pythonNotebookChallenge.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Libraries that will be used to build the model, train and deal with the dataset must be import to be used"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "import scipy\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import itertools\n",
    "from pandas.plotting import scatter_matrix\n",
    "from sklearn import model_selection\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.naive_bayes import GaussianNB, MultinomialNB\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset that will be used in the problem must be read, in order to do it the pandas library will be used to read the CSV file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load data GPAM_Challenge dataset \n",
    "dataset = pd.read_csv('olist_order_reviews_dataset.csv') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset contains many rows without comments, and it needs to be treated it before using it to train the model, if not there will be \"Trash\" information that will not help to predict the product's rating based in blank comment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "#take the blank comments row out\n",
    "dataset = dataset.dropna(subset=['review_comment_message'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print information of the datased. Such as it's shape, destribution, the first 10 rows and check about the rating distribution ( amount of 1, 2, 3, 4 and 5 stars) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(41753, 7)\n",
      "                           review_id                          order_id  \\\n",
      "3   e64fb393e7b32834bb789ff8bb30750e  658677c97b385a9be170737859d3511b   \n",
      "4   f7c4243c7fe1938f181bec41a392bdeb  8e6bfb81e283fa7e4f11123a3fb894f1   \n",
      "9   8670d52e15e00043ae7de4c01cc2fe06  b9bf720beb4ab3728760088589c62129   \n",
      "12  4b49719c8a200003f700d3d986ea1a19  9d6f15f95d01e79bd1349cc208361f09   \n",
      "15  3948b09f7c818e2d86c9a546758b2335  e51478e7e277a83743b6f9991dbfa3fb   \n",
      "16  9314d6f9799f5bfba510cc7bcd468c01  0dacf04c5ad59fd5a0cc1faa07c34e39   \n",
      "19  373cbeecea8286a2b66c97b1b157ec46  583174fbe37d3d5f0d6661be3aad1786   \n",
      "22  d21bbc789670eab777d27372ab9094cc  4fc44d78867142c627497b60a7e0228a   \n",
      "24  0e0190b9db53b689b285d3f3916f8441  79832b7cb59ac6f887088ffd686e1d5e   \n",
      "27  fe3db7c069d694bab50cc43463f91608  2ca73e2ff9e3a186ad1e1ffb9b1d9c10   \n",
      "\n",
      "    review_score     review_comment_title  \\\n",
      "3              5                      NaN   \n",
      "4              5                      NaN   \n",
      "9              4                recomendo   \n",
      "12             4                      NaN   \n",
      "15             5          Super recomendo   \n",
      "16             2                      NaN   \n",
      "19             1  Não chegou meu produto    \n",
      "22             5                    Ótimo   \n",
      "24             5                      NaN   \n",
      "27             5                      NaN   \n",
      "\n",
      "                               review_comment_message review_creation_date  \\\n",
      "3               Recebi bem antes do prazo estipulado.  2017-04-21 00:00:00   \n",
      "4   Parabéns lojas lannister adorei comprar pela I...  2018-03-01 00:00:00   \n",
      "9   aparelho eficiente. no site a marca do aparelh...  2018-05-22 00:00:00   \n",
      "12    Mas um pouco ,travando...pelo valor ta Boa.\\r\\n  2018-02-16 00:00:00   \n",
      "15  Vendedor confiável, produto ok e entrega antes...  2018-05-23 00:00:00   \n",
      "16  GOSTARIA DE SABER O QUE HOUVE, SEMPRE RECEBI E...  2018-01-18 00:00:00   \n",
      "19                                            Péssimo  2018-08-15 00:00:00   \n",
      "22                                       Loja nota 10  2018-07-10 00:00:00   \n",
      "24              obrigado pela atençao amim dispensada  2017-12-01 00:00:00   \n",
      "27  A compra foi realizada facilmente.\\r\\nA entreg...  2018-03-23 00:00:00   \n",
      "\n",
      "   review_answer_timestamp  \n",
      "3      2017-04-21 22:02:06  \n",
      "4      2018-03-02 10:26:53  \n",
      "9      2018-05-23 16:45:47  \n",
      "12     2018-02-20 10:52:22  \n",
      "15     2018-05-24 03:00:01  \n",
      "16     2018-01-20 21:25:45  \n",
      "19     2018-08-15 04:10:37  \n",
      "22     2018-07-11 14:10:25  \n",
      "24     2017-12-09 22:58:58  \n",
      "27     2018-04-01 00:27:51  \n",
      "       review_score\n",
      "count  41753.000000\n",
      "mean       3.640409\n",
      "std        1.626383\n",
      "min        1.000000\n",
      "25%        2.000000\n",
      "50%        4.000000\n",
      "75%        5.000000\n",
      "max        5.000000\n",
      "review_score\n",
      "1     9179\n",
      "2     2229\n",
      "3     3665\n",
      "4     6034\n",
      "5    20646\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(dataset.shape)\n",
    "print(dataset.head(10))\n",
    "print(dataset.describe())\n",
    "print(dataset.groupby('review_score').size())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "An histogram is a good way to visualize the rating distribution. How many data there are in each class (1, 2, 3, 4, 5). As the data distribution is not uniform accuracy is not a very good metric to be used when we check the histogram."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "#check data distribution\n",
    "dataset.hist()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extract the values of the dataset to an array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "array = dataset.values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Slice the comments value into a single array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split comments column\n",
    "X = array[:,4] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The rating column will be split out of the dataset array and cast into int, just to be sure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "#split rating column\n",
    "Y = array[:,2]\n",
    "Y = Y.astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset has to be split into train and also test data. the proportion used is 30% to the test set and 70 to the training set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_size = 0.3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset needs to have a seed to be used when splitting the data 5 was chosen randomly "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "#random seed to split the dataset\n",
    "seed = 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The text needs to be vectorized in order to be better understood by machines, to do so a vectorized will be created and the comments (X) will fit it.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Vectorizer to split text\n",
    "vectorizer = TfidfVectorizer(ngram_range=(1,2))\n",
    "x_vectorized = vectorizer.fit_transform(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now the splitting of the dataset will be done between training set and validation set.\n",
    "The vectorized dataset will be used the validation size will be used (30%) and the seed to split(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "#split the datset between train and validation\n",
    "X_train, X_validation, Y_train, Y_validation = model_selection.train_test_split(x_vectorized,\n",
    "                                                Y,\n",
    "                                                test_size=validation_size,\n",
    "                                                random_state=seed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Just to check the size of each dataset training and validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set has 29227 samples.\n",
      "Test set has 12526 samples.\n"
     ]
    }
   ],
   "source": [
    "print(\"Training set has {} samples.\".format(X_train.shape[0]))\n",
    "print(\"Test set has {} samples.\".format(X_validation.shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The models to train will be created and tested to check which one will have better metrics on the dataset "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Linear Classifier\n",
    "linearClassificer = LinearSVC()\n",
    "linearClassificer.fit(X_train, Y_train)\n",
    "linearPrediction = linearClassificer.predict(X_validation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check the linear classifier accuracy ( it is not the recommended metric, I've tried f1-score but I got some erros and I could solve them)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy\n",
      "LinearSVC: 0.6828197349513013\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy\")\n",
    "print(\"LinearSVC:\",accuracy_score(Y_validation,linearPrediction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#KNeighbors Classifier\n",
    "classifierKNN = KNeighborsClassifier(n_neighbors=2)\n",
    "classifierKNN.fit(X_train, Y_train)\n",
    "predictionKNN = classifierKNN.predict(X_validation)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy KNN: 0.3029698227686412\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy KNN:\", accuracy_score(Y_validation, predictionKNN))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "#MultinomialNB Classifier\n",
    "multinomialClassifier = MultinomialNB()\n",
    "multinomialClassifier.fit(X_train, Y_train)\n",
    "predictionMUB = multinomialClassifier.predict(X_validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MultinomialNB: 0.6699664697429347\n"
     ]
    }
   ],
   "source": [
    "print(\"MultinomialNB:\",accuracy_score(Y_validation, predictionMUB))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once the highest accuracy comes from the Liner Classifier, it was chosen to be used "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A function the plot the confusion matrix of the classifier needs to be built"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    if normalize:\n",
    "        cm = np.around((cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]),decimals=2)\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "create confusion matrix with the best model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "confusionMatrixLSVC = confusion_matrix(Y_validation, linearPrediction)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The plotted data will be able to be analysed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[2388   58   99   57  175]\n",
      " [ 469   32   53   47   83]\n",
      " [ 416   28  164  140  339]\n",
      " [ 155   17  110  235 1285]\n",
      " [ 139    3   66  234 5734]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "classes = ['1', '2', '3', '4', '5']\n",
    "plt.figure()\n",
    "\n",
    "plot_confusion_matrix(confusionMatrixLSVC,\n",
    "                      classes,\n",
    "                      normalize=False,\n",
    "                      title='Confusion Matrix - SVM')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As the confusion matrix picture shows, the main diagonal shows the amount of correct predicted values the  other values are the values that were wrongly predicted\n"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}