BP_data_analysis.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "BP_data_analysis.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/bvm84/open_colab_ml/blob/master/BP_data_analysis.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rNXX8ZU63fuD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from IPython.display import clear_output\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "import sys\n",
        "import os\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import feather\n",
        "import logging\n",
        "from pathlib import PurePath\n",
        "from pandas import DataFrame\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.feature_selection import SelectKBest\n",
        "from sklearn.feature_selection import chi2, f_regression\n",
        "from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier\n",
        "from sklearn.linear_model import LinearRegression\n",
        "sys.path.append('/content/drive/My Drive/Colab/l2_ML/')\n",
        "% matplotlib inline\n",
        "np.random.seed(0)\n",
        "clear_output()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9BmGQ7tAHu9e",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Db():\n",
        "\n",
        "    def __init__(self):\n",
        "        self._df = None\n",
        "\n",
        "    @staticmethod\n",
        "    def create_df(wav_folder, l2_folder, pid, features_list):\n",
        "        allowed_features = ('HR', 'MAB2', 'MAD3P', 'MADP', 'MASEP', 'MASEPMax', 'MASRP', 'MB0',\n",
        "                            'MB1', 'MB2', 'MD3P', 'MDP', 'MFDP', 'MFSEP', 'MFSRP', 'MPI', 'MRR',\n",
        "                            'MS4S1', 'MS5S2', 'MSEP', 'MSEPMax', 'MSNR', 'MSRP', 'MWaveType',\n",
        "                            'MdD3P', 'MdDP', 'MdSEP', 'MdSEPMax', 'MdSRP', 'BR')\n",
        "        for feature in features_list:\n",
        "            if feature not in allowed_features:\n",
        "                raise ValueError(\"Unknown feature: {0}, expected one of {1}\".format(features_list, allowed_features))\n",
        "\n",
        "        pid_cid_sys_dia_features = []\n",
        "        wav_glob_ext = '*.wav'\n",
        "        for filename in glob.iglob(os.path.join(wav_folder, '') + wav_glob_ext, recursive=True):\n",
        "            pid = pid\n",
        "            cid = os.path.splitext(PurePath(filename).parts[-1])[0].split('_')[0]\n",
        "            sys = int(os.path.splitext(PurePath(filename).parts[-1])[0].split('_')[1])\n",
        "            dia = int(os.path.splitext(PurePath(filename).parts[-1])[0].split('_')[2])\n",
        "            json_filename = l2_folder.joinpath(cid).with_suffix('.json')\n",
        "            with open(json_filename, 'r') as json_file:\n",
        "                json_data = json.load(json_file)\n",
        "            # pprint.pprint(json_data)\n",
        "            features = []\n",
        "            for feature in features_list:\n",
        "                try:\n",
        "                    features.append(json_data['value'][feature])\n",
        "                except KeyError:\n",
        "                    features.append(np.nan)\n",
        "            pid_cid_sys_dia_features.append([pid, cid, sys, dia] + features)\n",
        "            # print(cid_sys_dia_features)\n",
        "        df = DataFrame(pid_cid_sys_dia_features,\n",
        "                       columns=['pid', 'cid', 'sys', 'dia'] + features_list)\n",
        "        return df\n",
        "\n",
        "    @staticmethod\n",
        "    def get_pid_dirs(db_folder):\n",
        "        pid_dirs = []\n",
        "        try:\n",
        "            for item in db_folder.iterdir():\n",
        "                # print(item)\n",
        "                if item.is_dir():\n",
        "                    pid_dirs.append(item)\n",
        "            return pid_dirs\n",
        "        except FileNotFoundError:\n",
        "            print('Invalid directory')\n",
        "\n",
        "    def create_dataframe(self, db_folder, features_list):\n",
        "        dfs = []\n",
        "        result_df = DataFrame()\n",
        "        for item in self.get_pid_dirs(db_folder):\n",
        "            wav_folder = item.joinpath('wav_out')\n",
        "            l2_folder = item.joinpath('l2')\n",
        "            pid = item.parts[-1]\n",
        "            dfs.append(self.create_df(wav_folder, l2_folder, pid, features_list))\n",
        "        result_df = pd.concat(dfs)\n",
        "        self._df = result_df\n",
        "\n",
        "    @property\n",
        "    def df(self):\n",
        "        return self._df\n",
        "\n",
        "    @df.setter\n",
        "    def set_df(self, df):\n",
        "        if isinstance(df, pd.DataFrame):\n",
        "            self._df = df\n",
        "        else:\n",
        "            print('Wrong dataframe')\n",
        "\n",
        "    def save_dataframe(self, df_filename):\n",
        "        try:\n",
        "            feather.write_dataframe(self._df, df_filename)\n",
        "        except Exception as e:\n",
        "            logging.error(e)\n",
        "\n",
        "    def read_dataframe(self, df_filename):\n",
        "        try:\n",
        "            self._df = feather.read_dataframe(df_filename)\n",
        "        except Exception as e:\n",
        "            logging.error(e)\n",
        "\n",
        "    def get_pid_ncid_dict(self, threshold_n=20):\n",
        "        pid_list = self._df['pid'].unique()\n",
        "        print(pid_list)\n",
        "        cid_count_list = []\n",
        "        for pid_item in pid_list:\n",
        "            temp_df = self._df.query('pid == @pid_item')\n",
        "            cid_count_list.append(len(temp_df))\n",
        "        pid_ncid_dict = {pid_list[i]: cid_count_list[i] for i in range(len(pid_list))}\n",
        "        print(pid_ncid_dict)\n",
        "        keys_to_delete = []\n",
        "        for key, value in pid_ncid_dict.items():\n",
        "            if value < threshold_n:\n",
        "                keys_to_delete.append(key)\n",
        "        for key in keys_to_delete:\n",
        "            pid_ncid_dict.pop(key)\n",
        "        print(pid_ncid_dict)\n",
        "        return pid_ncid_dict\n",
        "\n",
        "    def truncate_df(self, cid_quantity_tostore=20, do_filter=True):\n",
        "        if do_filter:\n",
        "            indexes_to_drop = []\n",
        "        pid_ncid_dict = self.get_pid_ncid_dict()\n",
        "        for key in pid_ncid_dict:\n",
        "            temp_df = self._df.query('pid == @key')\n",
        "            if len(temp_df) > cid_quantity_tostore:\n",
        "                remove_n = len(temp_df) - cid_quantity_tostore\n",
        "                drop_indexes = np.random.choice(temp_df.index, remove_n, replace=False)\n",
        "                indexes_to_drop.append(drop_indexes)\n",
        "        # print(len(np.concatenate(indexes_to_drop)))\n",
        "        cleared_df = self._df.drop(np.concatenate(indexes_to_drop))\n",
        "        print(len(pid_ncid_dict))\n",
        "        print(cleared_df)\n",
        "\n",
        "    def get_pid_df(self, pid):\n",
        "        if pid not in self._df['pid'].unique():\n",
        "            raise ValueError(\"No such pid in db: {0}, expected one of {1}\".format(pid, self._df['pid'].unique()))\n",
        "        else:\n",
        "            return self._df.query('pid == @pid')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p_I76i6U532v",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 263
        },
        "outputId": "9b4e8edf-0301-46e8-89e2-203801cad6ad"
      },
      "source": [
        "dataframe_filename = os.path.join('/content/drive/My Drive/Colab/l2_ML/all_json_df.file')\n",
        "inst = Db()\n",
        "inst.read_dataframe(dataframe_filename)\n",
        "df = inst.df\n",
        "print(df.columns)\n",
        "print(df.head())\n",
        "# df.describe()\n",
        "df.drop(['MB1'], axis=1, inplace=True)\n",
        "df.dropna(axis=0, how='any', inplace=True)\n",
        "# df.describe()\n",
        "index_0s = df.query('HR ==0').index\n",
        "index_dia_min = df.query('dia < 40').index\n",
        "index_dia_max = df.query('dia > 120').index\n",
        "df.drop(index_0s, axis=0, inplace=True)\n",
        "df.drop(index_dia_min, axis=0, inplace=True)\n",
        "df.drop(index_dia_max, axis=0, inplace=True)"
      ],
      "execution_count": 130,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Index(['pid', 'cid', 'sys', 'dia', 'HR', 'MAB2', 'MAD3P', 'MADP', 'MASEP',\n",
            "       'MASEPMax', 'MASRP', 'MB0', 'MB1', 'MB2', 'MD3P', 'MDP', 'MFDP',\n",
            "       'MFSEP', 'MFSRP', 'MPI', 'MRR', 'MS4S1', 'MS5S2', 'MSEP', 'MSEPMax',\n",
            "       'MSNR', 'MSRP', 'MWaveType', 'MdD3P', 'MdDP', 'MdSEP', 'MdSEPMax',\n",
            "       'MdSRP', 'BR'],\n",
            "      dtype='object')\n",
            "    pid    cid  sys  dia     HR  ...   MdDP  MdSEP  MdSEPMax  MdSRP     BR\n",
            "0  1003  46667  139   96  73.80  ...  818.0  816.0     825.0  819.0  0.327\n",
            "1  1003  47085  123   87  69.28  ...  867.0  864.0     879.0  869.0  0.213\n",
            "2  1003  47089  114   76  75.28  ...  797.0  797.0     797.0  797.0  0.423\n",
            "3  1003  47164  120   63  74.91  ...  800.0  801.0     813.0  802.0  0.230\n",
            "4  1003  47238  119   83  67.95  ...  890.0  883.0     890.0  887.0  0.253\n",
            "\n",
            "[5 rows x 34 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GfCaNXKGPqdx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "cac737df-bf2a-4e05-a7ce-8acc9f9ecdb6"
      },
      "source": [
        "def sys_zone(value):\n",
        "    if 110 <= value <= 130:\n",
        "        return 'Normal'\n",
        "    elif value <= 100:\n",
        "        return 'Low'\n",
        "    elif value >= 140:\n",
        "        return 'High'\n",
        "    else:\n",
        "        return np.nan\n",
        "def dia_zone(value):\n",
        "    if 70 <= value <= 90:\n",
        "        return 'Normal'\n",
        "    elif value <= 60:\n",
        "        return 'Low'\n",
        "    elif value >= 100:\n",
        "        return 'Low'\n",
        "    else:\n",
        "        return np.nan\n",
        "# print(df.loc[df['dia']==50])\n",
        "df['sys_labels'] = df['sys'].apply(sys_zone)\n",
        "df['dia_labels'] = df['dia'].apply(dia_zone)\n",
        "df_spaced = df.dropna()\n",
        "df_spaced.describe()\n",
        "df_4994 = df_spaced.query('pid==\"4994\"')\n",
        "df_4994.describe()\n",
        "print(df_4994)"
      ],
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "       pid    cid  sys  dia  ...  MdSRP     BR  sys_labels  dia_labels\n",
            "1324  4994  45888  120   80  ...  679.0  0.167      Normal      Normal\n",
            "1325  4994  46277  140   85  ...  698.0  0.243        High      Normal\n",
            "1326  4994  46314  142   74  ...  651.0  0.230        High      Normal\n",
            "1328  4994  46456  146   81  ...  645.0  0.233        High      Normal\n",
            "1329  4994  46467  143   81  ...  679.0  0.210        High      Normal\n",
            "1331  4994  46638  142   76  ...  685.0  0.217        High      Normal\n",
            "1332  4994  46682  143   82  ...  616.0  0.230        High      Normal\n",
            "1334  4994  46752  124   73  ...  726.0  0.223      Normal      Normal\n",
            "1335  4994  46771  140   74  ...  599.0  0.243        High      Normal\n",
            "1336  4994  46790  141   78  ...  614.0  0.270        High      Normal\n",
            "1337  4994  46799  141   86  ...  547.0  0.310        High      Normal\n",
            "1338  4994  46855  111   70  ...  746.0  0.250      Normal      Normal\n",
            "1340  4994  46886  110   75  ...  791.0  0.206      Normal      Normal\n",
            "1343  4994  46970  144   74  ...  731.0  0.244        High      Normal\n",
            "1348  4994  47101  141   76  ...  664.0  0.256        High      Normal\n",
            "1349  4994  47107  125   73  ...  817.0  0.222      Normal      Normal\n",
            "1351  4994  47151  126   70  ...  819.0  0.217      Normal      Normal\n",
            "1352  4994  47173  129   73  ...  751.0  0.256      Normal      Normal\n",
            "1353  4994  47195  144   78  ...  774.0  0.294        High      Normal\n",
            "1354  4994  47206  124   74  ...  734.0  0.389      Normal      Normal\n",
            "1355  4994  47222  130   75  ...  676.0  0.217      Normal      Normal\n",
            "1356  4994  47261  126   79  ...  763.0  0.228      Normal      Normal\n",
            "1358  4994  47306  129   73  ...  664.0  0.217      Normal      Normal\n",
            "1361  4994  47345  143   74  ...  687.0  0.178        High      Normal\n",
            "1362  4994  47357  147   80  ...  717.0  0.172        High      Normal\n",
            "1364  4994  47505  126   72  ...  728.0  0.178      Normal      Normal\n",
            "1367  4994  47821  130   74  ...  753.0  0.217      Normal      Normal\n",
            "1368  4994  47834  141   75  ...  654.0  0.217        High      Normal\n",
            "1371  4994  49314  150   81  ...  670.0  0.222        High      Normal\n",
            "1372  4994  49355  152   72  ...  664.0  0.239        High      Normal\n",
            "...    ...    ...  ...  ...  ...    ...    ...         ...         ...\n",
            "1431  4994  54132  151   78  ...  599.0  0.283        High      Normal\n",
            "1433  4994  54682  126   79  ...  641.0  0.222      Normal      Normal\n",
            "1437  4994  55236  121   70  ...  719.0  0.322      Normal      Normal\n",
            "1439  4994  55369  141   77  ...  590.0  0.189        High      Normal\n",
            "1442  4994  57582  124   74  ...  744.0  0.206      Normal      Normal\n",
            "1444  4994  58315  142   80  ...  756.0  0.300        High      Normal\n",
            "1449  4994  58643  142   77  ...  696.0  0.222        High      Normal\n",
            "1450  4994  59165  129   73  ...  797.0  0.217      Normal      Normal\n",
            "1451  4994  59274  143   78  ...  876.0  0.211        High      Normal\n",
            "1452  4994  62356  127   81  ...  737.0  0.194      Normal      Normal\n",
            "1454  4994  62733  130   78  ...  749.0  0.256      Normal      Normal\n",
            "1455  4994  62900  129   87  ...  680.0  0.283      Normal      Normal\n",
            "1457  4994  63141  130   77  ...  699.0  0.194      Normal      Normal\n",
            "1458  4994  63389  117   78  ...  719.0  0.233      Normal      Normal\n",
            "1459  4994  64195  140   77  ...  776.0  0.211        High      Normal\n",
            "1460  4994  64196  121   81  ...  752.0  0.205      Normal      Normal\n",
            "1461  4994  64197  123   86  ...  721.0  0.183      Normal      Normal\n",
            "1462  4994  66443  128   80  ...  783.0  0.217      Normal      Normal\n",
            "1464  4994  66514  141   84  ...  668.0  0.222        High      Normal\n",
            "1465  4994  66542  124   80  ...  761.0  0.228      Normal      Normal\n",
            "1466  4994  66608  143   84  ...  617.0  0.294        High      Normal\n",
            "1471  4994  67041  129   84  ...  763.0  0.228      Normal      Normal\n",
            "1472  4994  67053  129   84  ...  673.0  0.239      Normal      Normal\n",
            "1476  4994  78604  141   77  ...  717.0  0.256        High      Normal\n",
            "1480  4994  79457  120   73  ...  730.0  0.150      Normal      Normal\n",
            "1482  4994  79460  123   78  ...  766.0  0.133      Normal      Normal\n",
            "1483  4994  79461  128   76  ...  782.0  0.266      Normal      Normal\n",
            "1485  4994  79463  124   75  ...  783.0  0.233      Normal      Normal\n",
            "1486  4994  79600  126   72  ...  927.0  0.217      Normal      Normal\n",
            "1488  4994  83483  126   79  ...  798.0  0.206      Normal      Normal\n",
            "\n",
            "[93 rows x 35 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R6_iskYKXph-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        },
        "outputId": "12acc11d-78c6-4960-9ef9-526d04ffb346"
      },
      "source": [
        "print(df_4994.columns)\n",
        "#df_4994.drop(['pid', 'cid', 'dia_labels'], axis=1, inplace=True)\n",
        "# ax = sns.scatterplot(x=\"HR\", y=\"sys_labels\", data=df_4994)\n",
        "# sns.pairplot(df_4994, hue=\"sys_labels\")"
      ],
      "execution_count": 132,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Index(['pid', 'cid', 'sys', 'dia', 'HR', 'MAB2', 'MAD3P', 'MADP', 'MASEP',\n",
            "       'MASEPMax', 'MASRP', 'MB0', 'MB2', 'MD3P', 'MDP', 'MFDP', 'MFSEP',\n",
            "       'MFSRP', 'MPI', 'MRR', 'MS4S1', 'MS5S2', 'MSEP', 'MSEPMax', 'MSNR',\n",
            "       'MSRP', 'MWaveType', 'MdD3P', 'MdDP', 'MdSEP', 'MdSEPMax', 'MdSRP',\n",
            "       'BR', 'sys_labels', 'dia_labels'],\n",
            "      dtype='object')\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pWo3wqO6bKE9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# sns.pairplot(df_4994, hue=\"sys_labels\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RENSUacDrCSC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        },
        "outputId": "ba0d39a2-0297-4a2d-b86a-6a9b64a8471c"
      },
      "source": [
        "result = df_4994.corr()\n",
        "features = df_4994.drop(['pid', 'cid', 'sys', 'dia', 'sys_labels', 'dia_labels'], axis=1)\n",
        "result = features.corr()\n",
        "print(type(result))\n",
        "corr_threshold = 0.98\n",
        "drop_list =  []\n",
        "for index, row in result.iterrows():\n",
        "    # print(row)\n",
        "    for index_ser, value in row.iteritems():\n",
        "\n",
        "        if index != index_ser and abs(value) > corr_threshold and index_ser not in ['HR']:\n",
        "            drop_list.append(index_ser)\n",
        "df_test = DataFrame({'names': drop_list})\n",
        "uniques = df_test['names'].unique()\n",
        "print(uniques)\n",
        "cleared_features = features.drop(uniques, axis=1)\n",
        "'''\n",
        "for column_name, ser in result.iteritems():\n",
        "    if column_name != index and \n",
        "\n",
        "print(cleared_features.corr())\n",
        "cleared_features['sys'] = df_4994['sys']\n",
        "cleared_features['sys_labels'] = df_4994['sys_labels']\n",
        "sns.pairplot(cleared_features, hue=\"sys_labels\")\n",
        "'''"
      ],
      "execution_count": 134,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "['MRR' 'MdD3P' 'MdDP' 'MdSEP' 'MdSEPMax' 'MdSRP' 'MAD3P' 'MASEP'\n",
            " 'MASEPMax' 'MAB2']\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'\\nfor column_name, ser in result.iteritems():\\n    if column_name != index and \\n\\nprint(cleared_features.corr())\\ncleared_features[\\'sys\\'] = df_4994[\\'sys\\']\\ncleared_features[\\'sys_labels\\'] = df_4994[\\'sys_labels\\']\\nsns.pairplot(cleared_features, hue=\"sys_labels\")\\n'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 134
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IKTVmnfKOP3T",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "12006e54-3bf9-41a2-992c-a18d4f9da37e"
      },
      "source": [
        "X = cleared_features\n",
        "print(X.head())\n",
        "y = df_4994['sys']\n",
        "clf = SelectKBest(f_regression, k=5)\n",
        "X_new = clf.fit_transform(X, y)\n",
        "print(clf.get_support(indices=True))\n",
        "X_new = DataFrame(data = X_new, columns = X.columns[clf.get_support(indices=True)])\n",
        "print(X_new)\n",
        "print(y)\n",
        "LinearRegression\n",
        "clf = LinearRegression()\n",
        "# clf = RandomForestRegressor(n_estimators=20, max_depth=5, min_samples_leaf=5, min_samples_split=5)\n",
        "mean_cross_val_score = cross_val_score(clf, X_new, y, cv=5, scoring='neg_mean_squared_error').mean()\n",
        "# import sklearn\n",
        "# print(sorted(sklearn.metrics.SCORERS.keys()))\n",
        "print(cross_val_score(clf, X_new, y, cv=5, scoring='max_error'))"
      ],
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "         HR     MADP    MASRP   MB0  ...  MSNR   MSRP  MWaveType     BR\n",
            "1324  89.55  141.716  249.529  58.0  ...  20.0  206.0        2.0  0.167\n",
            "1325  85.84  481.296  554.263  58.0  ...  25.0  216.0        1.0  0.243\n",
            "1326  92.02  537.754  662.637  55.0  ...  30.0  207.0        1.0  0.230\n",
            "1328  92.88  194.704  243.614  57.0  ...  20.0  212.0        1.0  0.233\n",
            "1329  88.50  460.682  521.475  56.0  ...  25.0  211.0        1.0  0.210\n",
            "\n",
            "[5 rows x 19 columns]\n",
            "[ 0  1  6 15 16]\n",
            "        HR     MADP    MDP  MSNR   MSRP\n",
            "0    89.55  141.716  321.0  20.0  206.0\n",
            "1    85.84  481.296  321.0  25.0  216.0\n",
            "2    92.02  537.754  311.0  30.0  207.0\n",
            "3    92.88  194.704  318.0  20.0  212.0\n",
            "4    88.50  460.682  312.0  25.0  211.0\n",
            "5    87.59  510.256  314.0  30.0  208.0\n",
            "6    97.09  191.977  311.0  20.0  203.0\n",
            "7    82.87  174.312  318.0  20.0  221.0\n",
            "8   100.17  462.508  310.0  30.0  192.0\n",
            "9    98.36  385.840  299.0  25.0  191.0\n",
            "10  109.69  397.465  315.0  30.0  196.0\n",
            "11   80.65  475.427  337.0  30.0  228.0\n",
            "12   75.95  540.938  344.0  30.0  238.0\n",
            "13   82.42  314.397  309.0  25.0  217.0\n",
            "14   90.36  575.702  319.0  30.0  221.0\n",
            "15   73.35  551.708  354.0  30.0  245.0\n",
            "16   73.17  636.243  347.0  30.0  240.0\n",
            "17   80.21  589.982  329.0  30.0  227.0\n",
            "18   77.52  568.934  336.0  30.0  224.0\n",
            "19   82.08  132.465  358.0  20.0  239.0\n",
            "20   88.89  314.505  324.0  25.0  225.0\n",
            "21   78.64  450.663  315.0  25.0  222.0\n",
            "22   90.50  450.821  317.0  30.0  210.0\n",
            "23   87.34  234.585  321.0  20.0  219.0\n",
            "24   83.68  322.904  323.0  25.0  221.0\n",
            "25   82.30  528.803  327.0  30.0  218.0\n",
            "26   79.79  474.730  314.0  25.0  215.0\n",
            "27   91.60  476.322  308.0  30.0  209.0\n",
            "28   89.55  421.316  311.0  25.0  208.0\n",
            "29   90.50  447.621  314.0  25.0  212.0\n",
            "..     ...      ...    ...   ...    ...\n",
            "63  100.00  397.826  297.0  30.0  193.0\n",
            "64   93.60  430.940  312.0  30.0  196.0\n",
            "65   83.45  288.397  337.0  30.0  221.0\n",
            "66  101.69  300.003  307.0  30.0  178.0\n",
            "67   80.86  523.628  307.0  25.0  198.0\n",
            "68   79.47  516.680  307.0  35.0  212.0\n",
            "69   86.33  487.612  320.0  30.0  218.0\n",
            "70   75.19  531.312  346.0  35.0  232.0\n",
            "71   68.34  630.940  352.0  35.0  237.0\n",
            "72   81.63  524.384  328.0  35.0  219.0\n",
            "73   80.11  752.379  347.0  35.0  225.0\n",
            "74   88.24  397.567  319.0  30.0  216.0\n",
            "75   85.71  583.668  336.0  35.0  218.0\n",
            "76   83.80  933.359  343.0  35.0  231.0\n",
            "77   77.32  537.700  326.0  30.0  216.0\n",
            "78   79.89  533.930  322.0  30.0  216.0\n",
            "79   83.33  432.454  316.0  30.0  212.0\n",
            "80   76.82  411.262  333.0  30.0  227.0\n",
            "81   89.55  258.332  319.0  25.0  212.0\n",
            "82   79.05  568.211  319.0  35.0  216.0\n",
            "83   96.93  423.727  310.0  30.0  194.0\n",
            "84   78.95  370.291  326.0  30.0  226.0\n",
            "85   89.02  567.560  318.0  35.0  208.0\n",
            "86   83.80  654.891  310.0  35.0  202.0\n",
            "87   82.76  278.972  323.0  30.0  221.0\n",
            "88   78.12  615.845  314.0  35.0  216.0\n",
            "89   76.53  363.907  325.0  30.0  218.0\n",
            "90   76.73  289.097  329.0  30.0  223.0\n",
            "91   64.72  580.238  335.0  35.0  238.0\n",
            "92   75.19  289.296  348.0  25.0  239.0\n",
            "\n",
            "[93 rows x 5 columns]\n",
            "1324    120\n",
            "1325    140\n",
            "1326    142\n",
            "1328    146\n",
            "1329    143\n",
            "1331    142\n",
            "1332    143\n",
            "1334    124\n",
            "1335    140\n",
            "1336    141\n",
            "1337    141\n",
            "1338    111\n",
            "1340    110\n",
            "1343    144\n",
            "1348    141\n",
            "1349    125\n",
            "1351    126\n",
            "1352    129\n",
            "1353    144\n",
            "1354    124\n",
            "1355    130\n",
            "1356    126\n",
            "1358    129\n",
            "1361    143\n",
            "1362    147\n",
            "1364    126\n",
            "1367    130\n",
            "1368    141\n",
            "1371    150\n",
            "1372    152\n",
            "       ... \n",
            "1431    151\n",
            "1433    126\n",
            "1437    121\n",
            "1439    141\n",
            "1442    124\n",
            "1444    142\n",
            "1449    142\n",
            "1450    129\n",
            "1451    143\n",
            "1452    127\n",
            "1454    130\n",
            "1455    129\n",
            "1457    130\n",
            "1458    117\n",
            "1459    140\n",
            "1460    121\n",
            "1461    123\n",
            "1462    128\n",
            "1464    141\n",
            "1465    124\n",
            "1466    143\n",
            "1471    129\n",
            "1472    129\n",
            "1476    141\n",
            "1480    120\n",
            "1482    123\n",
            "1483    128\n",
            "1485    124\n",
            "1486    126\n",
            "1488    126\n",
            "Name: sys, Length: 93, dtype: int64\n",
            "[-23.23089609 -15.32713154 -14.97155177 -24.73676129 -14.89401398]\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}