      <th>tokens</th>\n",
       "      <th>wordcloud</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>artist</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1Kilo</th>\n",
       "      <td>[Chris],Ei!,Para de me olhar como se eu te dev...</td>\n",
       "      <td>ei para de me olhar como se eu te devesse algo...</td>\n",
       "      <td>[ei, para, de, me, olhar, como, se, eu, te, de...</td>\n",
       "      <td>ei olhar devesse algo real tao devendo tipo ba...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3030</th>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "      <td>pra ascender e envolver toda gente deixa o som...</td>\n",
       "      <td>[pra, ascender, e, envolver, toda, gente, deix...</td>\n",
       "      <td>ascender envolver toda gente deixa som reviver...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADL MC's</th>\n",
       "      <td>eu to juntando dinheiro, por isso,ando a pé me...</td>\n",
       "      <td>eu to juntando dinheiro por isso ando a pe mes...</td>\n",
       "      <td>[eu, to, juntando, dinheiro, por, isso, ando, ...</td>\n",
       "      <td>juntando dinheiro ando pe ligo ando dia inteir...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     lyrics  \\\n",
       "artist                                                        \n",
       "1Kilo     [Chris],Ei!,Para de me olhar como se eu te dev...   \n",
       "3030      Pra ascender e envolver toda gente,Deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro, por isso,ando a pé me...   \n",
       "\n",
       "                                               lyrics_clean  \\\n",
       "artist                                                        \n",
       "1Kilo     ei para de me olhar como se eu te devesse algo...   \n",
       "3030      pra ascender e envolver toda gente deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro por isso ando a pe mes...   \n",
       "\n",
       "                                                     tokens  \\\n",
       "artist                                                        \n",
       "1Kilo     [ei, para, de, me, olhar, como, se, eu, te, de...   \n",
       "3030      [pra, ascender, e, envolver, toda, gente, deix...   \n",
       "ADL MC's  [eu, to, juntando, dinheiro, por, isso, ando, ...   \n",
       "\n",
       "                                                  wordcloud  \n",
       "artist                                                       \n",
       "1Kilo     ei olhar devesse algo real tao devendo tipo ba...  \n",
       "3030      ascender envolver toda gente deixa som reviver...  \n",
       "ADL MC's  juntando dinheiro ando pe ligo ando dia inteir...  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#3.37seg\n",
    "update = [\"tá\",'pra','tô', 'cê','pro', 'então', \"meu\", \"em\", \n",
    "                   \"você\", \"de\", \"ao\", \"os\",'vou', 'vai', 'vem', 'mim', \n",
    "                   'uns', 'sei', 'quero', 'ser', 'ver', 'aqui','faz']\n",
    "stopwords_raw = [*stopwords_nltk, *update]\n",
    "\n",
    "stopwords = [unidecode.unidecode(word) for word in stopwords_raw]\n",
    "\n",
    "def remove_stopwords(word_list):\n",
    "    return ' '.join([word for word in word_list if word not in stopwords])\n",
    "\n",
    "df['wordcloud'] = df['tokens'].apply(lambda x: remove_stopwords(x))\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.3. Counting Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:13:49.124441Z",
     "start_time": "2021-01-28T20:13:48.855068Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lyrics</th>\n",
       "      <th>lyrics_clean</th>\n",
       "      <th>tokens</th>\n",
       "      <th>wordcloud</th>\n",
       "      <th>tokens-not-stop</th>\n",
       "      <th>Unique</th>\n",
       "      <th>Total</th>\n",
       "      <th>Relative</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>artist</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>1Kilo</th>\n",
       "      <td>[Chris],Ei!,Para de me olhar como se eu te dev...</td>\n",
       "      <td>ei para de me olhar como se eu te devesse algo...</td>\n",
       "      <td>[ei, para, de, me, olhar, como, se, eu, te, de...</td>\n",
       "      <td>ei olhar devesse algo real tao devendo tipo ba...</td>\n",
       "      <td>[ei, olhar, devesse, algo, real, tao, devendo,...</td>\n",
       "      <td>8068</td>\n",
       "      <td>84948</td>\n",
       "      <td>0.094976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3030</th>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "      <td>pra ascender e envolver toda gente deixa o som...</td>\n",
       "      <td>[pra, ascender, e, envolver, toda, gente, deix...</td>\n",
       "      <td>ascender envolver toda gente deixa som reviver...</td>\n",
       "      <td>[ascender, envolver, toda, gente, deixa, som, ...</td>\n",
       "      <td>6938</td>\n",
       "      <td>57760</td>\n",
       "      <td>0.120118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADL MC's</th>\n",
       "      <td>eu to juntando dinheiro, por isso,ando a pé me...</td>\n",
       "      <td>eu to juntando dinheiro por isso ando a pe mes...</td>\n",
       "      <td>[eu, to, juntando, dinheiro, por, isso, ando, ...</td>\n",
       "      <td>juntando dinheiro ando pe ligo ando dia inteir...</td>\n",
       "      <td>[juntando, dinheiro, ando, pe, ligo, ando, dia...</td>\n",
       "      <td>6914</td>\n",
       "      <td>42941</td>\n",
       "      <td>0.161012</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     lyrics  \\\n",
       "artist                                                        \n",
       "1Kilo     [Chris],Ei!,Para de me olhar como se eu te dev...   \n",
       "3030      Pra ascender e envolver toda gente,Deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro, por isso,ando a pé me...   \n",
       "\n",
       "                                               lyrics_clean  \\\n",
       "artist                                                        \n",
       "1Kilo     ei para de me olhar como se eu te devesse algo...   \n",
       "3030      pra ascender e envolver toda gente deixa o som...   \n",
       "ADL MC's  eu to juntando dinheiro por isso ando a pe mes...   \n",
       "\n",
       "                                                     tokens  \\\n",
       "artist                                                        \n",
       "1Kilo     [ei, para, de, me, olhar, como, se, eu, te, de...   \n",
       "3030      [pra, ascender, e, envolver, toda, gente, deix...   \n",
       "ADL MC's  [eu, to, juntando, dinheiro, por, isso, ando, ...   \n",
       "\n",
       "                                                  wordcloud  \\\n",
       "artist                                                        \n",
       "1Kilo     ei olhar devesse algo real tao devendo tipo ba...   \n",
       "3030      ascender envolver toda gente deixa som reviver...   \n",
       "ADL MC's  juntando dinheiro ando pe ligo ando dia inteir...   \n",
       "\n",
       "                                            tokens-not-stop  Unique  Total  \\\n",
       "artist                                                                       \n",
       "1Kilo     [ei, olhar, devesse, algo, real, tao, devendo,...    8068  84948   \n",
       "3030      [ascender, envolver, toda, gente, deixa, som, ...    6938  57760   \n",
       "ADL MC's  [juntando, dinheiro, ando, pe, ligo, ando, dia...    6914  42941   \n",
       "\n",
       "          Relative  \n",
       "artist              \n",
       "1Kilo     0.094976  \n",
       "3030      0.120118  \n",
       "ADL MC's  0.161012  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Unique'] = df['tokens'].apply(lambda x: len(set(x)))\n",
    "df['Total'] = df['tokens'].apply(lambda x: len(x))\n",
    "df['Relative'] = df['tokens'].apply(lambda x: len(set(x))/len(x))\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Plotting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1. Wordcloud"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-06T21:30:20.029863Z",
     "start_time": "2021-01-06T21:24:17.129375Z"
    }
   },
   "outputs": [],
   "source": [
    "#6:22min pra rodar \n",
    "path = os.getcwd()\n",
    "wordcloud_folder = 'wordcloud'\n",
    "\n",
    "for artist in df.index:\n",
    "    wordcloud = WordCloud(background_color=\"black\",\n",
    "                          max_words = 500,\n",
    "                          width=1600, height=800).generate(df.loc[artist]['wordcloud'])\n",
    "    \n",
    "    # mostrar a imagem final\n",
    "#     fig, ax = plt.subplots(figsize=(10,6))\n",
    "#     ax.imshow(wordcloud, interpolation='bilinear')\n",
    "#     ax.set_axis_off()\n",
    "#     plt.imshow(wordcloud);\n",
    "\n",
    "    if not os.path.isdir(wordcloud_folder):\n",
    "        os.mkdir(wordcloud_folder)\n",
    "    \n",
    "    else:\n",
    "        wordcloud.to_file(os.path.join(path,wordcloud_folder,f\"{artist}_wordcloud.png\"))     "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2. Scatterplot of Unique Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:10:43.871178Z",
     "start_time": "2021-01-28T20:10:43.766368Z"
    },
    "hide_input": true
   },
   "outputs": [],
   "source": [
    "from __future__ import division\n",
    "import sys\n",
    "from matplotlib import pyplot as plt\n",
    "from itertools import product\n",
    "import numpy as np\n",
    "from operator import itemgetter\n",
    "from matplotlib.path import get_path_collection_extents\n",
    "import matplotlib\n",
    "\n",
    "\n",
    "def get_bboxes_pathcollection(sc, ax):\n",
    "    \"\"\"Function to return a list of bounding boxes in display coordinates\n",
    "    for a scatter plot\n",
    "    Thank you to ImportanceOfBeingErnest\n",
    "    https://stackoverflow.com/a/55007838/1304161\"\"\"\n",
    "    #    ax.figure.canvas.draw() # need to draw before the transforms are set.\n",
    "    transform = sc.get_transform()\n",
    "    transOffset = sc.get_offset_transform()\n",
    "    offsets = sc._offsets\n",
    "    paths = sc.get_paths()\n",
    "    transforms = sc.get_transforms()\n",
    "\n",
    "    if not transform.is_affine:\n",
    "        paths = [transform.transform_path_non_affine(p) for p in paths]\n",
    "        transform = transform.get_affine()\n",
    "    if not transOffset.is_affine:\n",
    "        offsets = transOffset.transform_non_affine(offsets)\n",
    "        transOffset = transOffset.get_affine()\n",
    "\n",
    "    if isinstance(offsets, np.ma.MaskedArray):\n",
    "        offsets = offsets.filled(np.nan)\n",
    "\n",
    "    bboxes = []\n",
    "\n",
    "    if len(paths) and len(offsets):\n",
    "        if len(paths) < len(offsets):\n",
    "            # for usual scatters you have one path, but several offsets\n",
    "            paths = [paths[0]] * len(offsets)\n",
    "        if len(transforms) < len(offsets):\n",
    "            # often you may have a single scatter size, but several offsets\n",
    "            transforms = [transforms[0]] * len(offsets)\n",
    "\n",
    "        for p, o, t in zip(paths, offsets, transforms):\n",
    "            result = get_path_collection_extents(\n",
    "                transform.frozen(), [p], [t], [o], transOffset.frozen()\n",
    "            )\n",
    "            bboxes.append(result)\n",
    "\n",
    "    return bboxes\n",
    "\n",
    "\n",
    "def get_text_position(text):\n",
    "    x, y = text.get_position()\n",
    "    t_x, t_y = text.get_transform().transform((x, y))\n",
    "    return (t_x, t_y)\n",
    "\n",
    "\n",
    "def set_text_position(text, t_x, t_y):\n",
    "    x, y = text.get_transform().inverted().transform((t_x, t_y))\n",
    "    text.set_position((x, y))\n",
    "    return None\n",
    "\n",
    "\n",
    "def get_orig_coords(transform, t_x, t_y):\n",
    "    x, y = transform.inverted().transform((t_x, t_y))\n",
    "    return (x, y)\n",
    "\n",
    "\n",
    "def get_bboxes(objs, r=None, expand=(1, 1), ax=None, transform=None):\n",
    "    \"\"\"\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    objs : list, or PathCollection\n",
    "        List of objects to get bboxes from. Also works with mpl PathCollection.\n",
    "    r : renderer\n",
    "        Renderer. The default is None, then automatically deduced from ax.\n",
    "    expand : (float, float), optional\n",
    "        How much to expand bboxes in (x, y), in fractions. The default is (1, 1).\n",
    "    ax : Axes, optional\n",
    "        The default is None, then uses current axes.\n",
    "    transform : optional\n",
    "        Transform to apply to the objects, if they don't return they window extent.\n",
    "        The default is None, then applies the default ax transform.\n",
    "    Returns\n",
    "    -------\n",
    "    list\n",
    "        List of bboxes.\n",
    "    \"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "    r = r or get_renderer(ax.get_figure())\n",
    "    try:\n",
    "        return [i.get_window_extent(r).expanded(*expand) for i in objs]\n",
    "    except (AttributeError, TypeError):\n",
    "        try:\n",
    "            if all([isinstance(obj, matplotlib.transforms.BboxBase) for obj in objs]):\n",
    "                return objs\n",
    "            else:\n",
    "                raise ValueError(\"Something is wrong\")\n",
    "        except TypeError:\n",
    "            return get_bboxes_pathcollection(objs, ax)\n",
    "\n",
    "\n",
    "def get_midpoint(bbox):\n",
    "    cx = (bbox.x0 + bbox.x1) / 2\n",
    "    cy = (bbox.y0 + bbox.y1) / 2\n",
    "    return cx, cy\n",
    "\n",
    "\n",
    "def get_points_inside_bbox(x, y, bbox):\n",
    "    \"\"\"Return the indices of points inside the given bbox.\"\"\"\n",
    "    x1, y1, x2, y2 = bbox.xmin, bbox.ymin, bbox.xmax, bbox.ymax\n",
    "    x_in = np.logical_and(x > x1, x < x2)\n",
    "    y_in = np.logical_and(y > y1, y < y2)\n",
    "    return np.asarray(np.where(x_in & y_in)[0])\n",
    "\n",
    "\n",
    "def get_renderer(fig):\n",
    "    try:\n",
    "        return fig.canvas.get_renderer()\n",
    "    except AttributeError:\n",
    "        return fig.canvas.renderer\n",
    "\n",
    "\n",
    "def overlap_bbox_and_point(bbox, xp, yp):\n",
    "    \"\"\"Given a bbox that contains a given point, return the (x, y) displacement\n",
    "    necessary to make the bbox not overlap the point.\"\"\"\n",
    "    cx, cy = get_midpoint(bbox)\n",
    "\n",
    "    dir_x = np.sign(cx - xp)\n",
    "    dir_y = np.sign(cy - yp)\n",
    "\n",
    "    if dir_x == -1:\n",
    "        dx = xp - bbox.xmax\n",
    "    elif dir_x == 1:\n",
    "        dx = xp - bbox.xmin\n",
    "    else:\n",
    "        dx = 0\n",
    "\n",
    "    if dir_y == -1:\n",
    "        dy = yp - bbox.ymax\n",
    "    elif dir_y == 1:\n",
    "        dy = yp - bbox.ymin\n",
    "    else:\n",
    "        dy = 0\n",
    "    return dx, dy\n",
    "\n",
    "\n",
    "def move_texts(texts, delta_x, delta_y, bboxes=None, renderer=None, ax=None):\n",
    "    ax = ax or plt.gca()\n",
    "    if bboxes is None:\n",
    "        r = renderer or get_renderer(ax.get_figure())\n",
    "        bboxes = get_bboxes(texts, r, (1, 1), ax=ax)\n",
    "    ax_bbox = ax.patch.get_extents()\n",
    "    xmin = ax_bbox.xmin\n",
    "    xmax = ax_bbox.xmax\n",
    "    ymin = ax_bbox.ymin\n",
    "    ymax = ax_bbox.ymax\n",
    "    for i, (text, dx, dy) in enumerate(zip(texts, delta_x, delta_y)):\n",
    "        bbox = bboxes[i]\n",
    "        x1, y1, x2, y2 = bbox.xmin, bbox.ymin, bbox.xmax, bbox.ymax\n",
    "        if x1 + dx < xmin:\n",
    "            dx = 0\n",
    "        if x2 + dx > xmax:\n",
    "            dx = 0\n",
    "        if y1 + dy < ymin:\n",
    "            dy = 0\n",
    "        if y2 + dy > ymax:\n",
    "            dy = 0\n",
    "\n",
    "        x, y = get_text_position(text)\n",
    "        newx = x + dx\n",
    "        newy = y + dy\n",
    "        set_text_position(text, newx, newy)\n",
    "\n",
    "\n",
    "def optimally_align_text(\n",
    "    x,\n",
    "    y,\n",
    "    texts,\n",
    "    expand=(1.0, 1.0),\n",
    "    add_bboxes=[],\n",
    "    renderer=None,\n",
    "    ax=None,\n",
    "    direction=\"xy\",\n",
    "):\n",
    "    \"\"\"\n",
    "    For all text objects find alignment that causes the least overlap with\n",
    "    points and other texts and apply it\n",
    "    \"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "    r = renderer or get_renderer(ax.get_figure())\n",
    "    ax_bbox = ax.patch.get_extents()\n",
    "    xmin = ax_bbox.xmin\n",
    "    xmax = ax_bbox.xmax\n",
    "    ymin = ax_bbox.ymin\n",
    "    ymax = ax_bbox.ymax\n",
    "    bboxes = get_bboxes(texts, r, expand, ax=ax)\n",
    "    if \"x\" not in direction:\n",
    "        ha = [\"\"]\n",
    "    else:\n",
    "        ha = [\"left\", \"right\", \"center\"]\n",
    "    if \"y\" not in direction:\n",
    "        va = [\"\"]\n",
    "    else:\n",
    "        va = [\"bottom\", \"top\", \"center\"]\n",
    "    alignment = list(product(ha, va))\n",
    "    #    coords = np.array(zip(x, y))\n",
    "    for i, text in enumerate(texts):\n",
    "        #        tcoords = np.array(text.get_position()).T\n",
    "        #        nonself_coords = coords[~np.all(coords==tcoords, axis=1)]\n",
    "        #        nonself_x, nonself_y = np.split(nonself_coords, 2, axis=1)\n",
    "        counts = []\n",
    "        for h, v in alignment:\n",
    "            if h:\n",
    "                text.set_ha(h)\n",
    "            if v:\n",
    "                text.set_va(v)\n",
    "            bbox = text.get_window_extent(r).expanded(*expand)\n",
    "            c = len(get_points_inside_bbox(x, y, bbox))\n",
    "            intersections = [\n",
    "                bbox.intersection(bbox, bbox2) if i != j else None\n",
    "                for j, bbox2 in enumerate(bboxes + add_bboxes)\n",
    "            ]\n",
    "            intersections = sum(\n",
    "                [abs(b.width * b.height) if b is not None else 0 for b in intersections]\n",
    "            )\n",
    "            # Check for out-of-axes position\n",
    "            bbox = text.get_window_extent(r)\n",
    "            x1, y1, x2, y2 = bbox.xmin, bbox.ymin, bbox.xmax, bbox.ymax\n",
    "            if x1 < xmin or x2 > xmax or y1 < ymin or y2 > ymax:\n",
    "                axout = 1\n",
    "            else:\n",
    "                axout = 0\n",
    "            counts.append((axout, c, intersections))\n",
    "        # Most important: prefer alignments that keep the text inside the axes.\n",
    "        # If tied, take the alignments that minimize the number of x, y points\n",
    "        # contained inside the text.\n",
    "        # Break any remaining ties by minimizing the total area of intersections\n",
    "        # with all text bboxes and other objects to avoid.\n",
    "        a, value = min(enumerate(counts), key=itemgetter(1))\n",
    "        if \"x\" in direction:\n",
    "            text.set_ha(alignment[a][0])\n",
    "        if \"y\" in direction:\n",
    "            text.set_va(alignment[a][1])\n",
    "        bboxes[i] = text.get_window_extent(r).expanded(*expand)\n",
    "    return texts\n",
    "\n",
    "\n",
    "def repel_text(\n",
    "    texts, renderer=None, ax=None, expand=(1.2, 1.2), only_use_max_min=False, move=False\n",
    "):\n",
    "    \"\"\"\n",
    "    Repel texts from each other while expanding their bounding boxes by expand\n",
    "    (x, y), e.g. (1.2, 1.2) would multiply width and height by 1.2.\n",
    "    Requires a renderer to get the actual sizes of the text, and to that end\n",
    "    either one needs to be directly provided, or the axes have to be specified,\n",
    "    and the renderer is then got from the axes object.\n",
    "    \"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "    r = renderer or get_renderer(ax.get_figure())\n",
    "    bboxes = get_bboxes(texts, r, expand, ax=ax)\n",
    "    xmins = [bbox.xmin for bbox in bboxes]\n",
    "    xmaxs = [bbox.xmax for bbox in bboxes]\n",
    "    ymaxs = [bbox.ymax for bbox in bboxes]\n",
    "    ymins = [bbox.ymin for bbox in bboxes]\n",
    "\n",
    "    overlaps_x = np.zeros((len(bboxes), len(bboxes)))\n",
    "    overlaps_y = np.zeros_like(overlaps_x)\n",
    "    overlap_directions_x = np.zeros_like(overlaps_x)\n",
    "    overlap_directions_y = np.zeros_like(overlaps_y)\n",
    "    for i, bbox1 in enumerate(bboxes):\n",
    "        overlaps = get_points_inside_bbox(\n",
    "            xmins * 2 + xmaxs * 2, (ymins + ymaxs) * 2, bbox1\n",
    "        ) % len(bboxes)\n",
    "        overlaps = np.unique(overlaps)\n",
    "        for j in overlaps:\n",
    "            bbox2 = bboxes[j]\n",
    "            x, y = bbox1.intersection(bbox1, bbox2).size\n",
    "            overlaps_x[i, j] = x\n",
    "            overlaps_y[i, j] = y\n",
    "            direction = np.sign(bbox1.extents - bbox2.extents)[:2]\n",
    "            overlap_directions_x[i, j] = direction[0]\n",
    "            overlap_directions_y[i, j] = direction[1]\n",
    "\n",
    "    move_x = overlaps_x * overlap_directions_x\n",
    "    move_y = overlaps_y * overlap_directions_y\n",
    "\n",
    "    delta_x = move_x.sum(axis=1)\n",
    "    delta_y = move_y.sum(axis=1)\n",
    "\n",
    "    q = np.sum(overlaps_x), np.sum(overlaps_y)\n",
    "    if move:\n",
    "        move_texts(texts, delta_x, delta_y, bboxes, ax=ax)\n",
    "    return delta_x, delta_y, q\n",
    "\n",
    "\n",
    "def repel_text_from_bboxes(\n",
    "    add_bboxes,\n",
    "    texts,\n",
    "    renderer=None,\n",
    "    ax=None,\n",
    "    expand=(1.2, 1.2),\n",
    "    only_use_max_min=False,\n",
    "    move=False,\n",
    "):\n",
    "    \"\"\"\n",
    "    Repel texts from other objects' bboxes while expanding their (texts')\n",
    "    bounding boxes by expand (x, y), e.g. (1.2, 1.2) would multiply width and\n",
    "    height by 1.2.\n",
    "    Requires a renderer to get the actual sizes of the text, and to that end\n",
    "    either one needs to be directly provided, or the axes have to be specified,\n",
    "    and the renderer is then got from the axes object.\n",
    "    \"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "    r = renderer or get_renderer(ax.get_figure())\n",
    "\n",
    "    bboxes = get_bboxes(texts, r, expand, ax=ax)\n",
    "\n",
    "    overlaps_x = np.zeros((len(bboxes), len(add_bboxes)))\n",
    "    overlaps_y = np.zeros_like(overlaps_x)\n",
    "    overlap_directions_x = np.zeros_like(overlaps_x)\n",
    "    overlap_directions_y = np.zeros_like(overlaps_y)\n",
    "\n",
    "    for i, bbox1 in enumerate(bboxes):\n",
    "        for j, bbox2 in enumerate(add_bboxes):\n",
    "            try:\n",
    "                x, y = bbox1.intersection(bbox1, bbox2).size\n",
    "                direction = np.sign(bbox1.extents - bbox2.extents)[:2]\n",
    "                overlaps_x[i, j] = x\n",
    "                overlaps_y[i, j] = y\n",
    "                overlap_directions_x[i, j] = direction[0]\n",
    "                overlap_directions_y[i, j] = direction[1]\n",
    "            except AttributeError:\n",
    "                pass\n",
    "\n",
    "    move_x = overlaps_x * overlap_directions_x\n",
    "    move_y = overlaps_y * overlap_directions_y\n",
    "\n",
    "    delta_x = move_x.sum(axis=1)\n",
    "    delta_y = move_y.sum(axis=1)\n",
    "\n",
    "    q = np.sum(overlaps_x), np.sum(overlaps_y)\n",
    "    if move:\n",
    "        move_texts(texts, delta_x, delta_y, bboxes, ax=ax)\n",
    "    return delta_x, delta_y, q\n",
    "\n",
    "\n",
    "def repel_text_from_points(\n",
    "    x, y, texts, renderer=None, ax=None, expand=(1.2, 1.2), move=False\n",
    "):\n",
    "    \"\"\"\n",
    "    Repel texts from all points specified by x and y while expanding their\n",
    "    (texts'!) bounding boxes by expandby  (x, y), e.g. (1.2, 1.2)\n",
    "    would multiply both width and height by 1.2.\n",
    "    Requires a renderer to get the actual sizes of the text, and to that end\n",
    "    either one needs to be directly provided, or the axes have to be specified,\n",
    "    and the renderer is then got from the axes object.\n",
    "    \"\"\"\n",
    "    assert len(x) == len(y)\n",
    "    ax = ax or plt.gca()\n",
    "    r = renderer or get_renderer(ax.get_figure())\n",
    "    bboxes = get_bboxes(texts, r, expand, ax=ax)\n",
    "\n",
    "    # move_x[i,j] is the x displacement of the i'th text caused by the j'th point\n",
    "    move_x = np.zeros((len(bboxes), len(x)))\n",
    "    move_y = np.zeros((len(bboxes), len(x)))\n",
    "    for i, bbox in enumerate(bboxes):\n",
    "        xy_in = get_points_inside_bbox(x, y, bbox)\n",
    "        for j in xy_in:\n",
    "            xp, yp = x[j], y[j]\n",
    "            dx, dy = overlap_bbox_and_point(bbox, xp, yp)\n",
    "\n",
    "            move_x[i, j] = dx\n",
    "            move_y[i, j] = dy\n",
    "\n",
    "    delta_x = move_x.sum(axis=1)\n",
    "    delta_y = move_y.sum(axis=1)\n",
    "    q = np.sum(np.abs(move_x)), np.sum(np.abs(move_y))\n",
    "    if move:\n",
    "        move_texts(texts, delta_x, delta_y, bboxes, ax=ax)\n",
    "    return delta_x, delta_y, q\n",
    "\n",
    "\n",
    "def repel_text_from_axes(texts, ax=None, bboxes=None, renderer=None, expand=None):\n",
    "    ax = ax or plt.gca()\n",
    "    r = renderer or get_renderer(ax.get_figure())\n",
    "    if expand is None:\n",
    "        expand = (1, 1)\n",
    "    if bboxes is None:\n",
    "        bboxes = get_bboxes(texts, r, expand=expand, ax=ax)\n",
    "    ax_bbox = ax.patch.get_extents()\n",
    "    xmin = ax_bbox.xmin\n",
    "    xmax = ax_bbox.xmax\n",
    "    ymin = ax_bbox.ymin\n",
    "    ymax = ax_bbox.ymax\n",
    "    for i, bbox in enumerate(bboxes):\n",
    "        x1, y1, x2, y2 = bbox.xmin, bbox.ymin, bbox.xmax, bbox.ymax\n",
    "        dx, dy = 0, 0\n",
    "        if x1 < xmin:\n",
    "            dx = xmin - x1\n",
    "        if x2 > xmax:\n",
    "            dx = xmax - x2\n",
    "        if y1 < ymin:\n",
    "            dy = ymin - y1\n",
    "        if y2 > ymax:\n",
    "            dy = ymax - y2\n",
    "        if dx or dy:\n",
    "            x, y = get_text_position(texts[i])\n",
    "            newx, newy = x + dx, y + dy\n",
    "            set_text_position(texts[i], newx, newy)\n",
    "    return texts\n",
    "\n",
    "\n",
    "def float_to_tuple(a):\n",
    "    try:\n",
    "        a = float(a)\n",
    "        return (a, a)\n",
    "    except TypeError:\n",
    "        assert len(a) == 2\n",
    "        try:\n",
    "            b = float(a[0]), float(a[1])\n",
    "        except TypeError:\n",
    "            raise TypeError(\"Force values must be castable to floats\")\n",
    "        return b\n",
    "\n",
    "\n",
    "def adjust_text(\n",
    "    texts,\n",
    "    x=None,\n",
    "    y=None,\n",
    "    add_objects=None,\n",
    "    ax=None,\n",
    "    expand_text=(1.05, 1.2),\n",
    "    expand_points=(1.05, 1.2),\n",
    "    expand_objects=(1.05, 1.2),\n",
    "    expand_align=(1.05, 1.2),\n",
    "    autoalign=\"xy\",\n",
    "    va=\"center\",\n",
    "    ha=\"center\",\n",
    "    force_text=(0.1, 0.25),\n",
    "    force_points=(0.2, 0.5),\n",
    "    force_objects=(0.1, 0.25),\n",
    "    lim=500,\n",
    "    precision=0.01,\n",
    "    only_move={\"points\": \"xy\", \"text\": \"xy\", \"objects\": \"xy\"},\n",
    "    avoid_text=True,\n",
    "    avoid_points=True,\n",
    "    avoid_self=True,\n",
    "    save_steps=False,\n",
    "    save_prefix=\"\",\n",
    "    save_format=\"png\",\n",
    "    add_step_numbers=True,\n",
    "    *args,\n",
    "    **kwargs\n",
    "):\n",
    "    \"\"\"Iteratively adjusts the locations of texts.\n",
    "    Call adjust_text the very last, after all plotting (especially\n",
    "    anything that can change the axes limits) has been done. This is\n",
    "    because to move texts the function needs to use the dimensions of\n",
    "    the axes, and without knowing the final size of the plots the\n",
    "    results will be completely nonsensical, or suboptimal.\n",
    "    First moves all texts that are outside the axes limits\n",
    "    inside. Then in each iteration moves all texts away from each\n",
    "    other and from points. In the end hides texts and substitutes them\n",
    "    with annotations to link them to the respective points.\n",
    "    Parameters\n",
    "    ----------\n",
    "    texts : list\n",
    "        A list of :obj:`matplotlib.text.Text` objects to adjust.\n",
    "    Other Parameters\n",
    "    ----------------\n",
    "    x : array_like\n",
    "        x-coordinates of points to repel from; if not provided only uses text\n",
    "        coordinates.\n",
    "    y : array_like\n",
    "        y-coordinates of points to repel from; if not provided only uses text\n",
    "        coordinates\n",
    "    add_objects : list or PathCollection\n",
    "        a list of additional matplotlib objects to avoid; they must have a\n",
    "        `.get_window_extent()` method; alternatively, a PathCollection or a\n",
    "        list of Bbox objects.\n",
    "    ax : matplotlib axe, default is current axe (plt.gca())\n",
    "        axe object with the plot\n",
    "    expand_text : array_like, default (1.05, 1.2)\n",
    "        a tuple/list/... with 2 multipliers (x, y) by which to expand the\n",
    "        bounding box of texts when repelling them from each other.\n",
    "    expand_points : array_like, default (1.05, 1.2)\n",
    "        a tuple/list/... with 2 multipliers (x, y) by which to expand the\n",
    "        bounding box of texts when repelling them from points.\n",
    "    expand_objects : array_like, default (1.05, 1.2)\n",
    "        a tuple/list/... with 2 multipliers (x, y) by which to expand the\n",
    "        bounding box of texts when repelling them from other objects.\n",
    "    expand_align : array_like, default (1.05, 1.2)\n",
    "        a tuple/list/... with 2 multipliers (x, y) by which to expand the\n",
    "        bounding box of texts when autoaligning texts.\n",
    "    autoalign: str or boolean {'xy', 'x', 'y', True, False}, default 'xy'\n",
    "        Direction in wich the best alignement will be determined\n",
    "        - 'xy' or True, best alignment of all texts determined in all\n",
    "          directions automatically before running the iterative adjustment\n",
    "          (overriding va and ha),\n",
    "        - 'x', will only align horizontally,\n",
    "        - 'y', will only align vertically,\n",
    "        - False, do nothing (i.e. preserve va and ha)\n",
    "    va : str, default 'center'\n",
    "        vertical alignment of texts\n",
    "    ha : str, default 'center'\n",
    "        horizontal alignment of texts,\n",
    "    force_text : tuple, default (0.1, 0.25)\n",
    "        the repel force from texts is multiplied by this value\n",
    "    force_points : tuple, default (0.2, 0.5)\n",
    "        the repel force from points is multiplied by this value\n",
    "    force_objects : float, default (0.1, 0.25)\n",
    "        same as other forces, but for repelling additional objects\n",
    "    lim : int, default 500\n",
    "        limit of number of iterations\n",
    "    precision : float, default 0.01\n",
    "        iterate until the sum of all overlaps along both x and y are less than\n",
    "        this amount, as a fraction of the total widths and heights,\n",
    "        respectively. May need to increase for complicated situations.\n",
    "    only_move : dict, default {'points':'xy', 'text':'xy', 'objects':'xy'}\n",
    "        a dict to restrict movement of texts to only certain axes for certain\n",
    "        types of overlaps.\n",
    "        Valid keys are 'points', 'text', and 'objects'.\n",
    "        Valid values are '', 'x', 'y', and 'xy'.\n",
    "        For example, only_move={'points':'y', 'text':'xy', 'objects':'xy'}\n",
    "        forbids moving texts along the x axis due to overlaps with points.\n",
    "    avoid_text : bool, default True\n",
    "        whether to repel texts from each other.\n",
    "    avoid_points : bool, default True\n",
    "        whether to repel texts from points. Can be helpful to switch off in\n",
    "        extremely crowded plots.\n",
    "    avoid_self : bool, default True\n",
    "        whether to repel texts from its original positions.\n",
    "    save_steps : bool, default False\n",
    "        whether to save intermediate steps as images.\n",
    "    save_prefix : str, default ''\n",
    "        if `save_steps` is True, a path and/or prefix to the saved steps.\n",
    "    save_format : str, default 'png'\n",
    "        if `save_steps` is True, a format to save the steps into.\n",
    "    add_step_numbers : bool, default True\n",
    "        if `save_steps` is True, whether to add step numbers as titles to the\n",
    "        images of saving steps.\n",
    "    args and kwargs :\n",
    "        any arguments will be fed into obj:`ax.annotate` after all the\n",
    "        optimization is done just for plotting the connecting arrows if\n",
    "        required.\n",
    "    Return\n",
    "    ------\n",
    "    int\n",
    "        Number of iteration\n",
    "    \"\"\"\n",
    "    plt.draw()\n",
    "    ax = ax or plt.gca()\n",
    "    r = get_renderer(ax.get_figure())\n",
    "    transform = texts[0].get_transform()\n",
    "    if (x is not None) & (y is not None):\n",
    "        for ix, tupxy in enumerate(zip(x, y)):\n",
    "            t_x, t_y = transform.transform(tupxy)\n",
    "            x[ix] = t_x\n",
    "            y[ix] = t_y\n",
    "    orig_xy = [get_text_position(text) for text in texts]\n",
    "    orig_x = [xy[0] for xy in orig_xy]\n",
    "    orig_y = [xy[1] for xy in orig_xy]\n",
    "    force_objects = float_to_tuple(force_objects)\n",
    "    force_text = float_to_tuple(force_text)\n",
    "    force_points = float_to_tuple(force_points)\n",
    "\n",
    "    #    xdiff = np.diff(ax.get_xlim())[0]\n",
    "    #    ydiff = np.diff(ax.get_ylim())[0]\n",
    "\n",
    "    bboxes = get_bboxes(texts, r, (1.0, 1.0), ax)\n",
    "    sum_width = np.sum(list(map(lambda bbox: bbox.width, bboxes)))\n",
    "    sum_height = np.sum(list(map(lambda bbox: bbox.height, bboxes)))\n",
    "    if not any(list(map(lambda val: \"x\" in val, only_move.values()))):\n",
    "        precision_x = np.inf\n",
    "    else:\n",
    "        precision_x = precision * sum_width\n",
    "    #\n",
    "    if not any(list(map(lambda val: \"y\" in val, only_move.values()))):\n",
    "        precision_y = np.inf\n",
    "    else:\n",
    "        precision_y = precision * sum_height\n",
    "\n",
    "    if x is None:\n",
    "        if y is None:\n",
    "            if avoid_self:\n",
    "                x, y = orig_x, orig_y\n",
    "            else:\n",
    "                x, y = [], []\n",
    "        else:\n",
    "            raise ValueError(\"Please specify both x and y, or neither\")\n",
    "    if y is None:\n",
    "        raise ValueError(\"Please specify both x and y, or neither\")\n",
    "    if add_objects is None:\n",
    "        text_from_objects = False\n",
    "        add_bboxes = []\n",
    "    else:\n",
    "        try:\n",
    "            add_bboxes = get_bboxes(add_objects, r, (1, 1), ax)\n",
    "        except:\n",
    "            raise ValueError(\n",
    "                \"Can't get bounding boxes from add_objects - is'\\\n",
    "                             it a flat list of matplotlib objects?\"\n",
    "            )\n",
    "            return\n",
    "        text_from_objects = True\n",
    "    for text in texts:\n",
    "        text.set_va(va)\n",
    "        text.set_ha(ha)\n",
    "    if save_steps:\n",
    "        if add_step_numbers:\n",
    "            plt.title(\"Before\")\n",
    "        plt.savefig(\n",
    "            \"%s%s.%s\" % (save_prefix, \"000a\", save_format), format=save_format, dpi=150\n",
    "        )\n",
    "\n",
    "    if autoalign:\n",
    "        if autoalign is True:\n",
    "            autoalign = \"xy\"\n",
    "        for i in range(2):\n",
    "            texts = optimally_align_text(\n",
    "                x,\n",
    "                y,\n",
    "                texts,\n",
    "                expand=expand_align,\n",
    "                add_bboxes=add_bboxes,\n",
    "                direction=autoalign,\n",
    "                renderer=r,\n",
    "                ax=ax,\n",
    "            )\n",
    "\n",
    "    if save_steps:\n",
    "        if add_step_numbers:\n",
    "            plt.title(\"Autoaligned\")\n",
    "        plt.savefig(\n",
    "            \"%s%s.%s\" % (save_prefix, \"000b\", save_format), format=save_format, dpi=150\n",
    "        )\n",
    "\n",
    "    texts = repel_text_from_axes(texts, ax, renderer=r, expand=expand_points)\n",
    "    history = [(np.inf, np.inf)] * 10\n",
    "    for i in range(lim):\n",
    "        #        q1, q2 = [np.inf, np.inf], [np.inf, np.inf]\n",
    "\n",
    "        if avoid_text:\n",
    "            d_x_text, d_y_text, q1 = repel_text(\n",
    "                texts, renderer=r, ax=ax, expand=expand_text\n",
    "            )\n",
    "        else:\n",
    "            d_x_text, d_y_text, q1 = [0] * len(texts), [0] * len(texts), (0, 0)\n",
    "\n",
    "        if avoid_points:\n",
    "            d_x_points, d_y_points, q2 = repel_text_from_points(\n",
    "                x, y, texts, ax=ax, renderer=r, expand=expand_points\n",
    "            )\n",
    "        else:\n",
    "            d_x_points, d_y_points, q2 = [0] * len(texts), [0] * len(texts), (0, 0)\n",
    "\n",
    "        if text_from_objects:\n",
    "            d_x_objects, d_y_objects, q3 = repel_text_from_bboxes(\n",
    "                add_bboxes, texts, ax=ax, renderer=r, expand=expand_objects\n",
    "            )\n",
    "        else:\n",
    "            d_x_objects, d_y_objects, q3 = [0] * len(texts), [0] * len(texts), (0, 0)\n",
    "\n",
    "        if only_move:\n",
    "            if \"text\" in only_move:\n",
    "                if \"x\" not in only_move[\"text\"]:\n",
    "                    d_x_text = np.zeros_like(d_x_text)\n",
    "                if \"y\" not in only_move[\"text\"]:\n",
    "                    d_y_text = np.zeros_like(d_y_text)\n",
    "            if \"points\" in only_move:\n",
    "                if \"x\" not in only_move[\"points\"]:\n",
    "                    d_x_points = np.zeros_like(d_x_points)\n",
    "                if \"y\" not in only_move[\"points\"]:\n",
    "                    d_y_points = np.zeros_like(d_y_points)\n",
    "            if \"objects\" in only_move:\n",
    "                if \"x\" not in only_move[\"objects\"]:\n",
    "                    d_x_objects = np.zeros_like(d_x_objects)\n",
    "                if \"y\" not in only_move[\"objects\"]:\n",
    "                    d_y_objects = np.zeros_like(d_y_objects)\n",
    "\n",
    "        dx = (\n",
    "            np.array(d_x_text) * force_text[0]\n",
    "            + np.array(d_x_points) * force_points[0]\n",
    "            + np.array(d_x_objects) * force_objects[0]\n",
    "        )\n",
    "        dy = (\n",
    "            np.array(d_y_text) * force_text[1]\n",
    "            + np.array(d_y_points) * force_points[1]\n",
    "            + np.array(d_y_objects) * force_objects[1]\n",
    "        )\n",
    "        qx = np.sum([q[0] for q in [q1, q2, q3]])\n",
    "        qy = np.sum([q[1] for q in [q1, q2, q3]])\n",
    "        histm = np.max(np.array(history), axis=0)\n",
    "        history.pop(0)\n",
    "        history.append((qx, qy))\n",
    "        move_texts(texts, dx, dy, bboxes=get_bboxes(texts, r, (1, 1), ax), ax=ax)\n",
    "        if save_steps:\n",
    "            if add_step_numbers:\n",
    "                plt.title(i + 1)\n",
    "            plt.savefig(\n",
    "                \"%s%s.%s\" % (save_prefix, \"{0:03}\".format(i + 1), save_format),\n",
    "                format=save_format,\n",
    "                dpi=150,\n",
    "            )\n",
    "        # Stop if we've reached the precision threshold, or if the x and y displacement\n",
    "        # are both greater than the max over the last 10 iterations (suggesting a\n",
    "        # failure to converge)\n",
    "        if (qx < precision_x and qy < precision_y) or np.all([qx, qy] >= histm):\n",
    "            break\n",
    "        # Now adding arrows from texts to their original locations if required\n",
    "    if \"arrowprops\" in kwargs:\n",
    "        bboxes = get_bboxes(texts, r, (1, 1), ax)\n",
    "        kwap = kwargs.pop(\"arrowprops\")\n",
    "        for j, (bbox, text) in enumerate(zip(bboxes, texts)):\n",
    "            ap = {\"patchA\": text}  # Ensure arrow is clipped by the text\n",
    "            ap.update(kwap)  # Add arrowprops from kwargs\n",
    "            ax.annotate(\n",
    "                \"\",  # Add an arrow from the text to the point\n",
    "                xy=get_orig_coords(transform, orig_x[j], orig_y[j]),\n",
    "                xytext=transform.inverted().transform(get_midpoint(bbox)),\n",
    "                arrowprops=ap,\n",
    "                xycoords=transform,\n",
    "                textcoords=transform,\n",
    "                *args,\n",
    "                **kwargs\n",
    "            )\n",
    "\n",
    "    if save_steps:\n",
    "        if add_step_numbers:\n",
    "            plt.title(i + 1)\n",
    "            plt.savefig(\n",
    "                \"%s%s.%s\" % (save_prefix, \"{0:03}\".format(i + 1), save_format),\n",
    "                format=save_format,\n",
    "                dpi=150,\n",
    "            )\n",
    "\n",
    "\n",
    "            return i + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-28T20:15:09.942563Z",
     "start_time": "2021-01-28T20:13:55.614467Z"
    },
    "hide_input": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x1080 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# https://github.com/Phlya/adjustText/blob/master/adjustText/__init__.py\n",
    "# 1:07 min pra executar\n",
    "x = df['Unique']\n",
    "y = df['Relative']\n",
    "annot = df.index\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, 15))\n",
    "ax.scatter(x, y)\n",
    "\n",
    "texts = []\n",
    "for i, txt in enumerate(annot):\n",
    "    texts.append(ax.text(x[i], y[i], txt))\n",
    "adjust_text(texts)\n",
    "plt.xlabel(\"#Palavras Únicas\")\n",
    "plt.ylabel(\"Relação #Palavras Únicas/#Palavras Totais\")\n",
    "ax.yaxis.set_major_formatter(mtick.PercentFormatter(1))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-06T22:31:40.362236Z",
     "start_time": "2021-01-06T22:31:40.255820Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lyrics</th>\n",
       "      <th>lyrics_clean</th>\n",
       "      <th>tokens</th>\n",
       "      <th>wordcloud</th>\n",
       "      <th>Unique</th>\n",
       "      <th>Total</th>\n",
       "      <th>Relative</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>artist</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>1Kilo</th>\n",
       "      <td>[Chris],Ei!,Para de me olhar como se eu te dev...</td>\n",
       "      <td>ei para de me olhar como se eu te devesse algo...</td>\n",
       "      <td>[ei, para, de, me, olhar, como, se, eu, te, de...</td>\n",
       "      <td>ei olhar devesse algo real tao devendo tipo ba...</td>\n",
       "      <td>8068</td>\n",
       "      <td>84948</td>\n",
       "      <td>0.094976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3030</th>\n",
       "      <td>Pra ascender e envolver toda gente,Deixa o som...</td>\n",
       "      <td>pra ascender e envolver toda gente deixa o som...</td>\n",
       "      <td>[pra, ascender, e, envolver, toda, gente, deix...</td>\n",
       "      <td>ascender envolver toda gente deixa som reviver...</td>\n",
       "      <td>6938</td>\n",
       "      <td>57760</td>\n",
       "      <td>0.120118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADL MC's</th>\n",
       "      <td>eu to juntando dinheiro, por isso,ando a pé me...</td>\n",
       "      <td>eu to juntando dinheiro por isso ando a pe mes...</td>\n",
       "      <td>[eu, to, juntando, dinheiro, por, isso, ando, ...</td>\n",
       "      <td>juntando dinheiro ando pe ligo ando dia inteir...</td>\n",
       "      <td>6914</td>\n",
       "      <td>42941</td>\n",
       "      <td>0.161012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C4bal</th>\n",
       "      <td>Ok ok, é o cabal no remix,Junto com o chris br...</td>\n",
       "      <td>ok ok e o cabal no remix junto com o chris bro...</td>\n",
       "      <td>[ok, ok, e, o, cabal, no, remix, junto, com, o...</td>\n",
       "      <td>ok ok cabal remix junto chris brow olha oque f...</td>\n",
       "      <td>7376</td>\n",
       "      <td>67189</td>\n",
       "      <td>0.109780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cacife Clan</th>\n",
       "      <td>A noite comigo hoje é sua,Aproveite antes que ...</td>\n",
       "      <td>a noite comigo hoje e sua aproveite antes que ...</td>\n",
       "      <td>[a, noite, comigo, hoje, e, sua, aproveite, an...</td>\n",
       "      <td>noite comigo hoje aproveite antes sol caia lua...</td>\n",
       "      <td>5923</td>\n",
       "      <td>44167</td>\n",
       "      <td>0.134105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ConeCrewDiretoria</th>\n",
       "      <td>Você sabe como é (4x),Com o meu rap undergroun...</td>\n",
       "      <td>voce sabe como e com o meu rap underground se ...</td>\n",
       "      <td>[voce, sabe, como, e, com, o, meu, rap, underg...</td>\n",
       "      <td>sabe rap underground necessario faco uniao som...</td>\n",
       "      <td>7859</td>\n",
       "      <td>50583</td>\n",
       "      <td>0.155368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Costa Gold</th>\n",
       "      <td>Tanta miséria que vemos por aí,,Ensinam a valo...</td>\n",
       "      <td>tanta miseria que vemos por ai ensinam a valor...</td>\n",
       "      <td>[tanta, miseria, que, vemos, por, ai, ensinam,...</td>\n",
       "      <td>tanta miseria vemos ai ensinam valorizar tudo ...</td>\n",
       "      <td>9959</td>\n",
       "      <td>75137</td>\n",
       "      <td>0.132545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dalsin</th>\n",
       "      <td>Eu era fogo perdido ela solução da chama,Eu mo...</td>\n",
       "      <td>eu era fogo perdido ela solucao da chama eu mo...</td>\n",
       "      <td>[eu, era, fogo, perdido, ela, solucao, da, cha...</td>\n",
       "      <td>fogo perdido solucao chama mortal plano terra ...</td>\n",
       "      <td>6460</td>\n",
       "      <td>42404</td>\n",
       "      <td>0.152344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Diomedes Chinaski</th>\n",
       "      <td>[Buneco],Quantos irmãos que eu já vi padecer n...</td>\n",
       "      <td>quantos irmaos que eu ja vi padecer na corrida...</td>\n",
       "      <td>[quantos, irmaos, que, eu, ja, vi, padecer, na...</td>\n",
       "      <td>quantos irmaos vi padecer corrida maluca briga...</td>\n",
       "      <td>6794</td>\n",
       "      <td>43329</td>\n",
       "      <td>0.156800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Djonga</th>\n",
       "      <td>Aliado era os cara da zona Leste,De pivete ,,S...</td>\n",
       "      <td>aliado era os cara da zona leste de pivete so ...</td>\n",
       "      <td>[aliado, era, os, cara, da, zona, leste, de, p...</td>\n",
       "      <td>aliado cara zona leste pivete lembrar neguin p...</td>\n",
       "      <td>5497</td>\n",
       "      <td>35575</td>\n",
       "      <td>0.154519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Emicida</th>\n",
       "      <td>Quebram-se barreiras,Abrem-se as portas,Ultrap...</td>\n",
       "      <td>quebram se barreiras abrem se as portas ultrap...</td>\n",
       "      <td>[quebram, se, barreiras, abrem, se, as, portas...</td>\n",
       "      <td>quebram barreiras abrem portas ultrapassam fro...</td>\n",
       "      <td>10113</td>\n",
       "      <td>75413</td>\n",
       "      <td>0.134102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fabio Brazza</th>\n",
       "      <td>[Gabriel Elias],Eu vou te levar, pra ver a coi...</td>\n",
       "      <td>eu vou te levar pra ver a coisa mais bonita qu...</td>\n",
       "      <td>[eu, vou, te, levar, pra, ver, a, coisa, mais,...</td>\n",
       "      <td>levar coisa bonita vi confesso desde jamais es...</td>\n",
       "      <td>7869</td>\n",
       "      <td>42455</td>\n",
       "      <td>0.185349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Facção Central</th>\n",
       "      <td>Na minha infancia quis um kart,não tive nenhum...</td>\n",
       "      <td>na minha infancia quis um kart nao tive nenhum...</td>\n",
       "      <td>[na, minha, infancia, quis, um, kart, nao, tiv...</td>\n",
       "      <td>infancia quis kart nenhum boy vira piloto form...</td>\n",
       "      <td>10319</td>\n",
       "      <td>82796</td>\n",
       "      <td>0.124632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Froid</th>\n",
       "      <td>[Froid],Original rude boy,Original rude boy,Or...</td>\n",
       "      <td>original rude boy original rude boy original r...</td>\n",
       "      <td>[original, rude, boy, original, rude, boy, ori...</td>\n",
       "      <td>original rude boy original rude boy original r...</td>\n",
       "      <td>6485</td>\n",
       "      <td>39783</td>\n",
       "      <td>0.163009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gabriel O Pensador</th>\n",
       "      <td>Eu não sei se ele era antigo e tava congelado,...</td>\n",
       "      <td>eu nao sei se ele era antigo e tava congelado ...</td>\n",
       "      <td>[eu, nao, sei, se, ele, era, antigo, e, tava, ...</td>\n",
       "      <td>antigo tava congelado alguem morrido ressuscit...</td>\n",
       "      <td>7496</td>\n",
       "      <td>62607</td>\n",
       "      <td>0.119731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Haikaiss</th>\n",
       "      <td>[Miu],Você, é o meu sonho distante tão relevan...</td>\n",
       "      <td>voce e o meu sonho distante tao relevante que ...</td>\n",
       "      <td>[voce, e, o, meu, sonho, distante, tao, releva...</td>\n",
       "      <td>sonho distante tao relevante viver viver sonho...</td>\n",
       "      <td>10472</td>\n",
       "      <td>79710</td>\n",
       "      <td>0.131376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hungria Hip Hop</th>\n",
       "      <td>Aí, garçom,Aí, garçom,Aí, garçom,Aí, garçom,Aí...</td>\n",
       "      <td>ai garcom ai garcom ai garcom ai garcom ai gar...</td>\n",
       "      <td>[ai, garcom, ai, garcom, ai, garcom, ai, garco...</td>\n",
       "      <td>ai garcom ai garcom ai garcom ai garcom ai gar...</td>\n",
       "      <td>4520</td>\n",
       "      <td>36688</td>\n",
       "      <td>0.123201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MC Hariel</th>\n",
       "      <td>[MC PP da VS],Aí nós passa no meio do baile co...</td>\n",
       "      <td>ai nos passa no meio do baile com o dedo pro a...</td>\n",
       "      <td>[ai, nos, passa, no, meio, do, baile, com, o, ...</td>\n",
       "      <td>ai passa meio baile dedo alto copao 700 talo l...</td>\n",
       "      <td>5310</td>\n",
       "      <td>54138</td>\n",
       "      <td>0.098083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MV Bill</th>\n",
       "      <td>Quando a noite cai,A pista é dominada, o arreg...</td>\n",
       "      <td>quando a noite cai a pista e dominada o arrego...</td>\n",
       "      <td>[quando, a, noite, cai, a, pista, e, dominada,...</td>\n",
       "      <td>noite cai pista dominada arrego pago bonde sai...</td>\n",
       "      <td>7943</td>\n",
       "      <td>64158</td>\n",
       "      <td>0.123804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Marcelo D2</th>\n",
       "      <td>Rio, cidade-desespero,A vida é boa mas só vive...</td>\n",
       "      <td>rio cidade desespero a vida e boa mas so vive ...</td>\n",
       "      <td>[rio, cidade, desespero, a, vida, e, boa, mas,...</td>\n",
       "      <td>rio cidade desespero vida boa vive medo olho a...</td>\n",
       "      <td>6538</td>\n",
       "      <td>55622</td>\n",
       "      <td>0.117543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mr. Thug (Diego Thug)</th>\n",
       "      <td>Faltam palavras para demonstrar nosso corre en...</td>\n",
       "      <td>faltam palavras para demonstrar nosso corre en...</td>\n",
       "      <td>[faltam, palavras, para, demonstrar, nosso, co...</td>\n",
       "      <td>faltam palavras demonstrar corre rimamos toda ...</td>\n",
       "      <td>5279</td>\n",
       "      <td>48486</td>\n",
       "      <td>0.108877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Oriente</th>\n",
       "      <td>Tava ouvindo um Bob Marley e te olhando dormir...</td>\n",
       "      <td>tava ouvindo um bob marley e te olhando dormir...</td>\n",
       "      <td>[tava, ouvindo, um, bob, marley, e, te, olhand...</td>\n",
       "      <td>tava ouvindo bob marley olhando dormir lembran...</td>\n",
       "      <td>6224</td>\n",
       "      <td>35924</td>\n",
       "      <td>0.173255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Projota</th>\n",
       "      <td>Eu sangro, por mais que você não perceba,Mais ...</td>\n",
       "      <td>eu sangro por mais que voce nao perceba mais q...</td>\n",
       "      <td>[eu, sangro, por, mais, que, voce, nao, perceb...</td>\n",
       "      <td>sangro perceba receba amor forma posso dar ain...</td>\n",
       "      <td>8133</td>\n",
       "      <td>78622</td>\n",
       "      <td>0.103444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Racionais MC's</th>\n",
       "      <td>60% dos jovens de periferia sem antecedentes,C...</td>\n",
       "      <td>60 dos jovens de periferia sem antecedentes cr...</td>\n",
       "      <td>[60, dos, jovens, de, periferia, sem, antecede...</td>\n",
       "      <td>60 jovens periferia antecedentes criminais sof...</td>\n",
       "      <td>8052</td>\n",
       "      <td>64163</td>\n",
       "      <td>0.125493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Rashid</th>\n",
       "      <td>Cada manhã com você tem um aroma de maçã,E cad...</td>\n",
       "      <td>cada manha com voce tem um aroma de maca e cad...</td>\n",
       "      <td>[cada, manha, com, voce, tem, um, aroma, de, m...</td>\n",
       "      <td>cada manha aroma maca cada beijo completar cad...</td>\n",
       "      <td>8691</td>\n",
       "      <td>67057</td>\n",
       "      <td>0.129606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Sidoka</th>\n",
       "      <td>Sabe, eu posso ser bem mais que julga,Quem sab...</td>\n",
       "      <td>sabe eu posso ser bem mais que julga quem sabe...</td>\n",
       "      <td>[sabe, eu, posso, ser, bem, mais, que, julga, ...</td>\n",
       "      <td>sabe posso bem julga sabe dando mole sim sabe ...</td>\n",
       "      <td>5040</td>\n",
       "      <td>58357</td>\n",
       "      <td>0.086365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>UCLÃ</th>\n",
       "      <td>Eu gostaria que apertasse o haxixe, mas eu nem...</td>\n",
       "      <td>eu gostaria que apertasse o haxixe mas eu nem ...</td>\n",
       "      <td>[eu, gostaria, que, apertasse, o, haxixe, mas,...</td>\n",
       "      <td>gostaria apertasse haxixe fumo oi tesoura ai t...</td>\n",
       "      <td>5838</td>\n",
       "      <td>46838</td>\n",
       "      <td>0.124642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Xamã</th>\n",
       "      <td>Você me falou que eu curto foto de outras mina...</td>\n",
       "      <td>voce me falou que eu curto foto de outras mina...</td>\n",
       "      <td>[voce, me, falou, que, eu, curto, foto, de, ou...</td>\n",
       "      <td>falou curto foto outras mina diz vacilo conqui...</td>\n",
       "      <td>6513</td>\n",
       "      <td>38873</td>\n",
       "      <td>0.167546</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                  lyrics  \\\n",
       "artist                                                                     \n",
       "1Kilo                  [Chris],Ei!,Para de me olhar como se eu te dev...   \n",
       "3030                   Pra ascender e envolver toda gente,Deixa o som...   \n",
       "ADL MC's               eu to juntando dinheiro, por isso,ando a pé me...   \n",
       "C4bal                  Ok ok, é o cabal no remix,Junto com o chris br...   \n",
       "Cacife Clan            A noite comigo hoje é sua,Aproveite antes que ...   \n",
       "ConeCrewDiretoria      Você sabe como é (4x),Com o meu rap undergroun...   \n",
       "Costa Gold             Tanta miséria que vemos por aí,,Ensinam a valo...   \n",
       "Dalsin                 Eu era fogo perdido ela solução da chama,Eu mo...   \n",
       "Diomedes Chinaski      [Buneco],Quantos irmãos que eu já vi padecer n...   \n",
       "Djonga                 Aliado era os cara da zona Leste,De pivete ,,S...   \n",
       "Emicida                Quebram-se barreiras,Abrem-se as portas,Ultrap...   \n",
       "Fabio Brazza           [Gabriel Elias],Eu vou te levar, pra ver a coi...   \n",
       "Facção Central         Na minha infancia quis um kart,não tive nenhum...   \n",
       "Froid                  [Froid],Original rude boy,Original rude boy,Or...   \n",
       "Gabriel O Pensador     Eu não sei se ele era antigo e tava congelado,...   \n",
       "Haikaiss               [Miu],Você, é o meu sonho distante tão relevan...   \n",
       "Hungria Hip Hop        Aí, garçom,Aí, garçom,Aí, garçom,Aí, garçom,Aí...   \n",
       "MC Hariel              [MC PP da VS],Aí nós passa no meio do baile co...   \n",
       "MV Bill                Quando a noite cai,A pista é dominada, o arreg...   \n",
       "Marcelo D2             Rio, cidade-desespero,A vida é boa mas só vive...   \n",
       "Mr. Thug (Diego Thug)  Faltam palavras para demonstrar nosso corre en...   \n",
       "Oriente                Tava ouvindo um Bob Marley e te olhando dormir...   \n",
       "Projota                Eu sangro, por mais que você não perceba,Mais ...   \n",
       "Racionais MC's         60% dos jovens de periferia sem antecedentes,C...   \n",
       "Rashid                 Cada manhã com você tem um aroma de maçã,E cad...   \n",
       "Sidoka                 Sabe, eu posso ser bem mais que julga,Quem sab...   \n",
       "UCLÃ                   Eu gostaria que apertasse o haxixe, mas eu nem...   \n",
       "Xamã                   Você me falou que eu curto foto de outras mina...   \n",
       "\n",
       "                                                            lyrics_clean  \\\n",
       "artist                                                                     \n",
       "1Kilo                  ei para de me olhar como se eu te devesse algo...   \n",
       "3030                   pra ascender e envolver toda gente deixa o som...   \n",
       "ADL MC's               eu to juntando dinheiro por isso ando a pe mes...   \n",
       "C4bal                  ok ok e o cabal no remix junto com o chris bro...   \n",
       "Cacife Clan            a noite comigo hoje e sua aproveite antes que ...   \n",
       "ConeCrewDiretoria      voce sabe como e com o meu rap underground se ...   \n",
       "Costa Gold             tanta miseria que vemos por ai ensinam a valor...   \n",
       "Dalsin                 eu era fogo perdido ela solucao da chama eu mo...   \n",
       "Diomedes Chinaski      quantos irmaos que eu ja vi padecer na corrida...   \n",
       "Djonga                 aliado era os cara da zona leste de pivete so ...   \n",
       "Emicida                quebram se barreiras abrem se as portas ultrap...   \n",
       "Fabio Brazza           eu vou te levar pra ver a coisa mais bonita qu...   \n",
       "Facção Central         na minha infancia quis um kart nao tive nenhum...   \n",
       "Froid                  original rude boy original rude boy original r...   \n",
       "Gabriel O Pensador     eu nao sei se ele era antigo e tava congelado ...   \n",
       "Haikaiss               voce e o meu sonho distante tao relevante que ...   \n",
       "Hungria Hip Hop        ai garcom ai garcom ai garcom ai garcom ai gar...   \n",
       "MC Hariel              ai nos passa no meio do baile com o dedo pro a...   \n",
       "MV Bill                quando a noite cai a pista e dominada o arrego...   \n",
       "Marcelo D2             rio cidade desespero a vida e boa mas so vive ...   \n",
       "Mr. Thug (Diego Thug)  faltam palavras para demonstrar nosso corre en...   \n",
       "Oriente                tava ouvindo um bob marley e te olhando dormir...   \n",
       "Projota                eu sangro por mais que voce nao perceba mais q...   \n",
       "Racionais MC's         60 dos jovens de periferia sem antecedentes cr...   \n",
       "Rashid                 cada manha com voce tem um aroma de maca e cad...   \n",
       "Sidoka                 sabe eu posso ser bem mais que julga quem sabe...   \n",
       "UCLÃ                   eu gostaria que apertasse o haxixe mas eu nem ...   \n",
       "Xamã                   voce me falou que eu curto foto de outras mina...   \n",
       "\n",
       "                                                                  tokens  \\\n",
       "artist                                                                     \n",
       "1Kilo                  [ei, para, de, me, olhar, como, se, eu, te, de...   \n",
       "3030                   [pra, ascender, e, envolver, toda, gente, deix...   \n",
       "ADL MC's               [eu, to, juntando, dinheiro, por, isso, ando, ...   \n",
       "C4bal                  [ok, ok, e, o, cabal, no, remix, junto, com, o...   \n",
       "Cacife Clan            [a, noite, comigo, hoje, e, sua, aproveite, an...   \n",
       "ConeCrewDiretoria      [voce, sabe, como, e, com, o, meu, rap, underg...   \n",
       "Costa Gold             [tanta, miseria, que, vemos, por, ai, ensinam,...   \n",
       "Dalsin                 [eu, era, fogo, perdido, ela, solucao, da, cha...   \n",
       "Diomedes Chinaski      [quantos, irmaos, que, eu, ja, vi, padecer, na...   \n",
       "Djonga                 [aliado, era, os, cara, da, zona, leste, de, p...   \n",
       "Emicida                [quebram, se, barreiras, abrem, se, as, portas...   \n",
       "Fabio Brazza           [eu, vou, te, levar, pra, ver, a, coisa, mais,...   \n",
       "Facção Central         [na, minha, infancia, quis, um, kart, nao, tiv...   \n",
       "Froid                  [original, rude, boy, original, rude, boy, ori...   \n",
       "Gabriel O Pensador     [eu, nao, sei, se, ele, era, antigo, e, tava, ...   \n",
       "Haikaiss               [voce, e, o, meu, sonho, distante, tao, releva...   \n",
       "Hungria Hip Hop        [ai, garcom, ai, garcom, ai, garcom, ai, garco...   \n",
       "MC Hariel              [ai, nos, passa, no, meio, do, baile, com, o, ...   \n",
       "MV Bill                [quando, a, noite, cai, a, pista, e, dominada,...   \n",
       "Marcelo D2             [rio, cidade, desespero, a, vida, e, boa, mas,...   \n",
       "Mr. Thug (Diego Thug)  [faltam, palavras, para, demonstrar, nosso, co...   \n",
       "Oriente                [tava, ouvindo, um, bob, marley, e, te, olhand...   \n",
       "Projota                [eu, sangro, por, mais, que, voce, nao, perceb...   \n",
       "Racionais MC's         [60, dos, jovens, de, periferia, sem, antecede...   \n",
       "Rashid                 [cada, manha, com, voce, tem, um, aroma, de, m...   \n",
       "Sidoka                 [sabe, eu, posso, ser, bem, mais, que, julga, ...   \n",
       "UCLÃ                   [eu, gostaria, que, apertasse, o, haxixe, mas,...   \n",
       "Xamã                   [voce, me, falou, que, eu, curto, foto, de, ou...   \n",
       "\n",
       "                                                               wordcloud  \\\n",
       "artist                                                                     \n",
       "1Kilo                  ei olhar devesse algo real tao devendo tipo ba...   \n",
       "3030                   ascender envolver toda gente deixa som reviver...   \n",
       "ADL MC's               juntando dinheiro ando pe ligo ando dia inteir...   \n",
       "C4bal                  ok ok cabal remix junto chris brow olha oque f...   \n",
       "Cacife Clan            noite comigo hoje aproveite antes sol caia lua...   \n",
       "ConeCrewDiretoria      sabe rap underground necessario faco uniao som...   \n",
       "Costa Gold             tanta miseria vemos ai ensinam valorizar tudo ...   \n",
       "Dalsin                 fogo perdido solucao chama mortal plano terra ...   \n",
       "Diomedes Chinaski      quantos irmaos vi padecer corrida maluca briga...   \n",
       "Djonga                 aliado cara zona leste pivete lembrar neguin p...   \n",
       "Emicida                quebram barreiras abrem portas ultrapassam fro...   \n",
       "Fabio Brazza           levar coisa bonita vi confesso desde jamais es...   \n",
       "Facção Central         infancia quis kart nenhum boy vira piloto form...   \n",
       "Froid                  original rude boy original rude boy original r...   \n",
       "Gabriel O Pensador     antigo tava congelado alguem morrido ressuscit...   \n",
       "Haikaiss               sonho distante tao relevante viver viver sonho...   \n",
       "Hungria Hip Hop        ai garcom ai garcom ai garcom ai garcom ai gar...   \n",
       "MC Hariel              ai passa meio baile dedo alto copao 700 talo l...   \n",
       "MV Bill                noite cai pista dominada arrego pago bonde sai...   \n",
       "Marcelo D2             rio cidade desespero vida boa vive medo olho a...   \n",
       "Mr. Thug (Diego Thug)  faltam palavras demonstrar corre rimamos toda ...   \n",
       "Oriente                tava ouvindo bob marley olhando dormir lembran...   \n",
       "Projota                sangro perceba receba amor forma posso dar ain...   \n",
       "Racionais MC's         60 jovens periferia antecedentes criminais sof...   \n",
       "Rashid                 cada manha aroma maca cada beijo completar cad...   \n",
       "Sidoka                 sabe posso bem julga sabe dando mole sim sabe ...   \n",
       "UCLÃ                   gostaria apertasse haxixe fumo oi tesoura ai t...   \n",
       "Xamã                   falou curto foto outras mina diz vacilo conqui...   \n",
       "\n",
       "                       Unique  Total  Relative  \n",
       "artist                                          \n",
       "1Kilo                    8068  84948  0.094976  \n",
       "3030                     6938  57760  0.120118  \n",
       "ADL MC's                 6914  42941  0.161012  \n",
       "C4bal                    7376  67189  0.109780  \n",
       "Cacife Clan              5923  44167  0.134105  \n",
       "ConeCrewDiretoria        7859  50583  0.155368  \n",
       "Costa Gold               9959  75137  0.132545  \n",
       "Dalsin                   6460  42404  0.152344  \n",
       "Diomedes Chinaski        6794  43329  0.156800  \n",
       "Djonga                   5497  35575  0.154519  \n",
       "Emicida                 10113  75413  0.134102  \n",
       "Fabio Brazza             7869  42455  0.185349  \n",
       "Facção Central          10319  82796  0.124632  \n",
       "Froid                    6485  39783  0.163009  \n",
       "Gabriel O Pensador       7496  62607  0.119731  \n",
       "Haikaiss                10472  79710  0.131376  \n",
       "Hungria Hip Hop          4520  36688  0.123201  \n",
       "MC Hariel                5310  54138  0.098083  \n",
       "MV Bill                  7943  64158  0.123804  \n",
       "Marcelo D2               6538  55622  0.117543  \n",
       "Mr. Thug (Diego Thug)    5279  48486  0.108877  \n",
       "Oriente                  6224  35924  0.173255  \n",
       "Projota                  8133  78622  0.103444  \n",
       "Racionais MC's           8052  64163  0.125493  \n",
       "Rashid                   8691  67057  0.129606  \n",
       "Sidoka                   5040  58357  0.086365  \n",
       "UCLÃ                     5838  46838  0.124642  \n",
       "Xamã                     6513  38873  0.167546  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df['Total']>35000]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {
    "height": "336.998px",
    "width": "292.983px"
   },
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "165px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}