rbc_notebook_sprint1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1 style='color:#00868b'>Read, balance and clean dataset<span class=\"tocSkip\"></span></h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook reads in the data, balances it, cleans it and then saves it to a csv file. This csv can then be used in a separate notebook to train a model. This notebook also allows the user to tweak the balancing and cleaning as they see fit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Read file**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"complaints-2020-01-22_08_24.csv\", encoding=\"utf-8\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Select columns to keep**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add your columns here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_selected = df.loc[:, ('Product', 'Consumer complaint narrative')]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Remove columns with the least complaints**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_balanced1 = df_selected\n",
    "\n",
    "df_balanced1 = df_balanced1[df_balanced1[\"Product\"] != \"Virtual currency\"]\n",
    "df_balanced1 = df_balanced1[df_balanced1[\"Product\"] != \"Other financial service\"]\n",
    "df_balanced1 = df_balanced1[df_balanced1[\"Product\"] != \"Prepaid card\"]\n",
    "df_balanced1 = df_balanced1[df_balanced1[\"Product\"] != \"Money transfers\"]\n",
    "df_balanced1 = df_balanced1[df_balanced1[\"Product\"] != \"Payday loan\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Make a copy and iteratively drop the products with the most complaints, adding a random sample % of them back afterwards**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(480700, 2)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_balanced2 = df_balanced1.copy()\n",
    "df_balanced1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "indexNames = df_balanced2[df_balanced2[\"Product\"] == \"Mortgage\"].index\n",
    "df_balanced2.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced2[df_balanced2[\"Product\"] == \"Debt collection\"].index\n",
    "df_balanced2.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced2[df_balanced2[\"Product\"] == \"Credit reporting, credit repair services, or other personal consumer reports\"].index\n",
    "df_balanced2.drop(indexNames, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_balanced2 = df_balanced2.append(df_balanced1.loc[df_balanced1[\"Product\"] == 'Mortgage'].sample(frac=0.5))\n",
    "df_balanced2 = df_balanced2.append(df_balanced1.loc[df_balanced1[\"Product\"] == 'Debt collection'].sample(frac=0.4))\n",
    "df_balanced2 = df_balanced2.append(df_balanced1.loc[df_balanced1[\"Product\"] == 'Credit reporting, credit repair services, or other personal consumer reports'].sample(frac=0.3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(287475, 2)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_balanced3 = df_balanced2.copy()\n",
    "df_balanced2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Credit reporting, credit repair services, or other personal consumer reports\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Debt collection\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Credit reporting\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Credit card or prepaid card\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Mortgage\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)\n",
    "indexNames = df_balanced3[df_balanced3[\"Product\"] == \"Student loan\"].index\n",
    "df_balanced3.drop(indexNames, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Credit reporting, credit repair services, or other personal consumer reports'].sample(frac=0.5))\n",
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Debt collection'].sample(frac=0.5))\n",
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Credit reporting'].sample(frac=0.6))\n",
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Credit card or prepaid card'].sample(frac=0.6))\n",
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Mortgage'].sample(frac=0.6))\n",
    "df_balanced3 = df_balanced3.append(df_balanced2.loc[df_balanced2[\"Product\"] == 'Student loan'].sample(frac=0.85))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(203874, 2)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_balanced3.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'categories')"
      ]
     },
     "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": [
    "diff_products = df_balanced3.groupby(\"Product\")[\"Product\"].count().sort_values(ascending=False)\n",
    "fig = plt.figure()\n",
    "plt.rcParams['figure.figsize'] = [30,15]\n",
    "ax = fig.add_axes([0,0,0.4,0.5])\n",
    "\n",
    "ax.barh(diff_products.index,diff_products,color='teal')\n",
    "ax.invert_yaxis()\n",
    "plt.title('Complaints per Products')\n",
    "plt.xlabel('amount')\n",
    "plt.ylabel('categories')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data cleaning & lemmatization**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "import string\n",
    "\n",
    "# Remove urls\n",
    "def clean_url(complaint):\n",
    "    # to do: more regex url garbage matching\n",
    "    complaint = re.sub('https?://(?:[-\\w.]|(?:%[\\da-fA-F]{2}))+', '', complaint)\n",
    "    complaint = re.sub('https? ?: ?// ?(?:[-\\w.]|(?:%[\\da-fA-F]{2}))+', '', complaint)\n",
    "    return complaint\n",
    "\n",
    "# Remove punctuation from complaint\n",
    "def clean_punctuation(complaint):\n",
    "    complaint = re.sub('[%s]' % re.escape(string.punctuation), '', complaint)\n",
    "    return complaint\n",
    "\n",
    "# Remove non-sensical characters from complaint\n",
    "def clean_nonsense(complaint):\n",
    "    complaint = re.sub('[''\"\"...]', '', complaint)\n",
    "    complaint = re.sub('\\n', '', complaint)\n",
    "    return complaint\n",
    "\n",
    "# Remove censored words from complaint\n",
    "def clean_censored(complaint):\n",
    "    complaint = re.sub('[XXXX]', '', complaint)\n",
    "    return complaint\n",
    "\n",
    "# Turn everything into lowercase\n",
    "def clean_lowercase(complaint):\n",
    "    complaint = complaint.lower()\n",
    "    return complaint\n",
    "\n",
    "# Remove numbers from complaint\n",
    "def clean_numbers(complaint):\n",
    "    complaint = re.sub(\"\\d+\", '', complaint)\n",
    "    return complaint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_url)\n",
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_punctuation)\n",
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_nonsense)\n",
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_censored)\n",
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_lowercase)\n",
    "df_balanced3[\"Consumer complaint narrative\"] = df_balanced3[\"Consumer complaint narrative\"].apply(clean_numbers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(203874, 2)\n"
     ]
    },
    {
     "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>Product</th>\n",
       "      <th>Consumer complaint narrative</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "      <td>they would not let me pay my loan off  days be...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "      <td>service finance are liars and are charging me ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Checking or savings account</td>\n",
       "      <td>over draft fees due to fraudulent charges subm...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>Vehicle loan or lease</td>\n",
       "      <td>on  i signed a car loan agreement to finance m...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>Money transfer, virtual currency, or money ser...</td>\n",
       "      <td>we hired   and    debt collection to handle co...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              Product  \\\n",
       "9           Payday loan, title loan, or personal loan   \n",
       "12          Payday loan, title loan, or personal loan   \n",
       "27                        Checking or savings account   \n",
       "38                              Vehicle loan or lease   \n",
       "44  Money transfer, virtual currency, or money ser...   \n",
       "\n",
       "                         Consumer complaint narrative  \n",
       "9   they would not let me pay my loan off  days be...  \n",
       "12  service finance are liars and are charging me ...  \n",
       "27  over draft fees due to fraudulent charges subm...  \n",
       "38  on  i signed a car loan agreement to finance m...  \n",
       "44  we hired   and    debt collection to handle co...  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df_balanced3.shape)\n",
    "df_balanced3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pickle for later use\n",
    "df_balanced3.to_pickle(\"corpus_balanced3_cleaned.pkl\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Run the lemmatization.py script in Pycharm to lemmatize the balanced & cleaned dataset.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This will produce a csv, <code>corpus_balanced3_cleaned_lemmatized.csv</code> that can be read into any notebook used for training using the code below:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<code> my_df = pd.read_csv(\"corpus_balanced3_cleaned_lemmatized.csv\", encoding=\"utf-8\")</code>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "my_df = pd.read_csv(\"corpus_balanced3_cleaned_lemmatized.csv\", encoding=\"utf-8\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "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>Product</th>\n",
       "      <th>Consumer complaint narrative</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "      <td>they would not let me pay my loan off day befo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Payday loan, title loan, or personal loan</td>\n",
       "      <td>service finance are liar and are charging me i...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Checking or savings account</td>\n",
       "      <td>over draft fee due to fraudulent charge submit...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Vehicle loan or lease</td>\n",
       "      <td>on i signed a car loan agreement to finance my...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Money transfer, virtual currency, or money ser...</td>\n",
       "      <td>we hired and debt collection to handle collect...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             Product  \\\n",
       "0          Payday loan, title loan, or personal loan   \n",
       "1          Payday loan, title loan, or personal loan   \n",
       "2                        Checking or savings account   \n",
       "3                              Vehicle loan or lease   \n",
       "4  Money transfer, virtual currency, or money ser...   \n",
       "\n",
       "                        Consumer complaint narrative  \n",
       "0  they would not let me pay my loan off day befo...  \n",
       "1  service finance are liar and are charging me i...  \n",
       "2  over draft fee due to fraudulent charge submit...  \n",
       "3  on i signed a car loan agreement to finance my...  \n",
       "4  we hired and debt collection to handle collect...  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Train in your separate notebook**\n",
    "\n",
    "Use <code>RandomForest_run_006.ipynb</code> as an example for structure. Make a copy, rename is and replace what you need. [notebook](RandomForest_run_006.ipynb)"
   ]
  },
  {
   "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": "3.7.7"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}