Lyrics_Classification_Project.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lyrics Classification Project"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 1. Importing libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "import re\n",
    "from bs4 import BeautifulSoup\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import random\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfTransformer\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.decomposition import LatentDirichletAllocation\n",
    "\n",
    "import spacy \n",
    "nlp = spacy.load('en_core_web_sm')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 2. To scrape the lyrics use lyrics_scraper.py to save lyrics as csv files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ETTA = 'Etta'\n",
    "BILLY = 'Billy'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 3. Extracting clean text from lyrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def lyrics_cleaner(lyrics):\n",
    "    \"\"\" Clean html leftovers with regex \"\"\"\n",
    "    lyrics_clean = []\n",
    "    for i in lyrics:\n",
    "        text = re.sub('\\r|\\s+|\\(|\\)|\\[|\\]|\\,|\\?|\\!|\\-|\\\"|\\:|\\;|\\.|\\“|\\\"|\\”|CHORUS', ' ', i)\n",
    "        lyrics_clean.append(text)\n",
    "    word_end = []\n",
    "    for j in lyrics_clean:\n",
    "        text2 = re.sub('''(?:_a)?in'\\s''', 'ing ', j)\n",
    "        word_end.append(text2)\n",
    "    return word_end\n",
    "\n",
    "def import_data(first_name):\n",
    "    \"\"\" Import csv, clean lyrics syntax, drop duplicates, convert back to list \"\"\"\n",
    "    l = lyrics_cleaner(pd.read_csv(f'{first_name}_cleaned.csv')['0'].tolist())\n",
    "    return l\n",
    "\n",
    "def drop_duplicates(spacyfied_lyr):\n",
    "    \"\"\" Take first 6 words of the lyrics, add as new column, drop duplicate rows, join back to original df \"\"\"\n",
    "    splitted_text =[]\n",
    "    joined_back= []\n",
    "    df = pd.DataFrame(spacyfied_lyr)\n",
    "    for i in spacyfied_lyr:\n",
    "        splitted = i.lower().split()\n",
    "        first_six = splitted[0:6]\n",
    "        splitted_text.append(first_six)\n",
    "    for j in splitted_text:\n",
    "        joined = ' '.join(j)\n",
    "        joined_back.append(joined)\n",
    "    drop_dup = pd.DataFrame(joined_back).drop_duplicates()\n",
    "    merged = pd.merge(df, drop_dup, how='inner', left_index=True, right_index=True).iloc[:,0].to_list()\n",
    "    return merged"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "etta_lyr = drop_duplicates(import_data(ETTA))\n",
    "del etta_lyr[176], etta_lyr[91]  # Unclean data on lyrics.com: Gucci Mane song instead of Etta's song -> del\n",
    "billy_lyr = drop_duplicates(import_data(BILLY))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 4. Balancing classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Etta James' songs: 224\n",
      "Number of Billy Joel's songs: 151\n"
     ]
    }
   ],
   "source": [
    "print (f'''Number of Etta James' songs: {len(etta_lyr)}\n",
    "Number of Billy Joel's songs: {len(billy_lyr)}''')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "random.seed(10)\n",
    "index = random.sample(range(len(etta_lyr)), len(etta_lyr)-len(billy_lyr))\n",
    "\n",
    "def dropped_songs(etta_lyr):\n",
    "    \"\"\" Keep dropped songs to use later for model evaluation \"\"\"\n",
    "    etta_dropped = []\n",
    "    for i in index:\n",
    "        song = etta_lyr[i]\n",
    "        etta_dropped.append(song)\n",
    "    return etta_dropped \n",
    "\n",
    "def random_delete(song_lyr):\n",
    "    \"\"\" Deletes randomly selected songs to even the classes \"\"\"\n",
    "    for i in sorted(index, reverse=True):\n",
    "        del song_lyr[i] \n",
    "    return song_lyr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "etta_dropped = dropped_songs(etta_lyr)\n",
    "etta_lyr = random_delete(etta_lyr)\n",
    "lyrics = etta_lyr + billy_lyr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Natural Language Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 6. Token preprocessing with Spacy "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def spacify_my_text(lyrics_text):\n",
    "    \"\"\" Use Spacy to lowercae all characters, lemmatize words and remove stop words \"\"\"\n",
    "    spacyfied = []\n",
    "    for sentence in lyrics_text:\n",
    "        parsed_sentence = nlp(sentence.lower())\n",
    "        treated_sentence = ''\n",
    "        for token in parsed_sentence:\n",
    "            if not token.is_stop:\n",
    "                treated_sentence += str(token.lemma_) + ' '\n",
    "        spacyfied.append(treated_sentence.strip())\n",
    "    return spacyfied\n",
    "\n",
    "lyr_spacy = spacify_my_text(lyrics)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 7. Train test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.Series(lyr_spacy)\n",
    "y = [ETTA] * len(etta_lyr) + [BILLY] * len(billy_lyr)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 8. Converting text into a numerical matrix with Bag of Words and using Multinominal Naive Bayes classifier to train the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cv_tf():\n",
    "    \"\"\" Convert text to matrix of token counts, normalize the representation of the matrix \"\"\"\n",
    "    cv = CountVectorizer(min_df=5)\n",
    "    tf = TfidfTransformer()\n",
    "    return cv, tf\n",
    "\n",
    "def fit_transform_X(X_input):\n",
    "    \"\"\" Count Vectorizer converts vector into a matrix, TD-IDF normalizes the matrix \"\"\"\n",
    "    X_vec = cv.fit_transform(X_input)\n",
    "    X_tf = tf.fit_transform(X_vec)\n",
    "    return X_tf\n",
    "\n",
    "def transform_X(X_input):\n",
    "    \"\"\" Transform vector to matrix, according to training fit_transform \"\"\"\n",
    "    X_vec = cv.transform(X_input)\n",
    "    X_tf = tf.transform(X_vec)\n",
    "    return X_tf\n",
    "\n",
    "def naive_bayes(X_tf, y_train):\n",
    "    \"\"\" Count Vectorizer and TD-IDF \"\"\"\n",
    "    m = MultinomialNB()\n",
    "    m.fit(X_tf, y_train)\n",
    "    return m\n",
    "\n",
    "def model_score(X, y, model):\n",
    "    \"\"\" Calculates the training and validation accuracy \"\"\"\n",
    "    score = model.score(X, y).round(4)\n",
    "    val_acc = cross_val_score(model, X, y, cv=5).round(4)\n",
    "    return score, val_acc    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv, tf = cv_tf()\n",
    "X_train_tf = fit_transform_X(X_train)\n",
    "model = naive_bayes(X_train_tf, y_train) \n",
    "score_train, val_acc_train = model_score(X_train_tf, y_train, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test_tf = transform_X(X_test)\n",
    "score_test, val_acc_test = model_score(X_test_tf, y_test, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training accuracy: 0.9602\n",
      "Training cross-validation accuracy mean: 0.8051\n",
      "Test accuracy: 0.75\n",
      "Test cross-validation accuracy mean: 0.6317\n"
     ]
    }
   ],
   "source": [
    "print(f'Training accuracy: {score_train}')\n",
    "print(f'Training cross-validation accuracy mean: {val_acc_train.mean().round(4)}')\n",
    "print(f'Test accuracy: {score_test}')\n",
    "print(f'Test cross-validation accuracy mean: {val_acc_test.mean().round(4)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 9. Word occurence mapping \n",
    "More insigihts from the analysis -> which words are likely to belong to which artist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23880b845f8>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def word_occurence():\n",
    "    \"\"\" Creates table of most occurring words using log probabilities of features \"\"\"\n",
    "    feature_log_probs = model.feature_log_prob_\n",
    "    diff = feature_log_probs[0]-feature_log_probs[1]\n",
    "    likelyhood = pd.DataFrame(data=diff, index = sorted(cv.vocabulary_.keys())\n",
    "                             ).sort_values(by=0).rename({0: ' '}, axis = 1)\n",
    "    head = likelyhood.head(15)\n",
    "    tail = likelyhood.tail(15)\n",
    "    important_words = pd.concat([head, tail])\n",
    "    fig = important_words.plot(kind='bar', \n",
    "                           figsize=(15,5), \n",
    "                           fontsize=13, \n",
    "                           legend=None, \n",
    "                           color=[np.where(important_words.iloc[:,0]>0, 'g', 'r')], \n",
    "                           alpha=0.5)\n",
    "    plt.title('Most occurring words in Etta James (left) and Billy Joel (right) lyrics ')\n",
    "    return fig \n",
    "\n",
    "word_occurence()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 10. Testing the model on 73 Etta James songs that were taken out to balance the classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 2 artists>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def prediction_tester(etta_dropped):\n",
    "    \"\"\" Predicts lyrics for Etta James songs \"\"\"\n",
    "    etta_tf = transform_X(spacify_my_text(etta_dropped))\n",
    "    y_true = len(etta_dropped) * [ETTA]\n",
    "    score = model.score(etta_tf, y_true).round(2)\n",
    "    prediction = model.predict(etta_tf)\n",
    "    unique, counts = np.unique(prediction, return_counts=True)\n",
    "    fig2 = plt.bar(unique, counts, color=('r', 'g'), alpha = 0.4)\n",
    "    plt.title(f'Model accuracy on 73 Etta songs: {score}')\n",
    "    return fig2\n",
    "prediction_tester(etta_dropped)                            "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Potential for improvement\n",
    "Billy Joel has more unique words per song -> balance lyrics distribution "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def lyrics_distribution(lyrics, colname):\n",
    "    df = pd.DataFrame(lyrics, columns=[colname])\n",
    "    df_cleanup = df[colname].str.lower().str.split()\n",
    "    df['LyricsCounter'] = df_cleanup.apply(set).apply(len)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "billy_df = lyrics_distribution(billy_lyr, BILLY)\n",
    "etta_df = lyrics_distribution(etta_lyr, ETTA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Lyrics length distribution - Etta (red), Billy (blue)')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEICAYAAACpqsStAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAZaUlEQVR4nO3de7xcZX3v8c+XEBJlByKExJAEEq4KeKSyD+YUD+4gWEERbC2CLYQeeOUcj/BSjtVS7dHYehS1BS+H+moUahAxUpQDjdIagV205SKp3GIk4SaXbJNyJ9GEBH7nj/VMmOzM7JnZc9vP7O/79ZrXnlmXZ/2edfnNs5611mxFBGZmlqdduh2AmZmNnpO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llzEm8jKQDJG1sYXm7SgpJc1tVZgPLPl7SI03M/xlJ30zvW71eviHp4+l9U3FWKHtA0qpWlTfWSXqDpNtbWN727aHCnZJe18LyF0q6Ib3f4fiQdKWkxaMsd7KkX0qaXqusVh6Xkr4i6dxmy2lGTyRxSY9IOr7ZciLioYjoa0VMndTuL4t614ukcyUN1lHeuRHx2WbjqlTviBiMiMObLXuU8Twu6beSNpa9vpTG7bRumklaZT4DfLHJMiqK4iGSi4FP1ztPqtOLqe7Ppy+Bt5SVuTQiTmxDuB8AfhwRG9pQ9ki+AHxS0q4dXu52PZHEW6GbG2E8kTSh2zG02YkR0Vf2+nC7FiRpNvAW4B+rjG/FPv3/gLeXWrh1+mz60p8KfAP4viS1IJaR/HfgW21exk4i4nHgQeBdnV52SU8n8XR6dWLZ50mSnpF0hKSDUivuTyQ9CvyoNKxs+r0lfVPSUJrve2n4dEk/lPSspKcl3VJnPJMlXSzpMUnrJf2tpMlp3PHpjOJjkv5D0jpJZ5XNu4+kH6TWzR2SPlvWsistf1VqAf1B2XwVy6sQ2wGSfiLpBUn/DOxdNm74ejknxfqCpIcknS7pDcD/Bf5riuHJNO2Vki6V9E+SNqXxO7VAJX1S0lOSHpZ0etnwn0o6u+xzeYt2p3prWPeMpMMl/UvaVvdKemfZuCvT6fANqS63SppXbR2NVqV1I+l/Au8DPp6GXZum/Yu0Tl+QtErSu0co+u3AzyJiS9myHpf0UUn3Ar9Jw2ZLujbtBw9L+mDZ9K+W9K20f68CjipfQET8BrgLOKHRekfEy8BVwD7pVffZ2kjHboVpDwDmAHcOG7WPpBvTurxZ0pwqyxppH0PSYZJ+nI71X5YfX8kg8E66pKeTOHAF8Mdln98FPBIR95UNOxZ4HZU3wlXAbsBhwAzgy2n4R4GHKHbM1wL/u854/hqYB/wn4GBgLvCJsvGzgVcB+wL/A/iapD3SuK8Bz6Y4/huwcFgdAA5Prb/v1VHecMuA24BpwEXAmZUmSvNfDJwQEVOAY4B7IuJe4DzgJymGaWWzvZ/ilHwKcGuFYmencfsC5wCXSzqoSpzlqtW7FOtuwHLgBxTb6gLgu8PKfj/F9tsLeBT4qzqW25BK6yYi/hb4LqnVGhHvSZOvoVinewL/B7hK0owqRb8BuL/C8NOBE4E9VZz5LAd+BsyiSMYflfS2NO1fUiTAA4CT2HG/KlkNvLGROsP2s66zKFqqTzY4ez3HbskbgAci4qVhw/8Y+CTFPv0LRtFSlzQFWJHimQ78EbBE0qFlk41q/bRKryfxbwEnSyr1557JzhvyUxHxm4j4bfnA9K39NuADEfFMRLwYEaWW31aKhLNfGv4vtQKRtAtwLvDhVN7zwOcoDriSzcBnImJrRFwPbAEOkTQROBX4ZET8Nu3I9eyQFcurENsBwJFpXWyJiJuBH45QbgBHSJocEUMR8YsacVwbEbdGxMvlrcYyL5ct+ybgn4A/rKN+tRxD8SX8xbQOfgzcwI7r/JqIuDMitgLfplgPzVieWv2l1580MnNEXJ3W6csRcRXwCNBfZfKpwAsVhn85Ih5P+/R8YI+I+GzaVx8ALuOVdXAaxT7yTET8iuKMYbgX0rLqdaGkZ4FNFA2Xv0it8kbUc+yWVFsP/xgR/5r2uY8Dx0qa2WAc7wbWRMQVEbEtIlZSdDG9t2yaRtdPS/V0Eo+Ix4A7gPdI2ovi9POqYZM9VmX2OcCTEfFchXEXAb8CbpT0oKSP1hHOa4FJwN2lA5yihVTe1/jksNbEb4A+itb3hGGxVou7XLXyhtsXeCqdOpf8qlKB6cvnDOCDwK8lLZe00xfDMLVirbTsfWvMU499gUdjx195+xVFi7Tk12Xvq62f0h01pYuVHxthme+KiKllr79vJGBJZ0sq30deR9GSrOQZijOY4crX9/7AfuVfLMDHKPZHgJnDpq+03adQnAXW66KImEpxFng0cImkhrpj6jx2S2quh3QcP0fj+9X+wDHD1t/7KNZbSaPrp6V6OoknSylOq94H3BIR5Qctww7wco8B0yp1P0TE8xFxQUTMpWgh/5mkt9aIYz3wInBo2QG+Z0TsWUcd1lO0VmeXDSvv32v2pyiHgL0lvaps2H7VJo6IGyLieIod+QHg72rEUSu+Sstel95vAl5dNu61Ze9rlbsOmCPtcFFtP+CJGvPtJN1RU7pY+YVG56dyrDsMS2dEX6O402LvlAh/CVS7KHgPFc6shpX7GLB22BfLlIg4OY3/NTvuS5W2++uBu6vEUFUU7qHophtNn/GIx26Ze4ADtfNF8+31krQnRRfVOnY20j72GHDjsPXXFxHnlU0zqvXTKr2UxCequHBYepWuzH8feDNFn+QV9RaWWgI/Bi6VNFXSREnHAkg6WdKBKTk8B7yUXiOV9xLFlfovqbhIqXTB6e11xLKV4hTu05JeJelwyvoLU9lPUfRrNiwiHqQ4EBZL2i3Vs+JBJ2lmqv+rKb6UNvFK3dcDs1P3TyN2KVv2AEV/7jVp3F3AH6R6H0JxPaAUd616/xuwDfhI2n7HUfT7Xt1gfK1Qad2sZ8fY+ygS8H9Q3KZ9LkVLvJofAf859f1XcyvwoqSPpONigop7y0sXMK+muLg6VdJ+FMfJdunL9UiKY6H8ts63UAdJhwG/C4zm3v26jt2IeITiesZRw0adLOm/SJpEcSvmTyNiqEIRVfcx4HrgcEnvT/vQRElHD+sTfytFN11X9FIS/yHw27LXYoCI2ESRAPdLfxtRSpRrKA6489PnQ4GbgI3Av1L0Qf60jvI+QnG6egdF8v8RxQXOenyA4o6R9cDfA9+h6OMu+RTFRbBnJf1+nWWWO52iD/lpiout1fofJ1Bc2B2iSKC/yysH/gpgLbBeUrVWUyWPU3wZDFG0vs6NiLVp3F9TJLYNwOXAlcPmrVrv1Bd6MnAKxYW1rwDvj4g1DcTWqBu0433i/5CGV1o33wDeqOKui2tSq/UrFPvHEEUCr/ogT0SsA35CUcdq02yj+OI6mqJ//UmKM6fSGean0rIeoUhEw5PlqcCKiFifPs8BngcqXWAsKd1xszGV+XWKfviGNHjs/h07X4y/kiJ5P0lxM0HFi/WMsI+lbpjfo8gFQxRnLp+j6BpF0iyKY7jibZ6doOq9Cb1D0l9SXIQ8u9uxtIqkvwGmRsQ53Y7FukfF7Ytfj4j5bShbFHe1nBkRq9Ows4EDI6LeO7KajaGuY1fFrbo/B94aHXzgR9KXgVURsaRTy9wphl5P4pL2pjhdel9E/Fu34xmtdFo6gaIF9GaKM4+zImJ5VwMza5NeOXbbrZe6U3Yi6QMUfWXX9cBOsAfFKeUmiq6Ui5zArVf12LHbVj3fEjcz62U93RI3M+t1Hf3Rp2nTpsXcuXMbnm/Tpk3svvvurQ9ojBuv9YbxW3fXe/ypp+4rV658MiL2qTSuo0l87ty53Hnn8N+oqW1wcJCBgYHWBzTGjdd6w/itu+s9/tRTd0kVn6AGd6eYmWXNSdzMLGNO4mZmGXMSNzPLmJO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llrKNPbJrVcsmK4v81zNq8Zfv7TrjghFr/JtRsbHJL3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMuYkbmaWMSdxM7OMOYmbmWXMSdzMLGNO4mZmGXMSNzPLmJO4mVnGnMTNzDLmXzHMzeLFY6MMMxsT3BI3M8uYk7iZWcacxM3MMuYkbmaWMSdxM7OM1Z3EJU2Q9HNJy9PneZJul7RW0ncl7da+MM3MrJJGWuIfAlaXff48cElEHAw8A5zTysDMzKy2upK4pNnAO4FvpM8CjgOuSZMsBU5tR4BmZladIqL2RNI1wOeAKcCfAmcDt0XEQWn8HOCGiDiiwryLgEUAM2bMOGrZsmUNB7lx40b6+voani93Fes9NNR8wTNnNl9Gm2x4YQsAE1/ewtZdJnVsudOndG5ZI/G+Pv7UU/cFCxasjIj+SuNqPrEp6V3AhohYKWmgNLjCpBW/DSJiCbAEoL+/PwYGBipNNqLBwUFGM1/uKta7FU9bnnFG82W0ySUr1gAwa/PDPDF5XseWe9rAIR1b1ki8r48/zda9nsfujwHeLekkYDKwB/AlYKqkXSNiGzAbWDfqKMzMbFRq9olHxJ9HxOyImAucDtwUEX8E3Ay8N022ELiubVGamVlFzdwn/mfA/5L0ALA3cFlrQjIzs3o19CuGETEIDKb3DwFHtz4kMzOrl5/YNDPLmJO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llzEnczCxjTuJmZhlzEjczy5iTuJlZxpzEzcwy5iRuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMuYkbmaWMSdxM7OMNfTf7s161SUr1nRt2ReccEjXlm35c0vczCxjTuJmZhlzEjczy5iTuJlZxpzEzcwy5iRuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMuYkbmaWMSdxM7OM1UzikiZLukPS3ZJWSfp0Gj5P0u2S1kr6rqTd2h+umZmVq6clvgU4LiLeCBwJvEPSfODzwCURcTDwDHBO+8I0M7NKaibxKGxMHyemVwDHAdek4UuBU9sSoZmZVaWIqD2RNAFYCRwEXAp8EbgtIg5K4+cAN0TEERXmXQQsApgxY8ZRy5YtazjIjRs30tfX1/B8uatY76Gh5gueObP5MtpkwwtbAJj48ha27jKpy9F0xvQpr9TT+/r4U0/dFyxYsDIi+iuNq+t/bEbES8CRkqYC1wKvrzRZlXmXAEsA+vv7Y2BgoJ5F7mBwcJDRzJe7ivVevLj5gs84o/ky2qT0vy5nbX6YJybP63I0nXHawCv/Y9P7+vjTbN0bujslIp4FBoH5wFRJpS+B2cC6UUdhZmajUs/dKfukFjiSXgUcD6wGbgbemyZbCFzXriDNzKyyerpTZgJLU7/4LsDVEbFc0i+AZZI+A/wcuKyNcZqZWQU1k3hE3AP8ToXhDwFHtyMoMzOrj5/YNDPLmJO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llzEnczCxjTuJmZhmr6wewrMfU+BGtWx98qmYRt511fouCMbNmuCVuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMuYkbmaWMSdxM7OMOYmbmWXMSdzMLGNO4mZmGXMSNzPLmJO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llzEnczCxjTuJmZhlzEjczy9iu3Q5gXFm8uLHpDz208XnMbFxxS9zMLGNO4mZmGXMSNzPLmPvEbVTmX/HVpsu47azzWxCJ2fjmlriZWcZqJnFJcyTdLGm1pFWSPpSG7yVphaS16e9r2h+umZmVq6clvg34SES8HpgPfFDSYcCFwI0RcTBwY/psZmYdVDOJR8RQRPx7ev8CsBqYBZwCLE2TLQVObVeQZmZWmSKi/omlucAtwBHAoxExtWzcMxGxU5eKpEXAIoAZM2YctWzZsoaD3LhxI319fQ3PN+YMDTU0+cZJk+jbsqVNwVS3afO2zixn2vSq4ya+vIWtu0zqSBzdNn3KK/XsmX29QeO13lBf3RcsWLAyIvorjav77hRJfcD3gA9HxPOS6povIpYASwD6+/tjYGCg3kVuNzg4yGjmG3MafPpy8NBDGbj//vbEMoJbH3yqI8tZM8LdKbM2P8wTk+d1JI5uO23gkO3ve2Zfb9B4rTc0X/e67k6RNJEigX87Ir6fBq+XNDONnwlsGHUUZmY2KvXcnSLgMmB1RFxcNup6YGF6vxC4rvXhmZnZSOrpTjkGOBO4V9JdadjHgYuAqyWdAzwK/GF7QjQzs2pqJvGI+ClQrQP8ba0Nx8zMGuEnNs3MMuYkbmaWMSdxM7OM+VcMx7BNm7d17J5tM8uTW+JmZhlzEjczy5iTuJlZxpzEzcwy5gub1jUj/Yu3p487mvk3La9Zhv/Fm413bombmWXMSdzMLGNO4mZmGXMSNzPLmJO4mVnGnMTNzDLmJG5mljEncTOzjDmJm5llzEnczCxjTuJmZhlzEjczy5h/AMusyy5ZsWb7+1mbt+zwuZ0uOOGQjizH2sstcTOzjDmJm5llzEnczCxjTuJmZhlzEjczy5iTuJlZxpzEzcwy5iRuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMlYziUu6XNIGSfeVDdtL0gpJa9Pf17Q3TDMzq6Selvg3gXcMG3YhcGNEHAzcmD6bmVmH1UziEXEL8PSwwacAS9P7pcCpLY7LzMzqoIioPZE0F1geEUekz89GxNSy8c9ERMUuFUmLgEUAM2bMOGrZsmUNB7lx40b6+voanq+lhoY6vsjndpvErs9v6vhyx4Jte+xeV903TZvegWg6Z+LLW9i6y6SOLGv6lM4spx5j4hjvknrqvmDBgpUR0V9pXNv/x2ZELAGWAPT398fAwEDDZQwODjKa+Vpq8eKOL/IH+x/IXjfd0fHljgVPH3d0XXVfc9b5HYimc2ZtfpgnJs/ryLJOGxg7/2NzTBzjXdJs3Ud7d8p6STMB0t8No47AzMxGbbRJ/HpgYXq/ELiuNeGYmVkj6rnF8DvArcChkh6XdA5wEXCCpLXACemzmZl1WM0+8Yg4o8qot7U4FjMza5Cf2DQzy5iTuJlZxpzEzcwy5iRuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZa/sPYJm10/wrvtp0Gbf12I9o2fjilriZWcacxM3MMuYkbmaWMSdxM7OMOYmbmWXMSdzMLGNO4mZmGXMSNzPLmJO4mVnGnMTNzDLmJG5mljEncTOzjPkHsGq4ZMUaAOY/+FTnF77/gZ1fpo0bpX27Gy444ZCuLbvXuCVuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZcxI3M8uYk7iZWcacxM3MMuYkbmaWMT+xaePe/Cu+2nQZt511fkviePq4o5l/0/KuxtEJw58WnbV5S1efIG23dj6h6pa4mVnGnMTNzDLmJG5mlrFs+sR7ub/M8teKfnWz0XBL3MwsY00lcUnvkHS/pAckXdiqoMzMrD6jTuKSJgCXAicChwFnSDqsVYGZmVltzbTEjwYeiIiHIuJFYBlwSmvCMjOzeigiRjej9F7gHRFxbvp8JvDmiDhv2HSLgEXp46HA/aNY3DTgyVEFmrfxWm8Yv3V3vcefeuq+f0TsU2lEM3enqMKwnb4RImIJsKSJ5SDpzojob6aMHI3XesP4rbvrPf40W/dmulMeB+aUfZ4NrGuiPDMza1AzSfxnwMGS5knaDTgduL41YZmZWT1G3Z0SEdsknQf8MzABuDwiVrUssh011R2TsfFabxi/dXe9x5/muptHe2HTzMy6z09smpllzEnczCxjYzqJj7fH+iU9IuleSXdJujMN20vSCklr09/XdDvOZkm6XNIGSfeVDatYTxW+kvaBeyS9qXuRN69K3RdLeiJt97sknVQ27s9T3e+X9Hvdibp5kuZIulnSakmrJH0oDe/p7T5CvVu3zSNiTL4oLpY+CBwA7AbcDRzW7bjaXOdHgGnDhn0BuDC9vxD4fLfjbEE9jwXeBNxXq57AScANFM8lzAdu73b8baj7YuBPK0x7WNrvJwHz0vEwodt1GGW9ZwJvSu+nAGtS/Xp6u49Q75Zt87HcEvdj/YVTgKXp/VLg1C7G0hIRcQvw9LDB1ep5CnBFFG4Dpkqa2ZlIW69K3as5BVgWEVsi4mHgAYrjIjsRMRQR/57evwCsBmbR49t9hHpX0/A2H8tJfBbwWNnnxxm58r0ggB9JWpl+rgBgRkQMQbFDANO7Fl17VavneNkPzkvdBpeXdZn1ZN0lzQV+B7idcbTdh9UbWrTNx3ISr+ux/h5zTES8ieKXIT8o6dhuBzQGjIf94GvAgcCRwBDwN2l4z9VdUh/wPeDDEfH8SJNWGJZt3SvUu2XbfCwn8XH3WH9ErEt/NwDXUpxGrS+dRqa/G7oXYVtVq2fP7wcRsT4iXoqIl4Gv88rpc0/VXdJEikT27Yj4fhrc89u9Ur1buc3HchIfV4/1S9pd0pTSe+DtwH0UdV6YJlsIXNedCNuuWj2vB85KdyvMB54rnX73imF9ve+h2O5Q1P10SZMkzQMOBu7odHytIEnAZcDqiLi4bFRPb/dq9W7pNu/21dsaV3ZPoria+yDwiW7H0+a6HkBxVfpuYFWpvsDewI3A2vR3r27H2oK6fofiFHIrRcvjnGr1pDi9vDTtA/cC/d2Ovw11/1aq2z3pIJ5ZNv0nUt3vB07sdvxN1PstFN0C9wB3pddJvb7dR6h3y7a5H7s3M8vYWO5OMTOzGpzEzcwy5iRuZpYxJ3Ezs4w5iZuZZcxJ3MwsY07iZmYZ+//EZ2ctejTEswAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "billy_df['LyricsCounter'].hist(alpha=0.5)\n",
    "etta_df['LyricsCounter'].hist(color='r', alpha=0.5)\n",
    "plt.title('Lyrics length distribution - Etta (red), Billy (blue)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}