1. Merging datasets.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Enabling reloading extension to use external python files\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os.path\n",
    "import os\n",
    "import columns\n",
    "import datasets\n",
    "import seasons\n",
    "import directory_utils\n",
    "import unidecode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'111': [], '222': [], '265': [], '333': []}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dict = {sheet_name: [] for sheet_name in datasets.SHEETS_NAMES}\n",
    "data_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_sheet_as_data_frame(xls, sheet_name):\n",
    "    sheet = pd.read_excel(xls, sheet_name=sheet_name)\n",
    "    \n",
    "    columns_to_drop = []\n",
    "    for column in sheet.columns:\n",
    "        if(columns.UNNAMED in column):\n",
    "            columns_to_drop.append(column)\n",
    "    df = sheet.drop(columns_to_drop, axis=1) \n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "spreadsheet_names = [str(1997+w) for w in range(0,20)]\n",
    "\n",
    "for name in spreadsheet_names:\n",
    "    file_path = datasets.ORIGINAL_DATA_SETS_PATH + \"/\" + name + \".xlsx\"\n",
    "    \n",
    "    if os.path.isfile(file_path):\n",
    "        with pd.ExcelFile(file_path) as xls:          \n",
    "            for sheet_name in datasets.SHEETS_NAMES:\n",
    "                if sheet_name in xls.sheet_names:\n",
    "                    df = read_sheet_as_data_frame(xls,sheet_name)\n",
    "                    data_dict[sheet_name].append(df)\n",
    "\n",
    "    else:\n",
    "        print('File \"' + file_path + '\" was not found.')       "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of \"333\":(26001, 10)\n",
      "Shape of \"111\":(1050445, 18)\n",
      "Shape of \"222\":(6914, 21)\n",
      "Shape of \"265\":(6864, 13)\n"
     ]
    }
   ],
   "source": [
    "data_frames = {sheet_name: pd.concat(data_frames) for sheet_name, data_frames in data_dict.items()}\n",
    "for name, data_frame in data_frames.items():\n",
    "    data_frame.columns = data_frame.columns.map(lambda name: unidecode.unidecode(name))\n",
    "    print('Shape of \\\"' + name + '\\\":' + str(data_frame.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "directory_utils.create_if_doesnt_exist(datasets.MERGED_CSVS_DIRECTORY_PATH)\n",
    "\n",
    "for sheet_name, data_frame in data_frames.items():\n",
    "    csv_output_file_path = datasets.MERGED_CSVS_DIRECTORY_PATH + '/' + sheet_name + '.csv'\n",
    "    data_frame.to_csv(csv_output_file_path , index = False, encoding=datasets.DEFAULT_CSV_ENCODING)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_frame_111 = data_frames[datasets.SHEET_111_NAME]\n",
    "data_frame_222 = data_frames[datasets.SHEET_222_NAME]\n",
    "data_frame_333 = data_frames[datasets.SHEET_333_NAME]\n",
    "data_frame_265 = data_frames[datasets.SHEET_265_NAME]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyzing 111 data frame"
   ]
  },
  {
   "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>Cod</th>\n",
       "      <th>Ano</th>\n",
       "      <th>Dia Juliano</th>\n",
       "      <th>Hora - minuto</th>\n",
       "      <th>Veloc. Instantanea do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. Instantanea do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instante da aquisicao (deg)</th>\n",
       "      <th>Radiacao Incidente total de 10 leituras (KW/m2)</th>\n",
       "      <th>Radiacao Refletida Total de 10 leituras (KW/m2)</th>\n",
       "      <th>Fluxo Positivo de calor no solo total de 10 leituras (W/m2)</th>\n",
       "      <th>Fluxo negativo de calor no solo . total de 10 leituras (W/m2)</th>\n",
       "      <th>Umidade relativa Media (%)</th>\n",
       "      <th>Temperatura do Ar Media (degC)</th>\n",
       "      <th>Temperatura do solo  1 Media (degC) a 3cm no solo</th>\n",
       "      <th>Temperatura do solo 2Media (degC) a 6cm no solo</th>\n",
       "      <th>Temperatura do solo 3Media (degC) Colocado sensor de UV em 31/07/09</th>\n",
       "      <th>Pressao (hPa)</th>\n",
       "      <th>Chuva total do periodo (mm)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>111</td>\n",
       "      <td>1997</td>\n",
       "      <td>1.0</td>\n",
       "      <td>10</td>\n",
       "      <td>1618.00</td>\n",
       "      <td>1115.000</td>\n",
       "      <td>8.32</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-443.0</td>\n",
       "      <td>92.3</td>\n",
       "      <td>19.84</td>\n",
       "      <td>21.04</td>\n",
       "      <td>22.68</td>\n",
       "      <td>24.59</td>\n",
       "      <td>943.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>111</td>\n",
       "      <td>1997</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20</td>\n",
       "      <td>0.99</td>\n",
       "      <td>0.763</td>\n",
       "      <td>20.59</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-443.0</td>\n",
       "      <td>93.1</td>\n",
       "      <td>19.63</td>\n",
       "      <td>20.97</td>\n",
       "      <td>22.65</td>\n",
       "      <td>24.52</td>\n",
       "      <td>943.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Cod   Ano  Dia Juliano  Hora - minuto  \\\n",
       "0  111  1997          1.0             10   \n",
       "1  111  1997          1.0             20   \n",
       "\n",
       "   Veloc. Instantanea do Vento a 5 metros (m/s)  \\\n",
       "0                                       1618.00   \n",
       "1                                          0.99   \n",
       "\n",
       "   Veloc. Instantanea do Vento a 2 metros (m/s)  \\\n",
       "0                                      1115.000   \n",
       "1                                         0.763   \n",
       "\n",
       "   Direcao do Vento no instante da aquisicao (deg)  \\\n",
       "0                                             8.32   \n",
       "1                                            20.59   \n",
       "\n",
       "   Radiacao Incidente total de 10 leituras (KW/m2)  \\\n",
       "0                                              0.0   \n",
       "1                                              0.0   \n",
       "\n",
       "   Radiacao Refletida Total de 10 leituras (KW/m2)  \\\n",
       "0                                              0.0   \n",
       "1                                              0.0   \n",
       "\n",
       "   Fluxo Positivo de calor no solo total de 10 leituras (W/m2)  \\\n",
       "0                                                0.0             \n",
       "1                                                0.0             \n",
       "\n",
       "   Fluxo negativo de calor no solo . total de 10 leituras (W/m2)  \\\n",
       "0                                             -443.0               \n",
       "1                                             -443.0               \n",
       "\n",
       "   Umidade relativa Media (%)  Temperatura do Ar Media (degC)  \\\n",
       "0                        92.3                           19.84   \n",
       "1                        93.1                           19.63   \n",
       "\n",
       "   Temperatura do solo  1 Media (degC) a 3cm no solo  \\\n",
       "0                                              21.04   \n",
       "1                                              20.97   \n",
       "\n",
       "   Temperatura do solo 2Media (degC) a 6cm no solo   \\\n",
       "0                                             22.68   \n",
       "1                                             22.65   \n",
       "\n",
       "  Temperatura do solo 3Media (degC) Colocado sensor de UV em 31/07/09  \\\n",
       "0                                              24.59                    \n",
       "1                                              24.52                    \n",
       "\n",
       "   Pressao (hPa) Chuva total do periodo (mm)  \n",
       "0          943.0                           0  \n",
       "1          943.0                           0  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_111.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Cod', u'Ano', u'Dia Juliano', u'Hora - minuto',\n",
       "       u'Veloc. Instantanea do Vento a 5 metros (m/s)',\n",
       "       u'Veloc. Instantanea do Vento a 2 metros (m/s)',\n",
       "       u'Direcao do Vento no instante da aquisicao (deg)',\n",
       "       u'Radiacao Incidente total de 10 leituras (KW/m2)',\n",
       "       u'Radiacao Refletida Total de 10 leituras (KW/m2)',\n",
       "       u'Fluxo Positivo de calor no solo total de 10 leituras (W/m2)',\n",
       "       u'Fluxo negativo de calor no solo . total de 10 leituras (W/m2)',\n",
       "       u'Umidade relativa Media (%)', u'Temperatura do Ar Media (degC)',\n",
       "       u'Temperatura do solo  1 Media (degC) a 3cm no solo',\n",
       "       u'Temperatura do solo 2Media (degC) a 6cm no solo ',\n",
       "       u'Temperatura do solo 3Media (degC) Colocado sensor de UV em 31/07/09',\n",
       "       u'Pressao (hPa)', u'Chuva total do periodo (mm)'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_111.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([111, 333, 265, 222])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_111.Cod.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2005]\n"
     ]
    }
   ],
   "source": [
    "print(data_frame_111[data_frame_111.Cod != 111][columns.YEAR].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Cod                                                                       0\n",
       "Ano                                                                       0\n",
       "Dia Juliano                                                               0\n",
       "Hora - minuto                                                             0\n",
       "Veloc. Instantanea do Vento a 5 metros (m/s)                              0\n",
       "Veloc. Instantanea do Vento a 2 metros (m/s)                              0\n",
       "Direcao do Vento no instante da aquisicao (deg)                           0\n",
       "Radiacao Incidente total de 10 leituras (KW/m2)                           0\n",
       "Radiacao Refletida Total de 10 leituras (KW/m2)                           0\n",
       "Fluxo Positivo de calor no solo total de 10 leituras (W/m2)               0\n",
       "Fluxo negativo de calor no solo . total de 10 leituras (W/m2)          1825\n",
       "Umidade relativa Media (%)                                             1826\n",
       "Temperatura do Ar Media (degC)                                         1826\n",
       "Temperatura do solo  1 Media (degC) a 3cm no solo                      2191\n",
       "Temperatura do solo 2Media (degC) a 6cm no solo                        2191\n",
       "Temperatura do solo 3Media (degC) Colocado sensor de UV em 31/07/09    2191\n",
       "Pressao (hPa)                                                          2191\n",
       "Chuva total do periodo (mm)                                            2191\n",
       "dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_111.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>Cod</th>\n",
       "      <th>Ano</th>\n",
       "      <th>Dia Juliano</th>\n",
       "      <th>Hora - minuto</th>\n",
       "      <th>Veloc. Instantanea do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. Instantanea do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instante da aquisicao (deg)</th>\n",
       "      <th>Radiacao Incidente total de 10 leituras (KW/m2)</th>\n",
       "      <th>Radiacao Refletida Total de 10 leituras (KW/m2)</th>\n",
       "      <th>Fluxo Positivo de calor no solo total de 10 leituras (W/m2)</th>\n",
       "      <th>Fluxo negativo de calor no solo . total de 10 leituras (W/m2)</th>\n",
       "      <th>Umidade relativa Media (%)</th>\n",
       "      <th>Temperatura do Ar Media (degC)</th>\n",
       "      <th>Temperatura do solo  1 Media (degC) a 3cm no solo</th>\n",
       "      <th>Temperatura do solo 2Media (degC) a 6cm no solo</th>\n",
       "      <th>Pressao (hPa)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.050445e+06</td>\n",
       "      <td>1.048620e+06</td>\n",
       "      <td>1.048619e+06</td>\n",
       "      <td>1.048619e+06</td>\n",
       "      <td>1.048254e+06</td>\n",
       "      <td>1.048254e+06</td>\n",
       "      <td>1.048254e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.114778e+02</td>\n",
       "      <td>2.006549e+03</td>\n",
       "      <td>1.835618e+02</td>\n",
       "      <td>1.192104e+03</td>\n",
       "      <td>1.472154e+03</td>\n",
       "      <td>9.129840e+02</td>\n",
       "      <td>2.723121e+02</td>\n",
       "      <td>7.622076e+02</td>\n",
       "      <td>2.706437e+02</td>\n",
       "      <td>3.653693e+02</td>\n",
       "      <td>-1.662485e+03</td>\n",
       "      <td>7.404259e+01</td>\n",
       "      <td>2.634531e+01</td>\n",
       "      <td>1.332214e+03</td>\n",
       "      <td>1.296305e+03</td>\n",
       "      <td>9.462274e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>9.895335e+00</td>\n",
       "      <td>5.762457e+00</td>\n",
       "      <td>1.053855e+02</td>\n",
       "      <td>6.927634e+02</td>\n",
       "      <td>1.353874e+03</td>\n",
       "      <td>1.134761e+03</td>\n",
       "      <td>6.693158e+02</td>\n",
       "      <td>1.611430e+03</td>\n",
       "      <td>8.889303e+02</td>\n",
       "      <td>1.147225e+03</td>\n",
       "      <td>2.796438e+03</td>\n",
       "      <td>2.748618e+01</td>\n",
       "      <td>2.413575e+02</td>\n",
       "      <td>2.725003e+03</td>\n",
       "      <td>2.694402e+03</td>\n",
       "      <td>8.089817e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.110000e+02</td>\n",
       "      <td>1.997000e+03</td>\n",
       "      <td>-5.336000e+01</td>\n",
       "      <td>1.000000e+01</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "      <td>-6.999000e+03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.110000e+02</td>\n",
       "      <td>2.002000e+03</td>\n",
       "      <td>9.300000e+01</td>\n",
       "      <td>6.100000e+02</td>\n",
       "      <td>1.010000e+00</td>\n",
       "      <td>5.390000e-01</td>\n",
       "      <td>9.380000e+01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-5.298000e+02</td>\n",
       "      <td>6.053000e+01</td>\n",
       "      <td>1.813000e+01</td>\n",
       "      <td>2.049000e+01</td>\n",
       "      <td>2.117000e+01</td>\n",
       "      <td>9.430000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.110000e+02</td>\n",
       "      <td>2.007000e+03</td>\n",
       "      <td>1.840000e+02</td>\n",
       "      <td>1.210000e+03</td>\n",
       "      <td>1.481000e+03</td>\n",
       "      <td>1.010000e+00</td>\n",
       "      <td>1.555000e+02</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>-2.460000e+02</td>\n",
       "      <td>7.750000e+01</td>\n",
       "      <td>2.114000e+01</td>\n",
       "      <td>2.448000e+01</td>\n",
       "      <td>2.493000e+01</td>\n",
       "      <td>9.470000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.110000e+02</td>\n",
       "      <td>2.012000e+03</td>\n",
       "      <td>2.750000e+02</td>\n",
       "      <td>1.810000e+03</td>\n",
       "      <td>2.265000e+03</td>\n",
       "      <td>1.716000e+03</td>\n",
       "      <td>2.576000e+02</td>\n",
       "      <td>4.338000e+02</td>\n",
       "      <td>9.890000e-01</td>\n",
       "      <td>1.788000e+02</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>8.920000e+01</td>\n",
       "      <td>2.510000e+01</td>\n",
       "      <td>3.198000e+01</td>\n",
       "      <td>2.988000e+01</td>\n",
       "      <td>9.520000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3.330000e+02</td>\n",
       "      <td>2.139000e+03</td>\n",
       "      <td>6.540000e+02</td>\n",
       "      <td>6.640000e+03</td>\n",
       "      <td>6.996000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.538000e+03</td>\n",
       "      <td>6.836000e+03</td>\n",
       "      <td>1.822660e+05</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>6.999000e+03</td>\n",
       "      <td>2.802000e+03</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Cod           Ano   Dia Juliano  Hora - minuto  \\\n",
       "count  1.050445e+06  1.050445e+06  1.050445e+06   1.050445e+06   \n",
       "mean   1.114778e+02  2.006549e+03  1.835618e+02   1.192104e+03   \n",
       "std    9.895335e+00  5.762457e+00  1.053855e+02   6.927634e+02   \n",
       "min    1.110000e+02  1.997000e+03 -5.336000e+01   1.000000e+01   \n",
       "25%    1.110000e+02  2.002000e+03  9.300000e+01   6.100000e+02   \n",
       "50%    1.110000e+02  2.007000e+03  1.840000e+02   1.210000e+03   \n",
       "75%    1.110000e+02  2.012000e+03  2.750000e+02   1.810000e+03   \n",
       "max    3.330000e+02  2.139000e+03  6.540000e+02   6.640000e+03   \n",
       "\n",
       "       Veloc. Instantanea do Vento a 5 metros (m/s)  \\\n",
       "count                                  1.050445e+06   \n",
       "mean                                   1.472154e+03   \n",
       "std                                    1.353874e+03   \n",
       "min                                   -6.999000e+03   \n",
       "25%                                    1.010000e+00   \n",
       "50%                                    1.481000e+03   \n",
       "75%                                    2.265000e+03   \n",
       "max                                    6.996000e+03   \n",
       "\n",
       "       Veloc. Instantanea do Vento a 2 metros (m/s)  \\\n",
       "count                                  1.050445e+06   \n",
       "mean                                   9.129840e+02   \n",
       "std                                    1.134761e+03   \n",
       "min                                    0.000000e+00   \n",
       "25%                                    5.390000e-01   \n",
       "50%                                    1.010000e+00   \n",
       "75%                                    1.716000e+03   \n",
       "max                                    6.999000e+03   \n",
       "\n",
       "       Direcao do Vento no instante da aquisicao (deg)  \\\n",
       "count                                     1.050445e+06   \n",
       "mean                                      2.723121e+02   \n",
       "std                                       6.693158e+02   \n",
       "min                                       0.000000e+00   \n",
       "25%                                       9.380000e+01   \n",
       "50%                                       1.555000e+02   \n",
       "75%                                       2.576000e+02   \n",
       "max                                       6.999000e+03   \n",
       "\n",
       "       Radiacao Incidente total de 10 leituras (KW/m2)  \\\n",
       "count                                     1.050445e+06   \n",
       "mean                                      7.622076e+02   \n",
       "std                                       1.611430e+03   \n",
       "min                                      -6.999000e+03   \n",
       "25%                                       0.000000e+00   \n",
       "50%                                       0.000000e+00   \n",
       "75%                                       4.338000e+02   \n",
       "max                                       6.999000e+03   \n",
       "\n",
       "       Radiacao Refletida Total de 10 leituras (KW/m2)  \\\n",
       "count                                     1.050445e+06   \n",
       "mean                                      2.706437e+02   \n",
       "std                                       8.889303e+02   \n",
       "min                                       0.000000e+00   \n",
       "25%                                       0.000000e+00   \n",
       "50%                                       0.000000e+00   \n",
       "75%                                       9.890000e-01   \n",
       "max                                       6.999000e+03   \n",
       "\n",
       "       Fluxo Positivo de calor no solo total de 10 leituras (W/m2)  \\\n",
       "count                                       1.050445e+06             \n",
       "mean                                        3.653693e+02             \n",
       "std                                         1.147225e+03             \n",
       "min                                         0.000000e+00             \n",
       "25%                                         0.000000e+00             \n",
       "50%                                         0.000000e+00             \n",
       "75%                                         1.788000e+02             \n",
       "max                                         6.999000e+03             \n",
       "\n",
       "       Fluxo negativo de calor no solo . total de 10 leituras (W/m2)  \\\n",
       "count                                       1.048620e+06               \n",
       "mean                                       -1.662485e+03               \n",
       "std                                         2.796438e+03               \n",
       "min                                        -6.999000e+03               \n",
       "25%                                        -5.298000e+02               \n",
       "50%                                        -2.460000e+02               \n",
       "75%                                         0.000000e+00               \n",
       "max                                         6.538000e+03               \n",
       "\n",
       "       Umidade relativa Media (%)  Temperatura do Ar Media (degC)  \\\n",
       "count                1.048619e+06                    1.048619e+06   \n",
       "mean                 7.404259e+01                    2.634531e+01   \n",
       "std                  2.748618e+01                    2.413575e+02   \n",
       "min                 -6.999000e+03                   -6.999000e+03   \n",
       "25%                  6.053000e+01                    1.813000e+01   \n",
       "50%                  7.750000e+01                    2.114000e+01   \n",
       "75%                  8.920000e+01                    2.510000e+01   \n",
       "max                  6.836000e+03                    1.822660e+05   \n",
       "\n",
       "       Temperatura do solo  1 Media (degC) a 3cm no solo  \\\n",
       "count                                       1.048254e+06   \n",
       "mean                                        1.332214e+03   \n",
       "std                                         2.725003e+03   \n",
       "min                                        -6.999000e+03   \n",
       "25%                                         2.049000e+01   \n",
       "50%                                         2.448000e+01   \n",
       "75%                                         3.198000e+01   \n",
       "max                                         6.999000e+03   \n",
       "\n",
       "       Temperatura do solo 2Media (degC) a 6cm no solo   Pressao (hPa)  \n",
       "count                                      1.048254e+06   1.048254e+06  \n",
       "mean                                       1.296305e+03   9.462274e+02  \n",
       "std                                        2.694402e+03   8.089817e+01  \n",
       "min                                       -6.999000e+03  -6.999000e+03  \n",
       "25%                                        2.117000e+01   9.430000e+02  \n",
       "50%                                        2.493000e+01   9.470000e+02  \n",
       "75%                                        2.988000e+01   9.520000e+02  \n",
       "max                                        6.999000e+03   2.802000e+03  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_111.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyzing 222 data frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Cod</th>\n",
       "      <th>Ano</th>\n",
       "      <th>Dia Juliano</th>\n",
       "      <th>Hora - minuto</th>\n",
       "      <th>Veloc. maxima do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. maxima do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instante da veloc. Max.(deg)</th>\n",
       "      <th>Horario de ocorrencia vr max. veloc. do vento</th>\n",
       "      <th>Radiacao Incidente Total do dia (KW/m2)</th>\n",
       "      <th>Radiacao Refletida Total do dia (KW/m2)</th>\n",
       "      <th>...</th>\n",
       "      <th>Fluxo Negativo de calor no solo. Total do dia (KW/m2)</th>\n",
       "      <th>Temperatura do Ar  Maxima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar Maxima(degC)</th>\n",
       "      <th>Temperatura do Ar Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar Minima</th>\n",
       "      <th>Temperatura do Solo 1 Maxima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Solo 1 Maxima.</th>\n",
       "      <th>Temperatura do Solo 1 Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Solo 1 minima</th>\n",
       "      <th>Chuva total do dia (mm)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>222</td>\n",
       "      <td>1997</td>\n",
       "      <td>1</td>\n",
       "      <td>2400</td>\n",
       "      <td>6031.0</td>\n",
       "      <td>2677.0</td>\n",
       "      <td>99.6</td>\n",
       "      <td>1923</td>\n",
       "      <td>270.8</td>\n",
       "      <td>53.81</td>\n",
       "      <td>...</td>\n",
       "      <td>-34.76</td>\n",
       "      <td>30.16</td>\n",
       "      <td>1255</td>\n",
       "      <td>18.50</td>\n",
       "      <td>551.0</td>\n",
       "      <td>34.53</td>\n",
       "      <td>1258</td>\n",
       "      <td>19.15</td>\n",
       "      <td>606.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>222</td>\n",
       "      <td>1997</td>\n",
       "      <td>2</td>\n",
       "      <td>2400</td>\n",
       "      <td>4756.0</td>\n",
       "      <td>1758.0</td>\n",
       "      <td>119.5</td>\n",
       "      <td>45</td>\n",
       "      <td>165.3</td>\n",
       "      <td>27.46</td>\n",
       "      <td>...</td>\n",
       "      <td>-26.86</td>\n",
       "      <td>26.15</td>\n",
       "      <td>1417</td>\n",
       "      <td>19.15</td>\n",
       "      <td>2355.0</td>\n",
       "      <td>27.66</td>\n",
       "      <td>1425</td>\n",
       "      <td>19.65</td>\n",
       "      <td>2352.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Cod   Ano Dia Juliano Hora - minuto  \\\n",
       "0  222  1997           1          2400   \n",
       "1  222  1997           2          2400   \n",
       "\n",
       "   Veloc. maxima do Vento a 5 metros (m/s)  \\\n",
       "0                                   6031.0   \n",
       "1                                   4756.0   \n",
       "\n",
       "   Veloc. maxima do Vento a 2 metros (m/s)  \\\n",
       "0                                   2677.0   \n",
       "1                                   1758.0   \n",
       "\n",
       "   Direcao do Vento no instante da veloc. Max.(deg)  \\\n",
       "0                                              99.6   \n",
       "1                                             119.5   \n",
       "\n",
       "  Horario de ocorrencia vr max. veloc. do vento  \\\n",
       "0                                          1923   \n",
       "1                                            45   \n",
       "\n",
       "   Radiacao Incidente Total do dia (KW/m2)  \\\n",
       "0                                    270.8   \n",
       "1                                    165.3   \n",
       "\n",
       "   Radiacao Refletida Total do dia (KW/m2)  ...  \\\n",
       "0                                    53.81  ...   \n",
       "1                                    27.46  ...   \n",
       "\n",
       "   Fluxo Negativo de calor no solo. Total do dia (KW/m2)  \\\n",
       "0                                             -34.76       \n",
       "1                                             -26.86       \n",
       "\n",
       "  Temperatura do Ar  Maxima do dia (degC)  \\\n",
       "0                                   30.16   \n",
       "1                                   26.15   \n",
       "\n",
       "  Horario de ocorrencia da Temp. Ar Maxima(degC)  \\\n",
       "0                                           1255   \n",
       "1                                           1417   \n",
       "\n",
       "  Temperatura do Ar Minima do dia (degC)  \\\n",
       "0                                  18.50   \n",
       "1                                  19.15   \n",
       "\n",
       "   Horario de ocorrencia da Temp. Ar Minima  \\\n",
       "0                                     551.0   \n",
       "1                                    2355.0   \n",
       "\n",
       "   Temperatura do Solo 1 Maxima do dia (degC)  \\\n",
       "0                                       34.53   \n",
       "1                                       27.66   \n",
       "\n",
       "  Horario de ocorrencia da Temp. Solo 1 Maxima.  \\\n",
       "0                                          1258   \n",
       "1                                          1425   \n",
       "\n",
       "  Temperatura do Solo 1 Minima do dia (degC)  \\\n",
       "0                                      19.15   \n",
       "1                                      19.65   \n",
       "\n",
       "   Horario de ocorrencia da Temp. Solo 1 minima  Chuva total do dia (mm)  \n",
       "0                                         606.0                      0.0  \n",
       "1                                        2352.0                      0.0  \n",
       "\n",
       "[2 rows x 21 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_222.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Cod', u'Ano', u'Dia Juliano', u'Hora - minuto',\n",
       "       u'Veloc. maxima do Vento a 5 metros (m/s)',\n",
       "       u'Veloc. maxima do Vento a 2 metros (m/s)',\n",
       "       u'Direcao do Vento no instante da veloc. Max.(deg)',\n",
       "       u'Horario de ocorrencia vr max. veloc. do vento',\n",
       "       u'Radiacao Incidente Total do dia (KW/m2)',\n",
       "       u'Radiacao Refletida Total do dia (KW/m2)',\n",
       "       u'Fluxo positivo de calor no solo Total do dia (KW/m2)',\n",
       "       u'Fluxo Negativo de calor no solo. Total do dia (KW/m2)',\n",
       "       u'Temperatura do Ar  Maxima do dia (degC)',\n",
       "       u'Horario de ocorrencia da Temp. Ar Maxima(degC)',\n",
       "       u'Temperatura do Ar Minima do dia (degC)',\n",
       "       u'Horario de ocorrencia da Temp. Ar Minima',\n",
       "       u'Temperatura do Solo 1 Maxima do dia (degC)',\n",
       "       u'Horario de ocorrencia da Temp. Solo 1 Maxima.',\n",
       "       u'Temperatura do Solo 1 Minima do dia (degC)',\n",
       "       u'Horario de ocorrencia da Temp. Solo 1 minima',\n",
       "       u'Chuva total do dia (mm)'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_222.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([222], dtype=object)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_222.Cod.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[]\n"
     ]
    }
   ],
   "source": [
    "print(data_frame_222[data_frame_222.Cod != 222][columns.YEAR].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Cod                                                        0\n",
       "Ano                                                        0\n",
       "Dia Juliano                                                0\n",
       "Hora - minuto                                              0\n",
       "Veloc. maxima do Vento a 5 metros (m/s)                    0\n",
       "Veloc. maxima do Vento a 2 metros (m/s)                    0\n",
       "Direcao do Vento no instante da veloc. Max.(deg)           0\n",
       "Horario de ocorrencia vr max. veloc. do vento              0\n",
       "Radiacao Incidente Total do dia (KW/m2)                    0\n",
       "Radiacao Refletida Total do dia (KW/m2)                    0\n",
       "Fluxo positivo de calor no solo Total do dia (KW/m2)       0\n",
       "Fluxo Negativo de calor no solo. Total do dia (KW/m2)      0\n",
       "Temperatura do Ar  Maxima do dia (degC)                    1\n",
       "Horario de ocorrencia da Temp. Ar Maxima(degC)             1\n",
       "Temperatura do Ar Minima do dia (degC)                     1\n",
       "Horario de ocorrencia da Temp. Ar Minima                   1\n",
       "Temperatura do Solo 1 Maxima do dia (degC)                 1\n",
       "Horario de ocorrencia da Temp. Solo 1 Maxima.              1\n",
       "Temperatura do Solo 1 Minima do dia (degC)                 1\n",
       "Horario de ocorrencia da Temp. Solo 1 minima               2\n",
       "Chuva total do dia (mm)                                  151\n",
       "dtype: int64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_222.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>Veloc. maxima do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. maxima do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instante da veloc. Max.(deg)</th>\n",
       "      <th>Radiacao Incidente Total do dia (KW/m2)</th>\n",
       "      <th>Radiacao Refletida Total do dia (KW/m2)</th>\n",
       "      <th>Fluxo positivo de calor no solo Total do dia (KW/m2)</th>\n",
       "      <th>Temperatura do Ar Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar Minima</th>\n",
       "      <th>Temperatura do Solo 1 Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Solo 1 minima</th>\n",
       "      <th>Chuva total do dia (mm)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6914.000000</td>\n",
       "      <td>6913.000000</td>\n",
       "      <td>6913.000000</td>\n",
       "      <td>6913.000000</td>\n",
       "      <td>6912.00000</td>\n",
       "      <td>6763.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>3666.239481</td>\n",
       "      <td>2584.439435</td>\n",
       "      <td>288.337367</td>\n",
       "      <td>210.470264</td>\n",
       "      <td>120.182050</td>\n",
       "      <td>108.991185</td>\n",
       "      <td>51.907359</td>\n",
       "      <td>791.598438</td>\n",
       "      <td>783.914700</td>\n",
       "      <td>783.52474</td>\n",
       "      <td>252.199053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2038.105213</td>\n",
       "      <td>1860.333669</td>\n",
       "      <td>625.759617</td>\n",
       "      <td>275.510678</td>\n",
       "      <td>424.527474</td>\n",
       "      <td>716.880089</td>\n",
       "      <td>466.418455</td>\n",
       "      <td>656.412518</td>\n",
       "      <td>2249.099354</td>\n",
       "      <td>611.03749</td>\n",
       "      <td>992.209132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-931.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-6999.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3109.000000</td>\n",
       "      <td>5.090000</td>\n",
       "      <td>131.900000</td>\n",
       "      <td>110.600000</td>\n",
       "      <td>42.647500</td>\n",
       "      <td>-201.700000</td>\n",
       "      <td>14.020000</td>\n",
       "      <td>501.000000</td>\n",
       "      <td>15.480000</td>\n",
       "      <td>559.00000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>4109.000000</td>\n",
       "      <td>2854.000000</td>\n",
       "      <td>181.100000</td>\n",
       "      <td>207.700000</td>\n",
       "      <td>65.625000</td>\n",
       "      <td>26.230000</td>\n",
       "      <td>16.910000</td>\n",
       "      <td>558.000000</td>\n",
       "      <td>19.770000</td>\n",
       "      <td>649.00000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>5011.000000</td>\n",
       "      <td>3910.250000</td>\n",
       "      <td>282.600000</td>\n",
       "      <td>274.100000</td>\n",
       "      <td>98.500000</td>\n",
       "      <td>81.275000</td>\n",
       "      <td>19.040000</td>\n",
       "      <td>652.000000</td>\n",
       "      <td>22.910000</td>\n",
       "      <td>732.25000</td>\n",
       "      <td>2.540000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6973.000000</td>\n",
       "      <td>6973.000000</td>\n",
       "      <td>6897.000000</td>\n",
       "      <td>6589.000000</td>\n",
       "      <td>6909.000000</td>\n",
       "      <td>6924.000000</td>\n",
       "      <td>6991.000000</td>\n",
       "      <td>2400.000000</td>\n",
       "      <td>6999.000000</td>\n",
       "      <td>2400.00000</td>\n",
       "      <td>6858.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Veloc. maxima do Vento a 5 metros (m/s)  \\\n",
       "count                              6914.000000   \n",
       "mean                               3666.239481   \n",
       "std                                2038.105213   \n",
       "min                                   0.000000   \n",
       "25%                                3109.000000   \n",
       "50%                                4109.000000   \n",
       "75%                                5011.000000   \n",
       "max                                6973.000000   \n",
       "\n",
       "       Veloc. maxima do Vento a 2 metros (m/s)  \\\n",
       "count                              6914.000000   \n",
       "mean                               2584.439435   \n",
       "std                                1860.333669   \n",
       "min                                   0.000000   \n",
       "25%                                   5.090000   \n",
       "50%                                2854.000000   \n",
       "75%                                3910.250000   \n",
       "max                                6973.000000   \n",
       "\n",
       "       Direcao do Vento no instante da veloc. Max.(deg)  \\\n",
       "count                                       6914.000000   \n",
       "mean                                         288.337367   \n",
       "std                                          625.759617   \n",
       "min                                            0.000000   \n",
       "25%                                          131.900000   \n",
       "50%                                          181.100000   \n",
       "75%                                          282.600000   \n",
       "max                                         6897.000000   \n",
       "\n",
       "       Radiacao Incidente Total do dia (KW/m2)  \\\n",
       "count                              6914.000000   \n",
       "mean                                210.470264   \n",
       "std                                 275.510678   \n",
       "min                                   0.000000   \n",
       "25%                                 110.600000   \n",
       "50%                                 207.700000   \n",
       "75%                                 274.100000   \n",
       "max                                6589.000000   \n",
       "\n",
       "       Radiacao Refletida Total do dia (KW/m2)  \\\n",
       "count                              6914.000000   \n",
       "mean                                120.182050   \n",
       "std                                 424.527474   \n",
       "min                                   0.000000   \n",
       "25%                                  42.647500   \n",
       "50%                                  65.625000   \n",
       "75%                                  98.500000   \n",
       "max                                6909.000000   \n",
       "\n",
       "       Fluxo positivo de calor no solo Total do dia (KW/m2)  \\\n",
       "count                                        6914.000000      \n",
       "mean                                          108.991185      \n",
       "std                                           716.880089      \n",
       "min                                          -931.000000      \n",
       "25%                                          -201.700000      \n",
       "50%                                            26.230000      \n",
       "75%                                            81.275000      \n",
       "max                                          6924.000000      \n",
       "\n",
       "       Temperatura do Ar Minima do dia (degC)  \\\n",
       "count                             6913.000000   \n",
       "mean                                51.907359   \n",
       "std                                466.418455   \n",
       "min                                  0.000000   \n",
       "25%                                 14.020000   \n",
       "50%                                 16.910000   \n",
       "75%                                 19.040000   \n",
       "max                               6991.000000   \n",
       "\n",
       "       Horario de ocorrencia da Temp. Ar Minima  \\\n",
       "count                               6913.000000   \n",
       "mean                                 791.598438   \n",
       "std                                  656.412518   \n",
       "min                                    0.000000   \n",
       "25%                                  501.000000   \n",
       "50%                                  558.000000   \n",
       "75%                                  652.000000   \n",
       "max                                 2400.000000   \n",
       "\n",
       "       Temperatura do Solo 1 Minima do dia (degC)  \\\n",
       "count                                 6913.000000   \n",
       "mean                                   783.914700   \n",
       "std                                   2249.099354   \n",
       "min                                  -6999.000000   \n",
       "25%                                     15.480000   \n",
       "50%                                     19.770000   \n",
       "75%                                     22.910000   \n",
       "max                                   6999.000000   \n",
       "\n",
       "       Horario de ocorrencia da Temp. Solo 1 minima  Chuva total do dia (mm)  \n",
       "count                                    6912.00000              6763.000000  \n",
       "mean                                      783.52474               252.199053  \n",
       "std                                       611.03749               992.209132  \n",
       "min                                         0.00000                 0.000000  \n",
       "25%                                       559.00000                 0.000000  \n",
       "50%                                       649.00000                 0.000000  \n",
       "75%                                       732.25000                 2.540000  \n",
       "max                                      2400.00000              6858.000000  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_222.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyzing 333 data frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>Cod</th>\n",
       "      <th>Ano</th>\n",
       "      <th>Dia Juliano</th>\n",
       "      <th>Hora - minuto</th>\n",
       "      <th>Veloc. Instantanea do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. Instantanea do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instantanea  (deg)</th>\n",
       "      <th>Umidade relativa Instantanea  (%)</th>\n",
       "      <th>Temperatura do Ar Instantanea (degC)</th>\n",
       "      <th>Pressao</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>333</td>\n",
       "      <td>1997</td>\n",
       "      <td>1.0</td>\n",
       "      <td>700</td>\n",
       "      <td>3344.0</td>\n",
       "      <td>1973.0</td>\n",
       "      <td>89.1</td>\n",
       "      <td>89.7</td>\n",
       "      <td>20.47</td>\n",
       "      <td>942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>333</td>\n",
       "      <td>1997</td>\n",
       "      <td>1.0</td>\n",
       "      <td>900</td>\n",
       "      <td>2618.0</td>\n",
       "      <td>1544.0</td>\n",
       "      <td>185.6</td>\n",
       "      <td>80.2</td>\n",
       "      <td>23.38</td>\n",
       "      <td>944</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Cod   Ano  Dia Juliano Hora - minuto  \\\n",
       "0  333  1997          1.0           700   \n",
       "1  333  1997          1.0           900   \n",
       "\n",
       "   Veloc. Instantanea do Vento a 5 metros (m/s)  \\\n",
       "0                                        3344.0   \n",
       "1                                        2618.0   \n",
       "\n",
       "   Veloc. Instantanea do Vento a 2 metros (m/s)  \\\n",
       "0                                        1973.0   \n",
       "1                                        1544.0   \n",
       "\n",
       "   Direcao do Vento no instantanea  (deg)  Umidade relativa Instantanea  (%)  \\\n",
       "0                                    89.1                               89.7   \n",
       "1                                   185.6                               80.2   \n",
       "\n",
       "   Temperatura do Ar Instantanea (degC) Pressao  \n",
       "0                                 20.47     942  \n",
       "1                                 23.38     944  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_333.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Cod', u'Ano', u'Dia Juliano', u'Hora - minuto',\n",
       "       u'Veloc. Instantanea do Vento a 5 metros (m/s)',\n",
       "       u'Veloc. Instantanea do Vento a 2 metros (m/s)',\n",
       "       u'Direcao do Vento no instantanea  (deg)',\n",
       "       u'Umidade relativa Instantanea  (%)',\n",
       "       u'Temperatura do Ar Instantanea (degC)', u'Pressao'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_333.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([333], dtype=object)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_333.Cod.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[]\n"
     ]
    }
   ],
   "source": [
    "print(data_frame_333[data_frame_333.Cod != 333][columns.YEAR].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Cod                                             0\n",
       "Ano                                             0\n",
       "Dia Juliano                                     0\n",
       "Hora - minuto                                   0\n",
       "Veloc. Instantanea do Vento a 5 metros (m/s)    0\n",
       "Veloc. Instantanea do Vento a 2 metros (m/s)    0\n",
       "Direcao do Vento no instantanea  (deg)          0\n",
       "Umidade relativa Instantanea  (%)               0\n",
       "Temperatura do Ar Instantanea (degC)            0\n",
       "Pressao                                         0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_333.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>Dia Juliano</th>\n",
       "      <th>Veloc. Instantanea do Vento a 5 metros (m/s)</th>\n",
       "      <th>Veloc. Instantanea do Vento a 2 metros (m/s)</th>\n",
       "      <th>Direcao do Vento no instantanea  (deg)</th>\n",
       "      <th>Umidade relativa Instantanea  (%)</th>\n",
       "      <th>Temperatura do Ar Instantanea (degC)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>26001.000000</td>\n",
       "      <td>26001.000000</td>\n",
       "      <td>26001.000000</td>\n",
       "      <td>26001.000000</td>\n",
       "      <td>26001.000000</td>\n",
       "      <td>26001.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>180.482874</td>\n",
       "      <td>1434.967665</td>\n",
       "      <td>1041.404223</td>\n",
       "      <td>300.222295</td>\n",
       "      <td>69.928546</td>\n",
       "      <td>27.642355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>105.504686</td>\n",
       "      <td>1268.984601</td>\n",
       "      <td>1124.401471</td>\n",
       "      <td>727.433045</td>\n",
       "      <td>19.715930</td>\n",
       "      <td>173.852086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-53.360000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.770000</td>\n",
       "      <td>-165.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>89.000000</td>\n",
       "      <td>1.010000</td>\n",
       "      <td>0.763000</td>\n",
       "      <td>96.500000</td>\n",
       "      <td>54.920000</td>\n",
       "      <td>19.140000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>179.000000</td>\n",
       "      <td>1481.000000</td>\n",
       "      <td>1167.000000</td>\n",
       "      <td>158.400000</td>\n",
       "      <td>72.300000</td>\n",
       "      <td>22.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>272.000000</td>\n",
       "      <td>2187.000000</td>\n",
       "      <td>1743.000000</td>\n",
       "      <td>275.000000</td>\n",
       "      <td>85.600000</td>\n",
       "      <td>26.680000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>366.000000</td>\n",
       "      <td>6895.000000</td>\n",
       "      <td>6679.000000</td>\n",
       "      <td>6952.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>6999.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Dia Juliano  Veloc. Instantanea do Vento a 5 metros (m/s)  \\\n",
       "count  26001.000000                                  26001.000000   \n",
       "mean     180.482874                                   1434.967665   \n",
       "std      105.504686                                   1268.984601   \n",
       "min      -53.360000                                      0.000000   \n",
       "25%       89.000000                                      1.010000   \n",
       "50%      179.000000                                   1481.000000   \n",
       "75%      272.000000                                   2187.000000   \n",
       "max      366.000000                                   6895.000000   \n",
       "\n",
       "       Veloc. Instantanea do Vento a 2 metros (m/s)  \\\n",
       "count                                  26001.000000   \n",
       "mean                                    1041.404223   \n",
       "std                                     1124.401471   \n",
       "min                                        0.000000   \n",
       "25%                                        0.763000   \n",
       "50%                                     1167.000000   \n",
       "75%                                     1743.000000   \n",
       "max                                     6679.000000   \n",
       "\n",
       "       Direcao do Vento no instantanea  (deg)  \\\n",
       "count                            26001.000000   \n",
       "mean                               300.222295   \n",
       "std                                727.433045   \n",
       "min                                  0.000000   \n",
       "25%                                 96.500000   \n",
       "50%                                158.400000   \n",
       "75%                                275.000000   \n",
       "max                               6952.000000   \n",
       "\n",
       "       Umidade relativa Instantanea  (%)  Temperatura do Ar Instantanea (degC)  \n",
       "count                       26001.000000                          26001.000000  \n",
       "mean                           69.928546                             27.642355  \n",
       "std                            19.715930                            173.852086  \n",
       "min                             7.770000                           -165.200000  \n",
       "25%                            54.920000                             19.140000  \n",
       "50%                            72.300000                             22.500000  \n",
       "75%                            85.600000                             26.680000  \n",
       "max                           100.000000                           6999.000000  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_333.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyzing 265 data frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>Cod</th>\n",
       "      <th>Ano</th>\n",
       "      <th>Dia Juliano</th>\n",
       "      <th>Hora - minuto</th>\n",
       "      <th>Chuva das 7 h as 7 h</th>\n",
       "      <th>Temperatura do Ar maxima do dia degC</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar maxima</th>\n",
       "      <th>Temperatura do Ar Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar minima</th>\n",
       "      <th>Umidade maxima</th>\n",
       "      <th>Hora</th>\n",
       "      <th>Umidade minima</th>\n",
       "      <th>Hora.1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>265</td>\n",
       "      <td>1997</td>\n",
       "      <td>1.0</td>\n",
       "      <td>700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>31.11</td>\n",
       "      <td>1356</td>\n",
       "      <td>18.50</td>\n",
       "      <td>551.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>265</td>\n",
       "      <td>1997</td>\n",
       "      <td>2.0</td>\n",
       "      <td>700</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.16</td>\n",
       "      <td>1255</td>\n",
       "      <td>19.21</td>\n",
       "      <td>510.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Cod   Ano  Dia Juliano Hora - minuto  Chuva das 7 h as 7 h  \\\n",
       "0  265  1997          1.0           700                   0.0   \n",
       "1  265  1997          2.0           700                   0.0   \n",
       "\n",
       "   Temperatura do Ar maxima do dia degC  \\\n",
       "0                                 31.11   \n",
       "1                                 30.16   \n",
       "\n",
       "  Horario de ocorrencia da Temp. Ar maxima  \\\n",
       "0                                     1356   \n",
       "1                                     1255   \n",
       "\n",
       "   Temperatura do Ar Minima do dia (degC)  \\\n",
       "0                                   18.50   \n",
       "1                                   19.21   \n",
       "\n",
       "   Horario de ocorrencia da Temp. Ar minima  Umidade maxima  Hora  \\\n",
       "0                                     551.0             NaN   NaN   \n",
       "1                                     510.0             NaN   NaN   \n",
       "\n",
       "   Umidade minima  Hora.1  \n",
       "0             NaN     NaN  \n",
       "1             NaN     NaN  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_265.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'Cod', u'Ano', u'Dia Juliano', u'Hora - minuto',\n",
       "       u'Chuva das 7 h as 7 h', u'Temperatura do Ar maxima do dia degC',\n",
       "       u'Horario de ocorrencia da Temp. Ar maxima',\n",
       "       u'Temperatura do Ar Minima do dia (degC)',\n",
       "       u'Horario de ocorrencia da Temp. Ar minima', u'Umidade maxima', u'Hora',\n",
       "       u'Umidade minima', u'Hora.1'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_265.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([265], dtype=object)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_265.Cod.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[]\n"
     ]
    }
   ],
   "source": [
    "print(data_frame_265[data_frame_265.Cod != 265][columns.YEAR].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Cod                                          0\n",
       "Ano                                          0\n",
       "Dia Juliano                                  0\n",
       "Hora - minuto                                0\n",
       "Chuva das 7 h as 7 h                         0\n",
       "Temperatura do Ar maxima do dia degC         0\n",
       "Horario de ocorrencia da Temp. Ar maxima     0\n",
       "Temperatura do Ar Minima do dia (degC)       5\n",
       "Horario de ocorrencia da Temp. Ar minima     5\n",
       "Umidade maxima                              57\n",
       "Hora                                        57\n",
       "Umidade minima                              57\n",
       "Hora.1                                      57\n",
       "dtype: int64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_265.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>Dia Juliano</th>\n",
       "      <th>Chuva das 7 h as 7 h</th>\n",
       "      <th>Temperatura do Ar maxima do dia degC</th>\n",
       "      <th>Temperatura do Ar Minima do dia (degC)</th>\n",
       "      <th>Horario de ocorrencia da Temp. Ar minima</th>\n",
       "      <th>Umidade maxima</th>\n",
       "      <th>Hora</th>\n",
       "      <th>Umidade minima</th>\n",
       "      <th>Hora.1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>6864.000000</td>\n",
       "      <td>6864.000000</td>\n",
       "      <td>6864.000000</td>\n",
       "      <td>6859.000000</td>\n",
       "      <td>6859.000000</td>\n",
       "      <td>6807.000000</td>\n",
       "      <td>6807.000000</td>\n",
       "      <td>6807.000000</td>\n",
       "      <td>6807.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>183.431911</td>\n",
       "      <td>259.223644</td>\n",
       "      <td>30.550498</td>\n",
       "      <td>64.955059</td>\n",
       "      <td>600.501385</td>\n",
       "      <td>92.501716</td>\n",
       "      <td>681.093874</td>\n",
       "      <td>47.692695</td>\n",
       "      <td>1427.192596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>105.584961</td>\n",
       "      <td>1013.569334</td>\n",
       "      <td>119.044092</td>\n",
       "      <td>555.781940</td>\n",
       "      <td>310.560816</td>\n",
       "      <td>7.417340</td>\n",
       "      <td>402.932153</td>\n",
       "      <td>105.982793</td>\n",
       "      <td>245.718799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-53.360000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.660000</td>\n",
       "      <td>-6999.000000</td>\n",
       "      <td>-6999.000000</td>\n",
       "      <td>51.080000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-6999.000000</td>\n",
       "      <td>-6999.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>92.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>26.140000</td>\n",
       "      <td>13.780000</td>\n",
       "      <td>509.000000</td>\n",
       "      <td>88.400000</td>\n",
       "      <td>519.000000</td>\n",
       "      <td>37.235000</td>\n",
       "      <td>1346.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>184.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>28.860000</td>\n",
       "      <td>16.910000</td>\n",
       "      <td>606.000000</td>\n",
       "      <td>94.000000</td>\n",
       "      <td>638.000000</td>\n",
       "      <td>46.550000</td>\n",
       "      <td>1449.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>275.000000</td>\n",
       "      <td>2.540000</td>\n",
       "      <td>31.380000</td>\n",
       "      <td>19.170000</td>\n",
       "      <td>701.000000</td>\n",
       "      <td>99.100000</td>\n",
       "      <td>702.000000</td>\n",
       "      <td>57.210000</td>\n",
       "      <td>1544.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>366.000000</td>\n",
       "      <td>6999.000000</td>\n",
       "      <td>6999.000000</td>\n",
       "      <td>6991.000000</td>\n",
       "      <td>2400.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>2359.000000</td>\n",
       "      <td>5044.000000</td>\n",
       "      <td>2353.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Dia Juliano  Chuva das 7 h as 7 h  \\\n",
       "count  6864.000000           6864.000000   \n",
       "mean    183.431911            259.223644   \n",
       "std     105.584961           1013.569334   \n",
       "min     -53.360000              0.000000   \n",
       "25%      92.000000              0.000000   \n",
       "50%     184.000000              0.000000   \n",
       "75%     275.000000              2.540000   \n",
       "max     366.000000           6999.000000   \n",
       "\n",
       "       Temperatura do Ar maxima do dia degC  \\\n",
       "count                           6864.000000   \n",
       "mean                              30.550498   \n",
       "std                              119.044092   \n",
       "min                               11.660000   \n",
       "25%                               26.140000   \n",
       "50%                               28.860000   \n",
       "75%                               31.380000   \n",
       "max                             6999.000000   \n",
       "\n",
       "       Temperatura do Ar Minima do dia (degC)  \\\n",
       "count                             6859.000000   \n",
       "mean                                64.955059   \n",
       "std                                555.781940   \n",
       "min                              -6999.000000   \n",
       "25%                                 13.780000   \n",
       "50%                                 16.910000   \n",
       "75%                                 19.170000   \n",
       "max                               6991.000000   \n",
       "\n",
       "       Horario de ocorrencia da Temp. Ar minima  Umidade maxima         Hora  \\\n",
       "count                               6859.000000     6807.000000  6807.000000   \n",
       "mean                                 600.501385       92.501716   681.093874   \n",
       "std                                  310.560816        7.417340   402.932153   \n",
       "min                                -6999.000000       51.080000     1.000000   \n",
       "25%                                  509.000000       88.400000   519.000000   \n",
       "50%                                  606.000000       94.000000   638.000000   \n",
       "75%                                  701.000000       99.100000   702.000000   \n",
       "max                                 2400.000000      100.000000  2359.000000   \n",
       "\n",
       "       Umidade minima       Hora.1  \n",
       "count     6807.000000  6807.000000  \n",
       "mean        47.692695  1427.192596  \n",
       "std        105.982793   245.718799  \n",
       "min      -6999.000000 -6999.000000  \n",
       "25%         37.235000  1346.000000  \n",
       "50%         46.550000  1449.000000  \n",
       "75%         57.210000  1544.000000  \n",
       "max       5044.000000  2353.000000  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_frame_265.describe()"
   ]
  },
  {
   "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": "3.8.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}