Customer_segmination.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1a00c042-14d8-4922-866f-5ff8fe43dc27",
   "metadata": {},
   "source": [
    "# Task I - Cory Wilson"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fc1c6d98",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "### Initial Imports and Dataframe creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "930f8ba2",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "### Importing required packages\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.cluster import KMeans\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "pd.options.display.float_format = '{:.5f}'.format\n",
    "pd.options.display.max_columns = None\n",
    "\n",
    "### Creating initial data set from the Churn CSV\n",
    "data = pd.read_csv('churn_clean.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83bea652-b199-465b-a59c-c608d36fe702",
   "metadata": {},
   "source": [
    "## Part I"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd55ec0f-06f8-4ac2-aa18-1fd54c8786a3",
   "metadata": {},
   "source": [
    "### A. 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ef9c437-9071-4d59-829e-16ad683e5772",
   "metadata": {},
   "source": [
    "Can customers be segmented in a way that will allow the company to make better decisions?"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8875fce3-827c-46c1-a482-cc6e7188b831",
   "metadata": {},
   "source": [
    "### A. 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a5a00db4-61c5-4531-b20b-b0349cc1f576",
   "metadata": {},
   "source": [
    "The main goal of this analysis is to establish different grouping or segments of our current and former subscribers."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebfc8d0a-de96-4499-b38e-deeffbed0bce",
   "metadata": {},
   "source": [
    "## Part II"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ade323a-d593-4a05-99c3-7ef42cb786b5",
   "metadata": {},
   "source": [
    "### B. 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19745d4b-159a-428a-a515-e0fe2e2f8b51",
   "metadata": {},
   "source": [
    "The k-Means algorithm works by initially segregating data into groups based on the k variable. K will determine the number of groups or clusters. k-Means is also an iterative algorithm. This means that iterations of the algorithm are ran until it feels the most accurate clusters have been found.\n",
    "\n",
    "k-Means works by initially assigning data points to clusters then creating centroids, which are simply the mean of all the data in a given cluster. The number of centroids the model uses is the same as the value of k. The algorithm then runs again, but the centroids move and means are recalculated. This process runs until the centroids stabilize. (Söder, 2008)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d33cf78e-be6e-47c7-b849-353bac686906",
   "metadata": {},
   "source": [
    "### B. 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2f7949f-ad53-492a-a867-3221d65b645f",
   "metadata": {},
   "source": [
    "One main assumption of k-Means clustering is that the data must be spherical in nature. This allows the algorithm to correctly select the clusters more accurately. Graphed spherical shaped data is easier for KMeans to create clusters, but as David Robinson points out in his piece on KMeans clustering, even if the data isn't spherical in nature, not all hope is lost. (Robinson, 2015)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0d05c11d-dd92-4471-acf6-73f49ff0aeaf",
   "metadata": {},
   "source": [
    "### B. 3. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b38298c-ba47-46bc-be39-da7c1086d06a",
   "metadata": {},
   "source": [
    "Packages used:\n",
    "- Pandas - Pandas allows a wide varitey of data analytics processes, it is the main library used to import files such as csv's into a format that Python can easily handle. It is also crucial in the creation of additional dataframes.\n",
    "- SkLearn - SKlearn has a multitude of uses from being the home of the algorithm we will be using, to spliting our data into training and test sets, to also giving us our metrics that will be used to judge how well the model performed.\n",
    "- Yellowbrick - Yellowbrick is used to easily create elbow diagrams.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ca3acbfe-2abe-403e-9619-173aa7f7ab9c",
   "metadata": {},
   "source": [
    "## Part III"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16e63f14-d182-4c8a-8005-13c574b5a92e",
   "metadata": {},
   "source": [
    "### C. 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ce8c210-ab2c-41f6-a22a-980687180e24",
   "metadata": {},
   "source": [
    "One of the major things that must be done is unnessesary columns must be dropped. K-Means cannot use categorical or binary data must be dropped. This is accomplished by using the .drop() method of the pandas package."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f2cdeeb-a0f0-40b0-a3fb-64f0f7838902",
   "metadata": {},
   "source": [
    "### C. 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1fb4b8c8-5892-4051-a2e5-f18bacdcf6b2",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The features that will be included are as follows:\n",
    "- Children - Continuous\n",
    "- Age - Continuous\n",
    "- Income - Continuous\n",
    "- Outage_sec_perweek - Continuous\n",
    "- Email - Continuous\n",
    "- Contacts - Continuous\n",
    "- Yearly_equip_failure - Continuous\n",
    "- Tenure - Continuous\n",
    "- MonthlyCharge - Continuous\n",
    "- Bandwidth_GB_Year - Continuous\n",
    "\n",
    "These variables are selected due to their type of data. Binary and categorical data cannot be used in a K-Means algorithm, the majority of columns in this data set fall into this category. The continuous nature of these variables will work within the bounds of K-Means data requirements. These variables will provide our model with a wide swath of data to help create clusters."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ede9300-4289-48ea-8d50-801b837f13bd",
   "metadata": {},
   "source": [
    "### C. 3."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9b24a68",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The code block presented below does a few things. The first thing it accomplishes is dropping columns that aren't necessary to this analysis. The newly created dataframe is then presented. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a0e75000-c98b-4d1b-a1eb-748b3afdbb67",
   "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>Children</th>\n",
       "      <th>Age</th>\n",
       "      <th>Income</th>\n",
       "      <th>Outage_sec_perweek</th>\n",
       "      <th>Email</th>\n",
       "      <th>Contacts</th>\n",
       "      <th>Yearly_equip_failure</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>MonthlyCharge</th>\n",
       "      <th>Bandwidth_GB_Year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>68</td>\n",
       "      <td>28561.99000</td>\n",
       "      <td>7.97832</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6.79551</td>\n",
       "      <td>172.45552</td>\n",
       "      <td>904.53611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>27</td>\n",
       "      <td>21704.77000</td>\n",
       "      <td>11.69908</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.15668</td>\n",
       "      <td>242.63255</td>\n",
       "      <td>800.98277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>50</td>\n",
       "      <td>9609.57000</td>\n",
       "      <td>10.75280</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>15.75414</td>\n",
       "      <td>159.94758</td>\n",
       "      <td>2054.70696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>18925.23000</td>\n",
       "      <td>14.91354</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17.08723</td>\n",
       "      <td>119.95684</td>\n",
       "      <td>2164.57941</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>83</td>\n",
       "      <td>40074.19000</td>\n",
       "      <td>8.14742</td>\n",
       "      <td>16</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1.67097</td>\n",
       "      <td>149.94832</td>\n",
       "      <td>271.49344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9995</th>\n",
       "      <td>3</td>\n",
       "      <td>23</td>\n",
       "      <td>55723.74000</td>\n",
       "      <td>9.41593</td>\n",
       "      <td>12</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>68.19713</td>\n",
       "      <td>159.97940</td>\n",
       "      <td>6511.25260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9996</th>\n",
       "      <td>4</td>\n",
       "      <td>48</td>\n",
       "      <td>34129.34000</td>\n",
       "      <td>6.74055</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>61.04037</td>\n",
       "      <td>207.48110</td>\n",
       "      <td>5695.95181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9997</th>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>45983.43000</td>\n",
       "      <td>6.59091</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>47.41689</td>\n",
       "      <td>169.97410</td>\n",
       "      <td>4159.30580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9998</th>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>16667.58000</td>\n",
       "      <td>12.07191</td>\n",
       "      <td>14</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71.09560</td>\n",
       "      <td>252.62400</td>\n",
       "      <td>6468.45675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9999</th>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>9020.92000</td>\n",
       "      <td>11.75472</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>63.35086</td>\n",
       "      <td>217.48400</td>\n",
       "      <td>5857.58617</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10000 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      Children  Age      Income  Outage_sec_perweek  Email  Contacts  \\\n",
       "0            0   68 28561.99000             7.97832     10         0   \n",
       "1            1   27 21704.77000            11.69908     12         0   \n",
       "2            4   50  9609.57000            10.75280      9         0   \n",
       "3            1   48 18925.23000            14.91354     15         2   \n",
       "4            0   83 40074.19000             8.14742     16         2   \n",
       "...        ...  ...         ...                 ...    ...       ...   \n",
       "9995         3   23 55723.74000             9.41593     12         2   \n",
       "9996         4   48 34129.34000             6.74055     15         2   \n",
       "9997         1   48 45983.43000             6.59091     10         0   \n",
       "9998         1   39 16667.58000            12.07191     14         1   \n",
       "9999         1   28  9020.92000            11.75472     17         1   \n",
       "\n",
       "      Yearly_equip_failure   Tenure  MonthlyCharge  Bandwidth_GB_Year  \n",
       "0                        1  6.79551      172.45552          904.53611  \n",
       "1                        1  1.15668      242.63255          800.98277  \n",
       "2                        1 15.75414      159.94758         2054.70696  \n",
       "3                        0 17.08723      119.95684         2164.57941  \n",
       "4                        1  1.67097      149.94832          271.49344  \n",
       "...                    ...      ...            ...                ...  \n",
       "9995                     0 68.19713      159.97940         6511.25260  \n",
       "9996                     0 61.04037      207.48110         5695.95181  \n",
       "9997                     0 47.41689      169.97410         4159.30580  \n",
       "9998                     0 71.09560      252.62400         6468.45675  \n",
       "9999                     0 63.35086      217.48400         5857.58617  \n",
       "\n",
       "[10000 rows x 10 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Dropping columns that aren't required for clustering\n",
    "clean_data = data.drop(columns=['CaseOrder','Customer_id',\n",
    "                                'Interaction','Churn','UID','Population',\n",
    "                                'Zip','Job','City','Lat','Lng','State',\n",
    "                                'County','TimeZone','item1','item2',\n",
    "                                'item3','item4','item5','item6','item7',\n",
    "                                'item8','Area','Marital','Gender',\n",
    "                                'Techie','Contract','Port_modem',\n",
    "                                'Tablet','InternetService','Phone',\n",
    "                                'Multiple','OnlineSecurity',\n",
    "                                'OnlineBackup','DeviceProtection',\n",
    "                                'TechSupport','StreamingTV',\n",
    "                                'StreamingMovies','PaperlessBilling',\n",
    "                                'PaymentMethod'])\n",
    "\n",
    "### Printing the cleaned data set to ensure proper formatting\n",
    "clean_data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5aac252-673e-4bb9-bee6-890e4313dc5c",
   "metadata": {},
   "source": [
    "### C. 4."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab09a173",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The code below creates a clean_data CSV file, which will be submitted with the project."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c000c528-df0c-4102-8d10-e7fe7acd8e20",
   "metadata": {},
   "outputs": [],
   "source": [
    "clean_data.to_csv('D212 Task 1 Clean Data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60ee8daf-bebd-4e52-93c5-0d18e8737e6e",
   "metadata": {},
   "source": [
    "## Part IV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "80d969c8-86af-437a-a7e1-1f96dab01c0f",
   "metadata": {},
   "source": [
    "### D. 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd364ac8-b8a1-41f3-b643-d5a7da454a83",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The code block listed in Section D. 2. is the block where the KMeans clustering takes place. The first section creates a new dataframe with original data to add the clusters the algorithm picks. This section is also where the data is scaled.\n",
    "\n",
    "The second section uses Yellowbrick to create an elbow plot to assist with the selection of k.\n",
    "\n",
    "The third section is where KMeans clustering takes place and the cluster chosen for each row is added in new column. The cleaned_data dataframe is also printed for verification in this section."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a53402e-f9ff-4d44-bb25-80f1d447750c",
   "metadata": {},
   "source": [
    "### D. 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "632b9560-2649-4276-ad14-5d16cb5bdecc",
   "metadata": {},
   "outputs": [],
   "source": [
    "### Creating a new dataframe to store the KMeans results with the original non-scaled data\n",
    "features = list(clean_data.columns)\n",
    "cleaned_data = clean_data[features]\n",
    "\n",
    "### Creating a scaler object and scaling the data to make it more uniformed\n",
    "scaler = StandardScaler()\n",
    "scaled_data = scaler.fit_transform(clean_data)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "504d98a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Distortion Score Elbow for KMeans Clustering'}, xlabel='k', ylabel='distortion score'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Using Yellowbrick to quickly create an elbow plot to fit the optimum value for k\n",
    "model = KMeans()\n",
    "visual = KElbowVisualizer(model, k=(1, 20))\n",
    "visual.fit(scaled_data)\n",
    "visual.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ffc2ec70",
   "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>Children</th>\n",
       "      <th>Age</th>\n",
       "      <th>Income</th>\n",
       "      <th>Outage_sec_perweek</th>\n",
       "      <th>Email</th>\n",
       "      <th>Contacts</th>\n",
       "      <th>Yearly_equip_failure</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>MonthlyCharge</th>\n",
       "      <th>Bandwidth_GB_Year</th>\n",
       "      <th>clusters</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>68</td>\n",
       "      <td>28561.99000</td>\n",
       "      <td>7.97832</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6.79551</td>\n",
       "      <td>172.45552</td>\n",
       "      <td>904.53611</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>27</td>\n",
       "      <td>21704.77000</td>\n",
       "      <td>11.69908</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.15668</td>\n",
       "      <td>242.63255</td>\n",
       "      <td>800.98277</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>50</td>\n",
       "      <td>9609.57000</td>\n",
       "      <td>10.75280</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>15.75414</td>\n",
       "      <td>159.94758</td>\n",
       "      <td>2054.70696</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>18925.23000</td>\n",
       "      <td>14.91354</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17.08723</td>\n",
       "      <td>119.95684</td>\n",
       "      <td>2164.57941</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>83</td>\n",
       "      <td>40074.19000</td>\n",
       "      <td>8.14742</td>\n",
       "      <td>16</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1.67097</td>\n",
       "      <td>149.94832</td>\n",
       "      <td>271.49344</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9995</th>\n",
       "      <td>3</td>\n",
       "      <td>23</td>\n",
       "      <td>55723.74000</td>\n",
       "      <td>9.41593</td>\n",
       "      <td>12</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>68.19713</td>\n",
       "      <td>159.97940</td>\n",
       "      <td>6511.25260</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9996</th>\n",
       "      <td>4</td>\n",
       "      <td>48</td>\n",
       "      <td>34129.34000</td>\n",
       "      <td>6.74055</td>\n",
       "      <td>15</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>61.04037</td>\n",
       "      <td>207.48110</td>\n",
       "      <td>5695.95181</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9997</th>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>45983.43000</td>\n",
       "      <td>6.59091</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>47.41689</td>\n",
       "      <td>169.97410</td>\n",
       "      <td>4159.30580</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9998</th>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>16667.58000</td>\n",
       "      <td>12.07191</td>\n",
       "      <td>14</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71.09560</td>\n",
       "      <td>252.62400</td>\n",
       "      <td>6468.45675</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9999</th>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>9020.92000</td>\n",
       "      <td>11.75472</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>63.35086</td>\n",
       "      <td>217.48400</td>\n",
       "      <td>5857.58617</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10000 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      Children  Age      Income  Outage_sec_perweek  Email  Contacts  \\\n",
       "0            0   68 28561.99000             7.97832     10         0   \n",
       "1            1   27 21704.77000            11.69908     12         0   \n",
       "2            4   50  9609.57000            10.75280      9         0   \n",
       "3            1   48 18925.23000            14.91354     15         2   \n",
       "4            0   83 40074.19000             8.14742     16         2   \n",
       "...        ...  ...         ...                 ...    ...       ...   \n",
       "9995         3   23 55723.74000             9.41593     12         2   \n",
       "9996         4   48 34129.34000             6.74055     15         2   \n",
       "9997         1   48 45983.43000             6.59091     10         0   \n",
       "9998         1   39 16667.58000            12.07191     14         1   \n",
       "9999         1   28  9020.92000            11.75472     17         1   \n",
       "\n",
       "      Yearly_equip_failure   Tenure  MonthlyCharge  Bandwidth_GB_Year  \\\n",
       "0                        1  6.79551      172.45552          904.53611   \n",
       "1                        1  1.15668      242.63255          800.98277   \n",
       "2                        1 15.75414      159.94758         2054.70696   \n",
       "3                        0 17.08723      119.95684         2164.57941   \n",
       "4                        1  1.67097      149.94832          271.49344   \n",
       "...                    ...      ...            ...                ...   \n",
       "9995                     0 68.19713      159.97940         6511.25260   \n",
       "9996                     0 61.04037      207.48110         5695.95181   \n",
       "9997                     0 47.41689      169.97410         4159.30580   \n",
       "9998                     0 71.09560      252.62400         6468.45675   \n",
       "9999                     0 63.35086      217.48400         5857.58617   \n",
       "\n",
       "      clusters  \n",
       "0            0  \n",
       "1            0  \n",
       "2            0  \n",
       "3            2  \n",
       "4            0  \n",
       "...        ...  \n",
       "9995         1  \n",
       "9996         1  \n",
       "9997         1  \n",
       "9998         1  \n",
       "9999         1  \n",
       "\n",
       "[10000 rows x 11 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Performing the KMeans clustering process and affixing the clusters to the cleaned_data dataframe\n",
    "clusters = KMeans(n_clusters=6)\n",
    "cleaned_data['clusters'] = clusters.fit_predict(scaled_data)\n",
    "\n",
    "### Printing the dataframe to ensure clusters were affixed correctly\n",
    "cleaned_data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a4cbf24-8074-41cd-89dd-e5a9ab71545f",
   "metadata": {},
   "source": [
    "## Part V"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9eb24857-016c-4ed3-a5a8-322405566047",
   "metadata": {},
   "source": [
    "### E. 1."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c1e180a8-42b5-4646-b89c-07dd0cd71adc",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The use of a clustering algorithm doesn't not return accuracy in a true sense. While accuracy can be returned, in this instance, it isn't viable data. The data points aren't clustered before clustering occurs, thus, giving nothing to compare it to."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a832af6-e7c6-4305-82ed-9d905107b168",
   "metadata": {},
   "source": [
    "### E. 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ef41c1d-3252-43e8-80e1-90ccf0c31526",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "Using KMeans we were able to find that our data set can be divided into 6 different clusters. These clusters are based on our variety of available data. The implications of this analysis could prove to be important, this gives the company 6 distinct groups to target with a different offers, sales, or promotions that will feel more customized to them. While clustering provides a deeper understanding of the data, without a more thorough investigation into the clusters and what they actually represent, you have just that, clusters."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b57ac3d7-d0ff-44d9-8310-a5a712251dea",
   "metadata": {},
   "source": [
    "### E. 3."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f290c37a-e64a-4807-a3ab-ca22da62d413",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "One limitation that makes this data set hard to work with is the shear number of features. Without using some form of dimensional reduction the set is hard to handle, and it also makes using KMeans harder as well. Being able to apply dimensional reduction to this set would possibly return better, more clear results."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9db469c-8d34-4bf5-b93d-eed694c39720",
   "metadata": {},
   "source": [
    "### E. 4."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0647e9f8-9a71-45a3-baaf-74ee0392c737",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "The best course of action moving forward would be applying some form of dimensional reduction, then analyzing the data for different segments using the resulting data set. The company could make more sense of the data with only a handful of features rather than the entire set of fifty features."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2deeb065",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "### Part VI"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7272a2e",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "### H: References"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49a21cae",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "Robinson, D. (2015, January 15). K-means clustering is not a free lunch. Variance Explained. Retrieved October 1, 2021, from http://varianceexplained.org/r/kmeans-free-lunch/.\n",
    "\n",
    "Söder, O. (2008, May 29). k-means clustering 1. How does k-means clustering work? K-means clustering 1. how does K-means clustering work? Retrieved October 1, 2021, from https://www.fon.hum.uva.nl/praat/manual/k-means_clustering_1__How_does_k-means_clustering_work_.html.\n",
    "\n",
    "### Coding Resources Used\n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html\n",
    "\n",
    "https://www.scikit-yb.org/en/latest/api/cluster/elbow.html"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}