[분석실무]4.고객의 행동 예측.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 고객 행동 예측\n",
    "* 군집화"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. 데이터 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "log_id         0\n",
       "customer_id    0\n",
       "usedate        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "uselog = pd.read_csv('use_log.csv')\n",
    "uselog.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_id             0\n",
       "name                    0\n",
       "class                   0\n",
       "gender                  0\n",
       "start_date              0\n",
       "end_date             2842\n",
       "campaign_id             0\n",
       "is_deleted              0\n",
       "class_name              0\n",
       "price                   0\n",
       "campaign_name           0\n",
       "mean                    0\n",
       "median                  0\n",
       "max                     0\n",
       "min                     0\n",
       "routine_flg             0\n",
       "calc_date               0\n",
       "membership_period       0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer = pd.read_csv('customer_join.csv')\n",
    "customer.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 군집화를 통한 회원 그룹화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>membership_period</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.083333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.583333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>4.5</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.916667</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       mean  median  max  min  membership_period\n",
       "0  4.833333     5.0    8    2                 47\n",
       "1  5.083333     5.0    7    3                 47\n",
       "2  4.583333     5.0    6    3                 47\n",
       "3  4.833333     4.5    7    2                 47\n",
       "4  3.916667     4.0    6    1                 47"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering = customer[['mean','median','max','min','membership_period']]\n",
    "customer_clustering.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* 'mean','median','max','min','membership_period'의 각 데이터 크기 차이가 크므로 표준화가 필요함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3 1 0 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:9: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  if __name__ == '__main__':\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>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>membership_period</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.083333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.583333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>4.5</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.916667</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       mean  median  max  min  membership_period  cluster\n",
       "0  4.833333     5.0    8    2                 47        3\n",
       "1  5.083333     5.0    7    3                 47        3\n",
       "2  4.583333     5.0    6    3                 47        3\n",
       "3  4.833333     4.5    7    2                 47        3\n",
       "4  3.916667     4.0    6    1                 47        3"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "sc = StandardScaler()\n",
    "customer_clustering_sc = sc.fit_transform(customer_clustering)\n",
    "\n",
    "kmeans = KMeans(n_clusters=4, random_state=0)\n",
    "clusters = kmeans.fit(customer_clustering_sc)\n",
    "customer_clustering['cluster'] = clusters.labels_\n",
    "print(customer_clustering['cluster'].unique())\n",
    "customer_clustering.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. 군집화 결과 분석"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>월평균값</th>\n",
       "      <th>월중앙값</th>\n",
       "      <th>월최대값</th>\n",
       "      <th>월최소값</th>\n",
       "      <th>회원기간</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cluster</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>0</th>\n",
       "      <td>1334</td>\n",
       "      <td>1334</td>\n",
       "      <td>1334</td>\n",
       "      <td>1334</td>\n",
       "      <td>1334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>763</td>\n",
       "      <td>763</td>\n",
       "      <td>763</td>\n",
       "      <td>763</td>\n",
       "      <td>763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>846</td>\n",
       "      <td>846</td>\n",
       "      <td>846</td>\n",
       "      <td>846</td>\n",
       "      <td>846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1249</td>\n",
       "      <td>1249</td>\n",
       "      <td>1249</td>\n",
       "      <td>1249</td>\n",
       "      <td>1249</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         월평균값  월중앙값  월최대값  월최소값  회원기간\n",
       "cluster                              \n",
       "0        1334  1334  1334  1334  1334\n",
       "1         763   763   763   763   763\n",
       "2         846   846   846   846   846\n",
       "3        1249  1249  1249  1249  1249"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering.columns = ['월평균값','월중앙값','월최대값','월최소값','회원기간','cluster']\n",
    "customer_clustering.groupby('cluster').count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>월평균값</th>\n",
       "      <th>월중앙값</th>\n",
       "      <th>월최대값</th>\n",
       "      <th>월최소값</th>\n",
       "      <th>회원기간</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cluster</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>0</th>\n",
       "      <td>5.524239</td>\n",
       "      <td>5.375187</td>\n",
       "      <td>8.745877</td>\n",
       "      <td>2.687406</td>\n",
       "      <td>14.843328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.054713</td>\n",
       "      <td>2.888598</td>\n",
       "      <td>4.756225</td>\n",
       "      <td>1.653997</td>\n",
       "      <td>9.263434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8.054608</td>\n",
       "      <td>8.039598</td>\n",
       "      <td>10.009456</td>\n",
       "      <td>6.160757</td>\n",
       "      <td>7.072104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.677494</td>\n",
       "      <td>4.671337</td>\n",
       "      <td>7.232986</td>\n",
       "      <td>2.153723</td>\n",
       "      <td>36.915933</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             월평균값      월중앙값       월최대값      월최소값       회원기간\n",
       "cluster                                                    \n",
       "0        5.524239  5.375187   8.745877  2.687406  14.843328\n",
       "1        3.054713  2.888598   4.756225  1.653997   9.263434\n",
       "2        8.054608  8.039598  10.009456  6.160757   7.072104\n",
       "3        4.677494  4.671337   7.232986  2.153723  36.915933"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering.groupby('cluster').mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. 시각화\n",
    "* 주성분분석"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "\n",
    "X = customer_clustering_sc\n",
    "pca = PCA(n_components=2)\n",
    "pca.fit(X)\n",
    "x_pca = pca.transform(X)\n",
    "pca_df = pd.DataFrame(x_pca)\n",
    "pca_df['cluster'] = customer_clustering['cluster']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "for i in customer_clustering['cluster'].unique():\n",
    "    tmp = pca_df.loc[pca_df['cluster']==i]\n",
    "    plt.scatter(tmp[0], tmp[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. 군집화 결과로 탈퇴 회원의 경향 파악\n",
    "* is_deleted열을 customer_clustering에 추가, cluster 및 is_deleted별로 집계"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>월평균값</th>\n",
       "      <th>월중앙값</th>\n",
       "      <th>월최대값</th>\n",
       "      <th>월최소값</th>\n",
       "      <th>회원기간</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.083333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.583333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.833333</td>\n",
       "      <td>4.5</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.916667</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       월평균값  월중앙값  월최대값  월최소값  회원기간  cluster\n",
       "0  4.833333   5.0     8     2    47        3\n",
       "1  5.083333   5.0     7     3    47        3\n",
       "2  4.583333   5.0     6     3    47        3\n",
       "3  4.833333   4.5     7     2    47        3\n",
       "4  3.916667   4.0     6     1    47        3"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>customer_id</th>\n",
       "      <th>name</th>\n",
       "      <th>class</th>\n",
       "      <th>gender</th>\n",
       "      <th>start_date</th>\n",
       "      <th>end_date</th>\n",
       "      <th>campaign_id</th>\n",
       "      <th>is_deleted</th>\n",
       "      <th>class_name</th>\n",
       "      <th>price</th>\n",
       "      <th>campaign_name</th>\n",
       "      <th>mean</th>\n",
       "      <th>median</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>routine_flg</th>\n",
       "      <th>calc_date</th>\n",
       "      <th>membership_period</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>OA832399</td>\n",
       "      <td>XXXX</td>\n",
       "      <td>C01</td>\n",
       "      <td>F</td>\n",
       "      <td>2015-05-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CA1</td>\n",
       "      <td>0</td>\n",
       "      <td>0_종일</td>\n",
       "      <td>10500</td>\n",
       "      <td>2_일반</td>\n",
       "      <td>4.833333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2019-04-30</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PL270116</td>\n",
       "      <td>XXXXX</td>\n",
       "      <td>C01</td>\n",
       "      <td>M</td>\n",
       "      <td>2015-05-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CA1</td>\n",
       "      <td>0</td>\n",
       "      <td>0_종일</td>\n",
       "      <td>10500</td>\n",
       "      <td>2_일반</td>\n",
       "      <td>5.083333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2019-04-30</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>OA974876</td>\n",
       "      <td>XXXXX</td>\n",
       "      <td>C01</td>\n",
       "      <td>M</td>\n",
       "      <td>2015-05-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CA1</td>\n",
       "      <td>0</td>\n",
       "      <td>0_종일</td>\n",
       "      <td>10500</td>\n",
       "      <td>2_일반</td>\n",
       "      <td>4.583333</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2019-04-30</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HD024127</td>\n",
       "      <td>XXXXX</td>\n",
       "      <td>C01</td>\n",
       "      <td>F</td>\n",
       "      <td>2015-05-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CA1</td>\n",
       "      <td>0</td>\n",
       "      <td>0_종일</td>\n",
       "      <td>10500</td>\n",
       "      <td>2_일반</td>\n",
       "      <td>4.833333</td>\n",
       "      <td>4.5</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2019-04-30</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HD661448</td>\n",
       "      <td>XXXXX</td>\n",
       "      <td>C03</td>\n",
       "      <td>F</td>\n",
       "      <td>2015-05-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CA1</td>\n",
       "      <td>0</td>\n",
       "      <td>2_야간</td>\n",
       "      <td>6000</td>\n",
       "      <td>2_일반</td>\n",
       "      <td>3.916667</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2019-04-30</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  customer_id   name class gender  start_date end_date campaign_id  \\\n",
       "0    OA832399   XXXX   C01      F  2015-05-01      NaN         CA1   \n",
       "1    PL270116  XXXXX   C01      M  2015-05-01      NaN         CA1   \n",
       "2    OA974876  XXXXX   C01      M  2015-05-01      NaN         CA1   \n",
       "3    HD024127  XXXXX   C01      F  2015-05-01      NaN         CA1   \n",
       "4    HD661448  XXXXX   C03      F  2015-05-01      NaN         CA1   \n",
       "\n",
       "   is_deleted class_name  price campaign_name      mean  median  max  min  \\\n",
       "0           0       0_종일  10500          2_일반  4.833333     5.0    8    2   \n",
       "1           0       0_종일  10500          2_일반  5.083333     5.0    7    3   \n",
       "2           0       0_종일  10500          2_일반  4.583333     5.0    6    3   \n",
       "3           0       0_종일  10500          2_일반  4.833333     4.5    7    2   \n",
       "4           0       2_야간   6000          2_일반  3.916667     4.0    6    1   \n",
       "\n",
       "   routine_flg   calc_date  membership_period  \n",
       "0            1  2019-04-30                 47  \n",
       "1            1  2019-04-30                 47  \n",
       "2            1  2019-04-30                 47  \n",
       "3            1  2019-04-30                 47  \n",
       "4            1  2019-04-30                 47  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>cluster</th>\n",
       "      <th>is_deleted</th>\n",
       "      <th>customer_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   cluster  is_deleted  customer_id\n",
       "0        0           0          785\n",
       "1        0           1          549\n",
       "2        1           1          763\n",
       "3        2           0          827\n",
       "4        2           1           19\n",
       "5        3           0         1230\n",
       "6        3           1           19"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering = pd.concat([customer_clustering, customer], axis=1)\n",
    "customer_clustering.groupby(['cluster', 'is_deleted'], as_index=False).count()[['cluster', 'is_deleted','customer_id']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>cluster</th>\n",
       "      <th>routine_flg</th>\n",
       "      <th>customer_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1247</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   cluster  routine_flg  customer_id\n",
       "0        0            0          228\n",
       "1        0            1         1106\n",
       "2        1            0          497\n",
       "3        1            1          266\n",
       "4        2            0           52\n",
       "5        2            1          794\n",
       "6        3            0            2\n",
       "7        3            1         1247"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_clustering.groupby(['cluster', 'routine_flg'], as_index=False).count()[['cluster', 'routine_flg','customer_id']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. 다음 달의 이용 횟수 예측을 위한 데이터 준비\n",
    "### A. uselog 데이터를 이용해 연월, 회원마다 집계"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>연월</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201804</td>\n",
       "      <td>AS002855</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201804</td>\n",
       "      <td>AS009013</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>201804</td>\n",
       "      <td>AS009373</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>201804</td>\n",
       "      <td>AS015315</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>201804</td>\n",
       "      <td>AS015739</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       연월 customer_id  count\n",
       "0  201804    AS002855      4\n",
       "1  201804    AS009013      2\n",
       "2  201804    AS009373      3\n",
       "3  201804    AS015315      6\n",
       "4  201804    AS015739      7"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "uselog['usedate'] = pd.to_datetime(uselog['usedate'])\n",
    "uselog['연월'] = uselog['usedate'].dt.strftime('%Y%m')\n",
    "uselog_months = uselog.groupby(['연월', 'customer_id'], as_index=False).count()\n",
    "uselog_months.rename(columns={'log_id' : 'count'}, inplace=True)\n",
    "\n",
    "del uselog_months['usedate']\n",
    "uselog_months.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### B. 2018년 10월부터 2019년 3월까지 과거 5개월분 이용횟수와 다음달의 이용횟수 저장"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda3\\lib\\site-packages\\pandas\\core\\frame.py:4133: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  errors=errors,\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>연월</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>count_pred</th>\n",
       "      <th>count_0</th>\n",
       "      <th>count_1</th>\n",
       "      <th>count_2</th>\n",
       "      <th>count_3</th>\n",
       "      <th>count_4</th>\n",
       "      <th>count_5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS002855</td>\n",
       "      <td>3</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS008805</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS009373</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015233</td>\n",
       "      <td>7</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015315</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       연월 customer_id  count_pred  count_0  count_1  count_2  count_3  \\\n",
       "0  201810    AS002855           3      7.0      3.0      5.0      5.0   \n",
       "1  201810    AS008805           2      2.0      5.0      7.0      8.0   \n",
       "2  201810    AS009373           5      6.0      6.0      7.0      4.0   \n",
       "3  201810    AS015233           7      9.0     11.0      5.0      7.0   \n",
       "4  201810    AS015315           4      7.0      3.0      6.0      3.0   \n",
       "\n",
       "   count_4  count_5  \n",
       "0      5.0      4.0  \n",
       "1      NaN      NaN  \n",
       "2      4.0      3.0  \n",
       "3      7.0      NaN  \n",
       "4      3.0      6.0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "year_months = list(uselog_months['연월'].unique()) # 201804 ~ 201903\n",
    "predict_data = pd.DataFrame()\n",
    "\n",
    "for i in range(6, len(year_months)): # 201810 ~ 201903\n",
    "    tmp = uselog_months.loc[uselog_months['연월']==year_months[i]]\n",
    "    tmp.rename(columns={'count' : 'count_pred'}, inplace=True)\n",
    "    for j in range(1,7):\n",
    "        tmp_before = uselog_months.loc[uselog_months['연월'] == year_months[i-j]]\n",
    "        del tmp_before['연월']\n",
    "        tmp_before.rename(columns={'count' : 'count_{}'.format(j-1)}, inplace=True)\n",
    "        tmp = pd.merge(tmp, tmp_before, on='customer_id', how='left')\n",
    "    predict_data = pd.concat([predict_data, tmp], ignore_index=True)\n",
    "predict_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>연월</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>count_pred</th>\n",
       "      <th>count_0</th>\n",
       "      <th>count_1</th>\n",
       "      <th>count_2</th>\n",
       "      <th>count_3</th>\n",
       "      <th>count_4</th>\n",
       "      <th>count_5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS002855</td>\n",
       "      <td>3</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS009373</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015315</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015739</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS019860</td>\n",
       "      <td>7</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       연월 customer_id  count_pred  count_0  count_1  count_2  count_3  \\\n",
       "0  201810    AS002855           3      7.0      3.0      5.0      5.0   \n",
       "1  201810    AS009373           5      6.0      6.0      7.0      4.0   \n",
       "2  201810    AS015315           4      7.0      3.0      6.0      3.0   \n",
       "3  201810    AS015739           5      6.0      5.0      8.0      6.0   \n",
       "4  201810    AS019860           7      5.0      7.0      4.0      6.0   \n",
       "\n",
       "   count_4  count_5  \n",
       "0      5.0      4.0  \n",
       "1      4.0      3.0  \n",
       "2      3.0      6.0  \n",
       "3      5.0      7.0  \n",
       "4      8.0      6.0  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_data = predict_data.dropna()\n",
    "predict_data = predict_data.reset_index(drop=True)\n",
    "predict_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. 독립변수 추가"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>연월</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>count_pred</th>\n",
       "      <th>count_0</th>\n",
       "      <th>count_1</th>\n",
       "      <th>count_2</th>\n",
       "      <th>count_3</th>\n",
       "      <th>count_4</th>\n",
       "      <th>count_5</th>\n",
       "      <th>start_date</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS002855</td>\n",
       "      <td>3</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2016-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS009373</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2015-11-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015315</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2015-07-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015739</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2017-06-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS019860</td>\n",
       "      <td>7</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2017-10-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       연월 customer_id  count_pred  count_0  count_1  count_2  count_3  \\\n",
       "0  201810    AS002855           3      7.0      3.0      5.0      5.0   \n",
       "1  201810    AS009373           5      6.0      6.0      7.0      4.0   \n",
       "2  201810    AS015315           4      7.0      3.0      6.0      3.0   \n",
       "3  201810    AS015739           5      6.0      5.0      8.0      6.0   \n",
       "4  201810    AS019860           7      5.0      7.0      4.0      6.0   \n",
       "\n",
       "   count_4  count_5  start_date  \n",
       "0      5.0      4.0  2016-11-01  \n",
       "1      4.0      3.0  2015-11-01  \n",
       "2      3.0      6.0  2015-07-01  \n",
       "3      5.0      7.0  2017-06-01  \n",
       "4      8.0      6.0  2017-10-01  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_data = pd.merge(predict_data, customer[['customer_id', 'start_date']], on='customer_id', how='left')\n",
    "predict_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:9: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  if __name__ == '__main__':\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>연월</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>count_pred</th>\n",
       "      <th>count_0</th>\n",
       "      <th>count_1</th>\n",
       "      <th>count_2</th>\n",
       "      <th>count_3</th>\n",
       "      <th>count_4</th>\n",
       "      <th>count_5</th>\n",
       "      <th>start_date</th>\n",
       "      <th>now_date</th>\n",
       "      <th>period</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS002855</td>\n",
       "      <td>3</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2016-11-01</td>\n",
       "      <td>2018-10-01</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS009373</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2015-11-01</td>\n",
       "      <td>2018-10-01</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015315</td>\n",
       "      <td>4</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2015-07-01</td>\n",
       "      <td>2018-10-01</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS015739</td>\n",
       "      <td>5</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2017-06-01</td>\n",
       "      <td>2018-10-01</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>201810</td>\n",
       "      <td>AS019860</td>\n",
       "      <td>7</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2017-10-01</td>\n",
       "      <td>2018-10-01</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       연월 customer_id  count_pred  count_0  count_1  count_2  count_3  \\\n",
       "0  201810    AS002855           3      7.0      3.0      5.0      5.0   \n",
       "1  201810    AS009373           5      6.0      6.0      7.0      4.0   \n",
       "2  201810    AS015315           4      7.0      3.0      6.0      3.0   \n",
       "3  201810    AS015739           5      6.0      5.0      8.0      6.0   \n",
       "4  201810    AS019860           7      5.0      7.0      4.0      6.0   \n",
       "\n",
       "   count_4  count_5 start_date   now_date period  \n",
       "0      5.0      4.0 2016-11-01 2018-10-01     23  \n",
       "1      4.0      3.0 2015-11-01 2018-10-01     35  \n",
       "2      3.0      6.0 2015-07-01 2018-10-01     39  \n",
       "3      5.0      7.0 2017-06-01 2018-10-01     16  \n",
       "4      8.0      6.0 2017-10-01 2018-10-01     12  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_data['now_date'] = pd.to_datetime(predict_data['연월'], format='%Y%m')\n",
    "predict_data['start_date'] = pd.to_datetime(predict_data['start_date'])\n",
    "\n",
    "from dateutil.relativedelta import relativedelta\n",
    "\n",
    "predict_data['period'] = None\n",
    "for i in range(len(predict_data)):\n",
    "    delta = relativedelta(predict_data['now_date'][i], predict_data['start_date'][i])\n",
    "    predict_data['period'][i] = delta.years*12 + delta.months\n",
    "predict_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. 회귀 모델"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import linear_model\n",
    "import sklearn.model_selection\n",
    "\n",
    "predict_data = predict_data.loc[predict_data['start_date'] >= pd.to_datetime('20180401')]\n",
    "\n",
    "model = linear_model.LinearRegression()\n",
    "X = predict_data[['count_0','count_1','count_2','count_3','count_4','count_5','period']]\n",
    "y = predict_data['count_pred']\n",
    "X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X,y)\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6028092457092975\n",
      "0.6225218765461253\n"
     ]
    }
   ],
   "source": [
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9. 독립변수별 계수 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>feature_names</th>\n",
       "      <th>coefficient</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>count_0</td>\n",
       "      <td>0.328388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>count_1</td>\n",
       "      <td>0.186886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>count_2</td>\n",
       "      <td>0.187746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>count_3</td>\n",
       "      <td>0.181768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>count_4</td>\n",
       "      <td>0.058669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>count_5</td>\n",
       "      <td>0.075501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>period</td>\n",
       "      <td>0.089578</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  feature_names  coefficient\n",
       "0       count_0     0.328388\n",
       "1       count_1     0.186886\n",
       "2       count_2     0.187746\n",
       "3       count_3     0.181768\n",
       "4       count_4     0.058669\n",
       "5       count_5     0.075501\n",
       "6        period     0.089578"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef = pd.DataFrame({'feature_names' : X.columns, 'coefficient' : model.coef_})\n",
    "coef"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 10. 다음 달의 이용 횟수 예측"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3.82796182, 2.0825743 ])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x1 = [3,4,4,6,8,7,8]\n",
    "x2 = [2,2,3,3,4,6,8]\n",
    "x_pred = [x1, x2]\n",
    "\n",
    "model.predict(x_pred)"
   ]
  },
  {
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}