4.Arima_models.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Disclaimer : This notebook still needs more documentation and refinement!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### importing the relevant libraries and preparing the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "import itertools\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "X = pd.read_csv(\"data/dengue_features_train.csv\")\n",
    "y = pd.read_csv(\"data/dengue_labels_train.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/susmeet/anaconda3/lib/python3.6/site-packages/statsmodels/compat/pandas.py:56: FutureWarning: The pandas.core.datetools module is deprecated and will be removed in a future version. Please use the pandas.tseries module instead.\n",
      "  from pandas.core import datetools\n"
     ]
    }
   ],
   "source": [
    "from statsmodels.tsa.arima_model import ARIMA\n",
    "from statsmodels.tsa.stattools import adfuller\n",
    "import statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#setting a datetime index\n",
    "y.index = pd.DatetimeIndex(X.week_start_date)\n",
    "\n",
    "#seperating the data for the 2 cities\n",
    "sj_cases = y.total_cases[y.city == 'sj']\n",
    "iq_cases = y.total_cases[y.city == 'iq']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#month sampling\n",
    "sj_monthly = sj_cases.resample('M').sum()\n",
    "iq_monthly = iq_cases.resample('M').sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Analysing the auto correlation and partial auto correlation for our data at different lags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f128d821e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(12,8))\n",
    "ax1 = fig.add_subplot(211)\n",
    "fig = sm.graphics.tsa.plot_acf(sj_monthly, lags=40, ax=ax1)\n",
    "ax2 = fig.add_subplot(212)\n",
    "fig = sm.graphics.tsa.plot_pacf(sj_monthly, lags=40, ax=ax2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " - **AR**: There is significant autocorrelation at lags upto 3 or even 4\n",
    " - **I**: Periodicity occurs at every 12th month, which is expected\n",
    " - **MA**: There is a clear partial auto correlation until lag 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tuning the hyperparameters for the Arima model\n",
    "<font color='red'>This section is a work in progress</font>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the p, d and q parameters to take any value between 0 and 2\n",
    "p = d = q = range(0, 2)\n",
    "\n",
    "# Generate all different combinations of p, q and q triplets\n",
    "pdq = list(itertools.product(p, d, q))\n",
    "\n",
    "# Generate all different combinations of seasonal p, q and q triplets\n",
    "# seasonal_pdq = [(x[0], x[1], x[2], 12) for x in list(itertools.product(p, d, q))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "pdq = [(3,0,1),\n",
    "       (3,0,2),\n",
    "       (3,0,3),\n",
    "       (3,0,4),\n",
    "       (4,0,1),\n",
    "       (4,0,2),\n",
    "       (4,0,3),\n",
    "       (4,0,4),\n",
    "       (1,0,3),\n",
    "       (2,0,3),\n",
    "       (1,0,4),\n",
    "       (2,0,4)\n",
    "      ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "seasonal_pdq = [\n",
    "       (3,2,2,12),\n",
    "       (4,2,1,12),\n",
    "       (4,2,2,12),\n",
    "       (2,2,3,12),\n",
    "       (1,2,4,12),\n",
    "       (2,2,4,12)\n",
    "      ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_aic = 100000\n",
    "optimal_pdq = 0\n",
    "optimal_seasonal_pdq = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "warnings.filterwarnings(\"ignore\") # specify to ignore warning messages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ARIMA(3, 0, 1)x(3, 2, 2, 12) - AIC:1948.1216709910318\n",
      "ARIMA(3, 0, 1)x(4, 2, 1, 12) - AIC:1792.98024584508\n",
      "ARIMA(3, 0, 1)x(4, 2, 2, 12) - AIC:1887.5702626520817\n",
      "ARIMA(3, 0, 1)x(2, 2, 3, 12) - AIC:2060.075682342034\n",
      "ARIMA(3, 0, 1)x(1, 2, 4, 12) - AIC:1756.576588342697\n",
      "ARIMA(3, 0, 1)x(2, 2, 4, 12) - AIC:1779.1513850989454\n",
      "ARIMA(3, 0, 2)x(3, 2, 2, 12) - AIC:2041.7954941730493\n",
      "ARIMA(3, 0, 2)x(4, 2, 1, 12) - AIC:1865.7682356187702\n",
      "ARIMA(3, 0, 2)x(4, 2, 2, 12) - AIC:1862.2750852137028\n",
      "ARIMA(3, 0, 2)x(2, 2, 3, 12) - AIC:2058.3515171557874\n",
      "ARIMA(3, 0, 2)x(1, 2, 4, 12) - AIC:1894.244422660541\n",
      "ARIMA(3, 0, 2)x(2, 2, 4, 12) - AIC:1895.304638265609\n",
      "ARIMA(3, 0, 3)x(3, 2, 2, 12) - AIC:2053.2150929897944\n",
      "ARIMA(3, 0, 3)x(4, 2, 1, 12) - AIC:1862.8544904174007\n",
      "ARIMA(3, 0, 3)x(4, 2, 2, 12) - AIC:1887.5840333685974\n",
      "ARIMA(3, 0, 3)x(2, 2, 3, 12) - AIC:2043.7894154512041\n",
      "ARIMA(3, 0, 3)x(1, 2, 4, 12) - AIC:1870.6500639464844\n",
      "ARIMA(3, 0, 3)x(2, 2, 4, 12) - AIC:1882.0531677961656\n",
      "ARIMA(3, 0, 4)x(3, 2, 2, 12) - AIC:2031.8569276997005\n",
      "ARIMA(3, 0, 4)x(4, 2, 1, 12) - AIC:1881.786412413644\n",
      "ARIMA(3, 0, 4)x(4, 2, 2, 12) - AIC:1884.3506420844833\n",
      "ARIMA(3, 0, 4)x(2, 2, 3, 12) - AIC:2027.8005710300704\n",
      "ARIMA(3, 0, 4)x(1, 2, 4, 12) - AIC:1852.8410489980383\n",
      "ARIMA(3, 0, 4)x(2, 2, 4, 12) - AIC:1854.6418254879684\n",
      "ARIMA(4, 0, 1)x(3, 2, 2, 12) - AIC:2043.0510347295115\n",
      "ARIMA(4, 0, 1)x(4, 2, 1, 12) - AIC:1727.009495786157\n",
      "ARIMA(4, 0, 1)x(4, 2, 2, 12) - AIC:1836.853834268281\n",
      "ARIMA(4, 0, 1)x(2, 2, 3, 12) - AIC:1943.8999823437202\n",
      "ARIMA(4, 0, 1)x(1, 2, 4, 12) - AIC:1753.1612396679031\n",
      "ARIMA(4, 0, 1)x(2, 2, 4, 12) - AIC:1878.1550464831332\n",
      "ARIMA(4, 0, 2)x(3, 2, 2, 12) - AIC:2037.038687957473\n",
      "ARIMA(4, 0, 2)x(4, 2, 1, 12) - AIC:1859.733385077086\n",
      "ARIMA(4, 0, 2)x(4, 2, 2, 12) - AIC:1838.3770636641996\n",
      "ARIMA(4, 0, 2)x(2, 2, 3, 12) - AIC:2066.3989966933395\n",
      "ARIMA(4, 0, 2)x(1, 2, 4, 12) - AIC:1892.6744665405035\n",
      "ARIMA(4, 0, 2)x(2, 2, 4, 12) - AIC:1896.6607735318912\n",
      "ARIMA(4, 0, 3)x(3, 2, 2, 12) - AIC:2045.4612513009074\n",
      "ARIMA(4, 0, 3)x(4, 2, 1, 12) - AIC:1880.595816365876\n",
      "ARIMA(4, 0, 3)x(4, 2, 2, 12) - AIC:1865.688592821342\n",
      "ARIMA(4, 0, 3)x(2, 2, 3, 12) - AIC:2049.524357453846\n",
      "ARIMA(4, 0, 3)x(1, 2, 4, 12) - AIC:1877.5521043907058\n",
      "ARIMA(4, 0, 3)x(2, 2, 4, 12) - AIC:1883.259504286094\n",
      "ARIMA(4, 0, 4)x(3, 2, 2, 12) - AIC:2039.0391769991\n",
      "ARIMA(4, 0, 4)x(4, 2, 1, 12) - AIC:1861.2768152497606\n",
      "ARIMA(4, 0, 4)x(4, 2, 2, 12) - AIC:1870.4981824181418\n",
      "ARIMA(4, 0, 4)x(2, 2, 3, 12) - AIC:2027.3512227742353\n",
      "ARIMA(4, 0, 4)x(1, 2, 4, 12) - AIC:1857.5146765936533\n",
      "ARIMA(4, 0, 4)x(2, 2, 4, 12) - AIC:1859.7139185449273\n",
      "ARIMA(1, 0, 3)x(3, 2, 2, 12) - AIC:1972.0887714382882\n",
      "ARIMA(1, 0, 3)x(4, 2, 1, 12) - AIC:1814.047928997457\n",
      "ARIMA(1, 0, 3)x(4, 2, 2, 12) - AIC:1862.8795125288768\n",
      "ARIMA(1, 0, 3)x(2, 2, 3, 12) - AIC:1916.9547037731268\n",
      "ARIMA(1, 0, 3)x(1, 2, 4, 12) - AIC:1747.4481231450031\n",
      "ARIMA(1, 0, 3)x(2, 2, 4, 12) - AIC:1718.9794281451739\n",
      "ARIMA(2, 0, 3)x(3, 2, 2, 12) - AIC:1999.5910661986954\n",
      "ARIMA(2, 0, 3)x(4, 2, 1, 12) - AIC:1779.3143733904813\n",
      "ARIMA(2, 0, 3)x(4, 2, 2, 12) - AIC:1848.6653586127552\n",
      "ARIMA(2, 0, 3)x(2, 2, 3, 12) - AIC:1938.0336605906223\n",
      "ARIMA(2, 0, 3)x(1, 2, 4, 12) - AIC:1880.026095248561\n",
      "ARIMA(2, 0, 3)x(2, 2, 4, 12) - AIC:1880.788349568685\n",
      "ARIMA(1, 0, 4)x(3, 2, 2, 12) - AIC:1964.8795633777997\n",
      "ARIMA(1, 0, 4)x(4, 2, 1, 12) - AIC:1762.2277740792324\n",
      "ARIMA(1, 0, 4)x(4, 2, 2, 12) - AIC:1890.6046060785434\n",
      "ARIMA(1, 0, 4)x(2, 2, 3, 12) - AIC:1970.3683346888786\n",
      "ARIMA(1, 0, 4)x(1, 2, 4, 12) - AIC:1730.7774632913745\n",
      "ARIMA(1, 0, 4)x(2, 2, 4, 12) - AIC:1859.745823627919\n",
      "ARIMA(2, 0, 4)x(3, 2, 2, 12) - AIC:2070.7358299526645\n",
      "ARIMA(2, 0, 4)x(4, 2, 1, 12) - AIC:1900.1673181373217\n",
      "ARIMA(2, 0, 4)x(4, 2, 2, 12) - AIC:1900.174081973942\n",
      "ARIMA(2, 0, 4)x(2, 2, 3, 12) - AIC:2031.8886245633553\n",
      "ARIMA(2, 0, 4)x(1, 2, 4, 12) - AIC:1860.2230526399826\n",
      "ARIMA(2, 0, 4)x(2, 2, 4, 12) - AIC:1866.590358540287\n"
     ]
    }
   ],
   "source": [
    "for param in pdq:\n",
    "    for param_seasonal in seasonal_pdq:\n",
    "        try:\n",
    "            mod = sm.tsa.statespace.SARIMAX(sj_monthly,\n",
    "                                            order=param,\n",
    "                                            seasonal_order=param_seasonal,\n",
    "                                            enforce_stationarity=False,\n",
    "                                            enforce_invertibility=False)\n",
    "\n",
    "            results = mod.fit()\n",
    "            \n",
    "            if results.aic < best_aic:\n",
    "                best_aic = results.aic\n",
    "                optimal_pdq = param\n",
    "                optimal_seasonal_pdq = param_seasonal\n",
    "\n",
    "            print('ARIMA{}x{} - AIC:{}'.format(param, param_seasonal, results.aic))\n",
    "        except:\n",
    "            continue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Following are the optimised hyperparameters for the arima or in this case sarimax model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1718.9794281451739"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_aic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 0, 3)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "optimal_pdq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2, 2, 4, 12)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "optimal_seasonal_pdq"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Fitting the model on optimised hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/susmeet/anaconda3/lib/python3.6/site-packages/statsmodels/base/model.py:496: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                 Statespace Model Results                                 \n",
      "==========================================================================================\n",
      "Dep. Variable:                        total_cases   No. Observations:                  217\n",
      "Model:             SARIMAX(4, 1, 3)x(1, 1, 1, 12)   Log Likelihood               -1279.931\n",
      "Date:                            Fri, 11 May 2018   AIC                           2579.862\n",
      "Time:                                    17:40:47   BIC                           2613.661\n",
      "Sample:                                04-30-1990   HQIC                          2593.516\n",
      "                                     - 04-30-2008                                         \n",
      "Covariance Type:                              opg                                         \n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "ar.L1          1.7038      0.631      2.701      0.007       0.467       2.940\n",
      "ar.L2         -0.7084      1.452     -0.488      0.626      -3.554       2.137\n",
      "ar.L3         -0.0468      1.213     -0.039      0.969      -2.424       2.330\n",
      "ar.L4          0.0143      0.398      0.036      0.971      -0.765       0.794\n",
      "ma.L1         -1.4089      0.648     -2.176      0.030      -2.678      -0.140\n",
      "ma.L2         -0.0685      1.297     -0.053      0.958      -2.611       2.474\n",
      "ma.L3          0.4896      0.686      0.713      0.476      -0.856       1.835\n",
      "ar.S.L12       0.0194      0.121      0.160      0.873      -0.218       0.257\n",
      "ma.S.L12      -0.9312      0.183     -5.100      0.000      -1.289      -0.573\n",
      "sigma2      2.036e+04   3537.207      5.755      0.000    1.34e+04    2.73e+04\n",
      "===================================================================================\n",
      "Ljung-Box (Q):                       30.19   Jarque-Bera (JB):              1723.16\n",
      "Prob(Q):                              0.87   Prob(JB):                         0.00\n",
      "Heteroskedasticity (H):               0.21   Skew:                             1.24\n",
      "Prob(H) (two-sided):                  0.00   Kurtosis:                        17.02\n",
      "===================================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Covariance matrix calculated using the outer product of gradients (complex-step).\n"
     ]
    }
   ],
   "source": [
    "mod = sm.tsa.statespace.SARIMAX(sj_monthly, order=(4,1,3), seasonal_order=(1,1,1,12))\n",
    "results = mod.fit()\n",
    "print (results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = results.predict()  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_absolute_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60.51614779317429"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mean_absolute_error(sj_monthly[:-1], pred.shift(-1).dropna().clip(0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This result is clearly not good enough, we will need to do more data engineering in order to make a useful arima model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### observing the seasonal differences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f1284bcd550>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAELCAYAAAD3HtBMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xd4W+XZ+PHvLcnytmMnTuJMJyEJ2SGYJGwoe7QBWijQ0rTQplDooO8Afl20vLwtHS8tLdDSAqVlBEoZgabQQMsokJA9nOlsxxlO7HhbsqTn98c5cmTHcmINy865P9ely9Kjo6Pbsnzu84zzPGKMQSmllOouV6oDUEop1TdpAlFKKRUTTSBKKaVioglEKaVUTDSBKKWUiokmEKWUUjHRBKKUUiomcScQERkuIv8SkQ0iUiYi37TLC0VkkYhssX8W2OUiIg+JSLmIrBGRGRH7mmtvv0VE5sYbm1JKqeSReC8kFJFioNgYs0JEcoHlwFXAF4FqY8xPRORuoMAYc5eIXA58HbgcmAX8yhgzS0QKgWVAKWDs/ZxqjKmJK0CllFJJ4Yl3B8aYvcBe+369iGwAhgJzgPPszZ4C3gHussv/ZKzMtVhE+tlJ6DxgkTGmGkBEFgGXAs919f4DBgwwJSUl8f4aSinlKMuXLz9ojCmKZx9xJ5BIIlICnAIsAQbZyQVjzF4RGWhvNhTYHfGyCrssWnmXSkpKWLZsWdyxK6WUk4jIznj3kbBOdBHJAf4KfMsYU9fVpp2UmS7KO3uveSKyTESWVVVVdT9YpZRScUtIAhGRNKzk8Ywx5iW7eL/dNBXuJzlgl1cAwyNePgyo7KL8KMaYx4wxpcaY0qKiuGpgSimlYpSIUVgCPA5sMMb8X8RTC4DwSKq5wKsR5V+wR2PNBmrtpq43gYtFpMAesXWxXaaUUqoXSkQfyJnATcBaEVlll/0/4CfACyJyC7ALuNZ+biHWCKxyoAn4EoAxplpE7gOW2tv9KNyhrpRSqveJexhvqpWWlhrtRFdKqe4RkeXGmNJ49qFXoiullIqJJhCllFIx0QTiQEu2HaK+pTXVYSil+jhNIA7T7A9y4x+W8PzS3cfeWCmluqAJxGGaW4MEQ4aaJn+qQ1FK9XGaQBzGFwgC0NASSHEkSqm+ThOIw/gDIQDqfZpAlFLx0QTiMOEEojUQpVS8NIE4jC+cQLQGopSKkyYQh/EHNYEopRJDE4jDaBOWUipRNIE4jE870ZVSCaIJxGG0BqKUShRNIA4TTiDNrUECdn+IUkrFQhOIw/iDwbb7jb5gF1sqpVTXNIE4TLgGAlDv0wkVlVKx0wTiMJEJRIfyKqXioQnEYXyRCUQ70pVScUhIAhGRJ0TkgIisiyi7V0T2iMgq+3Z5xHP3iEi5iGwSkUsiyi+1y8pF5O5ExKba87VrwtIEopSKXaJqIH8ELu2k/EFjzHT7thBARCYC1wOT7Nc8IiJuEXEDDwOXAROBG+xtVQL5tQailEoQTyJ2Yox5T0RKjnPzOcB8Y4wP2C4i5cBM+7lyY8w2ABGZb2+7PhExKos/qH0gSqnESHYfyB0issZu4iqwy4YCkcvhVdhl0cpVAvkDIUSs+1oDUUrFI5kJ5FFgDDAd2Av8wi6XTrY1XZQfRUTmicgyEVlWVVWViFgdwx8IkZeRBmgfiFIqPklLIMaY/caYoDEmBPyeI81UFcDwiE2HAZVdlHe278eMMaXGmNKioqLEB38C8wdCZKS5yEn3aA1EKRWXpCUQESmOeHg1EB6htQC4XkTSRWQUMBb4GFgKjBWRUSLixepoX5Cs+JzKFwiS7nFbCUQvJFRKxSEhnegi8hxwHjBARCqAHwDnich0rGaoHcBXAYwxZSLyAlbneAC43RgTtPdzB/Am4AaeMMaUJSI+dYQ/GMLrceH1uLQTXSkVl0SNwrqhk+LHu9j+fuD+TsoXAgsTEZPqnD8Qwuu2Eki9NmEppeKgV6I7jC9g1UByMzxaA1FKxUUTiMP47QSinehKqXhpAnEYfzBEejiBaA1EKRUHTSAO42u1+kByMjzaB6KUiosmEIfxB0Okp7nI8rppbg1iTKfXaiql1DFpAnGY8CiszDQ3wZChNagJRCkVG00gDhPuRM9IcwPW2uhKKRULTSAOE76QMNNrJZAWTSBKqRhpAnEYqwnLTWa4BuLXBKKUio0mEIcJN2GFE0hLQBOIUio2mkAcJBQybU1YGV6tgSil4qMJxEHCqxGmR9RAtBNdKRUrTSAO0lkC0U50pVSsNIE4iD9gJZDIUVjN/lBXL1FKqag0gThIWwJxaxOWUip+mkAcJLIGohcSKqXipQnEQXydNGG16CgspVSMNIE4SLgGku5xk+Gx/vRaA1FKxSohCUREnhCRAyKyLqKsUEQWicgW+2eBXS4i8pCIlIvIGhGZEfGaufb2W0RkbiJiU0f4g1ay8HpceNwuvG6XJhClVMwSVQP5I3Bph7K7gbeNMWOBt+3HAJcBY+3bPOBRsBIO8ANgFjAT+EE46ajE8EV0ogNkpLn0QkKlVMwSkkCMMe8B1R2K5wBP2fefAq6KKP+TsSwG+olIMXAJsMgYU22MqQEWcXRSUnGI7EQHyPS69ToQpVTMktkHMsgYsxfA/jnQLh8K7I7YrsIui1auEuRIH4idQNLc2oSllIpZKjrRpZMy00X50TsQmSciy0RkWVVVVUKDO5H5OtRAMtLc2oSllIpZMhPIfrtpCvvnAbu8Ahgesd0woLKL8qMYYx4zxpQaY0qLiooSHviJyt+hDyTTqzUQpVTskplAFgDhkVRzgVcjyr9gj8aaDdTaTVxvAheLSIHdeX6xXaYSpG0urLQjTVjaBwLLd1Zz/WMftSVYpdTx8SRiJyLyHHAeMEBEKrBGU/0EeEFEbgF2Adfamy8ELgfKgSbgSwDGmGoRuQ9Yam/3I2NMx455FYejaiBpbupaWlMZUq/w2uq9LN5WzcEGH0P6ZaY6HKX6jIQkEGPMDVGeuqCTbQ1we5T9PAE8kYiY1NE6jsLK8GofCMCq3YcBvahSqe7SK9EdJNyE1ZZAPG5aWp3dbOMPhFhfWQfo4lpKdZcmEAfx2WfYRzrR9Ur0jfvq2hKr0z8LpbpLE4iD+IIhvG4XItaI6Uwdxstqu/kKtAaiVHdpAnGQZn+QrHR32+PwhYRWt5Qzrdpd23ZfayBKdY8mEAdp9AXJ9h4ZN5FhT+nuc/Dw1TUVhxldlA1oDUSp7tIE4iCNvgBZ3vY1EHD2uuj761oYNzAX0BqIUt2lCcRBGv0BstOP1ECcvqytMYZGf5ABuV5AayBKdZcmEAdp8gfJjuwDsWsjTj1w+gIhgiHDgJx0wLmJVKlYaQJxEKsJK6IPxOE1kCY7cRZkeXGJcxOpUrHSBOIgjf4AOZ00YTm1D6TRFwAgO92jU9srFQNNIA7S5Au270Rva8Jy5iisRr+dQLxuMr2ethqJUur4aAJxkAZf5zUQp555N/qs3zsr3UOm1+XYmphSsdIE4hCBYAhfIKR9IBHCTVg56W69Kl+pGGgCcYgmO0l0NgqrxaEHzia7CSvL67GasByaSJWKlSYQh4jsMA4LN2GFD6ROE27CyvZ6yExzOTaRKhUrTSAO0dbeH9GJHu4PqWtxaAIJd6KHm7C0BqJUt2gCcYimthFHR2ogXo+L3HQPNU3+VIWVUm01kHQPmV63Y2tiSsVKE4hDNHTShAVQkO2lptGpCSSASyDd4yIzzeP4xbWU6q6kJxAR2SEia0VklYgss8sKRWSRiGyxfxbY5SIiD4lIuYisEZEZyY7PKZp8R3eiAxRkpVHT5Mx10cNzg4mILq6lVAx6qgZyvjFmujGm1H58N/C2MWYs8Lb9GOAyYKx9mwc82kPxnfAaI0YcRSrI9jq2CaspYnr7zDRtwlKqu1LVhDUHeMq+/xRwVUT5n4xlMdBPRIpTEeCJJtzen9OxCSvLS7VDm7Aa/IG2BbYyvVYTVijk3MW1lOqunkggBviHiCwXkXl22SBjzF4A++dAu3wosDvitRV2mYpT2zUPRzVheTns0Caspogr88NDmp28uJZS3eU59iZxO9MYUykiA4FFIrKxi22lk7KjTgntRDQPYMSIEYmJ8gTXNow37eg+kAZfAH8ghNfjrDEVjf4jc4Nlplm/e5M/0HaBpVKqa0k/YhhjKu2fB4CXgZnA/nDTlP3zgL15BTA84uXDgMpO9vmYMabUGFNaVFSUtNh3Hmpk16GmpO2/JzX6A2SkufC42//JC7KtxZQOO7AfpNEXaOsDCfcNaUe6SjV/IMRLKyowpvc3pyY1gYhItojkhu8DFwPrgAXAXHuzucCr9v0FwBfs0VizgdpwU1cq/NeLa/j6/JWpevuEijxYRirIshJItQMTiLXAlvWZhNeH1wkVVar9fd1evv3CalZX1KY6lGNKdhPWIOBlEQm/17PGmDdEZCnwgojcAuwCrrW3XwhcDpQDTcCXkhxflyqqm9hX13LULLZ9UaMvcFT/B0BBdhoANY3O6wdp9AXahjUfmdZFE4hKrQ176wGobvSlOJJjS+pR0RizDZjWSfkh4IJOyg1wezJjOl7BkGF/vY+QgWU7qjlv/MBjv6gXa/QHO62BFNpNWE4cyhu5QmOWw5f3Vb3Hxn11AH1icIuzek274WCDj6A9pPPj7dUpjiZ+TfZFcx21NWE5bChvKGRoao1ownL41PZOtWTbIZ5fuivVYbSzaZ9VA+kLCaRvt8skUeXhZgBccmIkkAZfkLyMo//c/bKsJiyndaK3BIIYY61GCBGLa2kNxDGW7qhm7pMfYwxce+pwXK7OBoH2rNqmVvbWtgBwuLn3JxCtgUSxz/4jnjW2iNUVh/t852pTlH6cdI+bbK+baof1gYTnBssKXwfi1RqIkzT5A3z5qWUEggZfIMQe+4Qx1cLNVwC1feCkThNIFOGzgDnThtAaNCzbUZPiiOLT5A8eNY1JWEG213E1kKa2K/OtxJGlCcRRyg80UNvcyhfPKAFga1VDagOybbSbr7K9bq2B9GX76lrwelxcOnkwXreLdzcfOPaLerGGiBFHHRVkeR03jLfj3GAZ2oTlKNsPNgJw4cRBAGytakxlOG027qujX1YaYwbm9Ik+EE0gUeytbaE4P4PsdA+zRhfyr01VqQ4pLtE60cGZU7pHrkYI2gfiNNsPNiIC04f3oyArjfIDqa+BBIIhlmyrZsLgPPplebUG0pftq21mcF4GAOeNH0j5gQbe21zFt19YRW0f+MNGavAFaA2aqNeyOHFK98jVCMFaXMvjEm3CcogdBxsZkp9JRpqbMUU5vaIJ6+nFO9l2sJG5Z5TQLzNN+0D6snANBOD88dZ0KTf/cSkvrdjD4m2HUhlat3283Yp32rB+nT7vxBl5myJWIwyzpnQ/MRPIr9/ews/e7GoaOmfZfqiJUQOyARhTlMO2FCeQ6kY/v1i0mbPHDuCSSYPol5WmNZC+KhQy7K9rYXB+JgCjBmRT0j+LNHseqd5Q3e2O9zYfJCPNRWlJQafPF+Wm0+AL9PmRZt1R32L9c0auET+0IJOdh3pHW3gi1Ta18pt/lfPUhzsJBHW2YWMM26saKBmQBcBJA3M42OBP6UCSf5Tto74lwF2XnoyIWDWQ5tZev7yAJpBOHGr00xo0bTUQEeF3N5Xy6h1nMrRfJpv316c4wu55f0sVs0b1b+so7qgoJx2AqvreP3VCoqzYVUNehqetmRJgQnFe2zQSJ5K/rqjAFwjR4AtQVll37Bec4GqaWqlrCTBqQA4AYwZaNZFUNmMt2V7NgJx0Jg3JAyA/y4sxUN/Suxc50wTSiQ17rX+ywflHDi7jB+cyblAuJw3MYfP+vlMD2XO4ma1VjZw9dkDUbQbkWlejH2zouwnkQH0LP3yt7LhqUcYY/rWpirPHFbWbnXhicR776lpOqOY8YwzPfryL0XZzTV9rfk2G8AisUXYNZEyRlUg27UvN/7UxhiXbDjFrVCH2vIH0y7Qv8G3u3d9FTSAdvLu5ilufXs7gvAxKRx7d5DNukNXhFuzlVcuwf2+xRo+dMy76tPcDToAayBvr9vHkBzt4b/OxR8uVVdZRVe/j/A7zm00ots7+wicQJ4KPt1dTfqCBW88dw5iibE0gWB3oACX9raQ6ojCL/tlelu1MzYwTu6ubqaxtYfbowrayIzNE9O5+EE0gHfx44QYG52ew4I4z6W8fWCONHZSLPxBiV3XfWCfkw62HKMpNZ+zAnKjbFOVav+fBht59ttOVcLPih1vbHyCNMUcl+3c2Wdf0nNshqU4ozgVg/QnUzPPsx7vIzfBw5bRiZo3uz9IdNY7vB9l+sBG3SxheaNVARISZowpZsi01CWSxPchl1uj+bWVtCaSXd6RrAonQ4AuweX89n5w6hIERbeORwgfivtAPYlWNq9tVjTvTPzucQPpuDSTcrPjh1oO8u7mKCd97g1PvW8SkH7zJ1HvfbDvrBHh74wGmDstvS5xh/XPSGZSXfsLUQKob/fx97T6uOWUoWV4Ps0f37xX9IKGQ4Uevree11UetFdcjNu2vbzcoBmDWqEL2HG6moqbnTwyXbKumMNvb7iQvP7NvLPSmCSTCmt2HCRk4ZUTnw13BqoFA3xiJtctez2TWqMIut/N6XORnpvXZJixjDJv31+N1u9i8v4EfLiijICuNSyYP5vrTRhAy8OBbmwH4aOshVu46zCenDul0XxOK81h/giSQF5fvxh8MceOskQBtTbKrdh9OZVg8t3QXT3ywnf94YTVrU7Bo0vrKOiYNyW9XNnOUdfbf07WQltYgb23Yz1knDWh3kheugfT2a840gURYscua7+qU4Z0PdwXISfcwtF8m/9p4oNd3ti6xZxGOrBpHMyDH22drIFUNPg43tXLltGIAth1s5FsXjeN/r57C9z85kS+eWcKC1ZUs31nDT/6+geL8DG46fWSn+5pYnEf5gYZe/7c9lkAwxFMf7uS0kgLGD7ZOeorzMxiQ42V1ReoSSPmBBn68cCMzRxXSP8fL7c+uaBtS3RNqGv3sOdzcNtop7OTBueRnprFke8/2Eb22upLa5lZumDmiXXl+pvaB9Gpbqxr495aD7cpW7DrMSQNzyLezfzQ3nzWKlbsP84lfvNMrrmCNprOqcTRFuel9NoFssZuvrj5lKPmZaQwvzOTqU4a2Pf/Vc0aTm+7h049+yOqKWr590bioQ5o/OW0ILpfwtWeW09qH+wr+tnYvew4385WzR7eViQhTh/VLyVl/S2uQR9/ZyhUPvY/bJfzi2mk8dMMpVNQ08cPX1vdYHOHa5cQOCcTlEk4rKeTDrYd6dC3ypxfv5KSBOe060AHS3C5y0j2aQHqre/66ljtfWNV29mOMYeWuGk4ZHr35KuyWs0ax8BtnEwwZfvBq2VFfuJ78AkYTChmWbD/EaSUFXfZ/hA3ISe+zTVjh/qiTB+fxy+un85sbZrRr3+6X5eXVO87ivjmT+O4VE7hmxrCo+5pQnMdPrpnC4m3VfO+Vdb3ib9ldxhh+9+42xhRlc+GEQe2emzosn/Kqhrbp7HvC6t2HueAX7/LAGxs5d1wR/7jzHIYXZnFaSSFfO+8kXlxewRvr9vZILGWVVvLs2IQFcPHEQVTUNLOyh5r4lu+sYXVFLTfNHtnp/2h+ZpoO4+0uEblURDaJSLmI3J2s9/nOFRM42ODj/xZt5unFO7nt6RXUNLUyo5Ohu50ZPziX/7x4PP8uP8hDb5ezbo/1xXxtdSUz7lvUtp5IT3h9TSWvrtrDyl01vL1hP/tqW/j1P8upqGnmkkmDj2sfA3LSe80orEMNvm41IW3e30BBVhoDcrycP34g0zo5CRg1IJubTi/hy2ePxn2MhYOumTGMO84/iflLd/PAG5v6VBLZsr+ez/1hCev31vHVc8cctUjStGH9MAaeWbyTT/z8HT7cejDKnhJjz+FmbnlqGSLw7Jdn8dgXShkUMUDlmxeOZcrQfO55aS0H6pL/P1NWWUdxfkbbUs6RLpsymHSPi5dWVCTlvQ82+NomLTXG8MDfNzIgx8unT+38hGZQXjple+p69dXovWpFQhFxAw8DFwEVwFIRWWCMSXgdd9rwflx36nCe/GAHAMMKMjlnXBEXnHz8a59/fvZIXltdyYNvbebBtzZzzSlDebNsH43+ICt21XD5lGKa/IGo63Akwro9tdzx7Mp2ZS6BkIFrThnarimnK5HTmURr3ukJ1Y1+rnjo34SM4a+3ndE21DKSLxAk3WPFaIxhTcVhxg7KPa6a1vH6j4vHUdPk57fvbqX8QD0/v3Ya/bKOPuikSn1LK08v3sVrqys586T+zB7dn/e3HOTpxTvJTvdw31WTubaTA9PUYdaZ90/e2Igx8I3nVvL6189ud9Fsohyob+FLT36MrzXIc1+Z1TYAJVKa28WDn53Olb9+n2/MX8ndl01g2rD8hP4tI5VV1h3V/xGWm5HGxZMG8/qavXz/ykl4PYk5v65tauWel9fwZtl+CrO9/PmWmZQfaODjHdX8z1WTo05yetPpI7nz+dW8UbaPy6cUJySWROtVCQSYCZQbY7YBiMh8YA6QlEbSuy47GYPhssnFnDe+qNtfWrdLmD9vNruqm3h+6W5+9942CrLSaAmEWF9ZR6bXzVf/tJx/33V+1GHB8frlW5vJy/Dw5JdmUtPoJz8rjX9uPMCBOh/3Xz35uH+nyOlMOjto9wR/IMS3X1hFdZOfdI+LG/+wmAmD89hV3cTe2hYumDCQqnofi7cd4pHPncpFEwfxl2UVlFXW8aM5kxIai4jwP1dNZkxRDv+7cAO/fGsL934qse8Rq9W7D3P7syuoqLE6g5/4YAe/f387IvDZ0uH896Und3qGDdZQ5aH9MtlzuJnvXD6BB9/azLdfWMUzX54V90HbGMOm/fUM7ZfJmopavvPyWvbX+Xh8bmmnySPspIE5/OhTk7n7pTVc9fAHfHLaEH712emIkNBEsru6iW1VDVw+OXqt/JpThvLa6kq++OTH3DhrBFd2GK23puIwgZBhxojja6kA+OHrZfyjbD9zTy/hb2srufKhfxMIGcYOzOH604ZHfd2npg3lN/8s5/8WbWbUgGzGD8rtFcvuRuptCWQosDvicQUwK1lvVpjt5aefmRbXPjxuF6OLcrjn8gmcM66Igiwvdz6/ig1766htbsUfDFFZ25KUBLJ692He2nCA/7hoHKdGNL2dVtL1sN3OhKczqWpITAIxxrBy92FeWbmHFbtqGF6QxU8+PZX8zDRW7z7Ma6srmTW6P+ePt6YTeXXVHu7/2wYO1Pu476rJjB+Uy3dfWcvOQ00M6ZfBhOI8/lG2n4w0N8MLs/jm/JXcdu4YHntvG7NHF/L5WZ2PqoqHiHDzWaN4d3NV0pt6jkdZZS2/emsLb288wOC8DF689XRKSwqpqLES7LhBuW2jd7ry+dkjOdzk5yvnjCbD6+Z7r6xjwepKPjVtCMYQ80HqTx/t5AcLytoe98/28vSXZ7X7bkZz3WnDuXDiIP74wXYe+mc5VfUtbN7fgDGG/jnpHG7yM/f0Er5+wdhj7ssfCPHSigq2H2rEHwhRkOVlZP8sHnp7C7kZaVGbjMCaseEbnziJV1ZVcsezK9mwt44rpgxhx6FGFq7dy+tr9pLtdfPh3Rccc7BNXUsrC9fs5aUVe/jGJ07i2xeP50tnlvD797cxpiiHK6cWt5tKpyO3S/jvS0/m1qeXc9mv3ueyyYN55HMz2pJqazDUrq8vFaQ3te+KyLXAJcaYL9uPbwJmGmO+3mG7ecA8gBEjRpy6c+fOHo+1K9+av5Il26vJzfCweX8Df7p5ZpdTiXS0ZX89A/MyujwY1Le0MufhD6hrbuVf/3keuRnHPnB0ZW1FLZ/8zb957KZTufg4+00609Ia5P+9tJaPth1ib20LmWlupg7LZ8WuGob0yyQ3w8O6PXWIgDEws6SQX1w3jYsefJdxg3L59kXjOG98582I/kAIl1iTXV718AfsrW1h7MAcnvjiaUmtNT36zlYeeGMjy757Ydu0Lz1t56FGrnr4A0SEa0uHces5YyiIUsvojmDIcPUjH7DzUBMel1DvCzB+UC4/v3Za2/Df49HSGuScn/6L4vwMLpo4iGEFWVw6eXC3m0ONMfzvwg388cMdXDRxEPmZadQ0tlLX0sqHWw9x35xJ3HR6SdTXr9p9mP94YRVbqxrxul14Pa62AQNpbuHPt8xi9nEMaw8EQ3z3lXXMX3rkfDbb62bOKUN5dsku7rxwHN+8MHoyW1tRy3W/+4jm1iATi/N45fYzY24S23WoiT8vtmqZD984gyumFvPe5iq++8o6nrp5Ztu09N0lIsuNMaUxvdjW22ogFUBknW4YcNTlqsaYx4DHAEpLS3tPBrRNHJLHK6sq2WuPlqw7znHuB+pauPultfxz4wHGD8rlha+eftRZztqKWh55p5ydh5rYeaiJZ748K+7kAUemM/n9+9vYVd1ETZOf608b0e0D84a9dby0cg9njx3ANy4YyyenDSEn3cNHWw9x3+vr6Zfp5a5LT+bGmSN4dfUevv9qGZ9+9ENCIXj4xhldvl/4H3BQXgZvfPMcWkOhHjmgh4dYfry9OuFt0cYY3tlcxa/f3kKjL8iLt51+1N/TFwhyy1PLMMDLt51BSYwHjM64XcL9V03htmeWM314P4b0y+SVlXu46fElvHjrGYzof3x//+eX7uZAvY9fXj+dM8ZEn7jzWESE71wxkf+8ZHxbPxdYB/Rbn17O914tozVo+NT0IQSCpl3fzRvr9vLN+asYkJPOE18s5fzxAxERfIEgWw80kpPuOe7fx+N28eNrprT1Yw7My2DK0HzS3C4O1LXw5Ifb+fLZo6Ku8vnTNzeSkebisS+cSunIwrj6U0b0z+KuS0/mo22H+MGCdaypOMyTH+xgzMAc8jJSewjvbTUQD7AZuADYAywFbjTGlEV7TWlpqVm2bFkPRXh83t9SxU2Pf9z2+H+vnsKNs0Z08QrL159byT/K9nHjrBE8s3gXU4bl8+xXZrX7R/rs7z5i7Z5ahvTLZN7Zo7muizbU7jDG8NDb5fx58c6260FG9s/ipdvO6HROsGgWrd/PV/60jNdU2BnEAAAgAElEQVTuOIspw44eKtnRt19YxUsr9jDvnNH8v8snxBx/MrUGQ0z/4T+4esZQpg3rx+Sh+W0TL8bKGMM/Nx7gobe3sLqiliH5Geyv93Hl1GJ++dnp7dr+l+2o5jO//YhfXT+dOdOPb1BEPDbvr+e6331Ekz/Ip2cM4wefnNhlTcIXCHLuT99heGEmL3z19KR1gLe0BvnW/FW8UbYPsBYAe/+u8xmQk87m/fVc9qv3mTosnz98obRb39nuWrqjmmt/+xG//Ox0rupkkMpHWw9xw+8X890rJvDliOtw4rVhbx3fmr+KTfvrmTYsn6dunhnXwI4TrgZijAmIyB3Am4AbeKKr5NFbhQ8uaW6hNWiOqwayvrKO11ZXcvv5Y/ivS06mdGQhtz+7ggf+vonvf3IiYB1Ilmyv5vtXTuTms0YlNGYR4ZsXjuVr54+hrrmVndVN3PDYYuY++TG/uWHGcZ/1hq8lCfepHMuP5kxm6tB8PlOamESYDGluF6UlhTyzZBdPL96F1+3i/11+Ml88M7a/QUtrkM/9YQnLd9YwvDCTBz49hatPGcbv3t3KLxZt5vzxA9sdmNbaQ8SPp+klEcYNyuXV28/kt+9u47mPd3HSwBxu6eT7tq2qgTS3i3c2V7GvroWfXzstackDICPNzcOfm8GzS3ZS3djKg29t5rklu7jjEydx3+vryfa6eXzuaVEHECTKKcP7kZHmYnXF4U4TyOP/3s7A3HQ+Pzux/XITivN4885zaPAFyPa6k/pZH69elUAAjDELgYWpjiMeA3LSGZibzqgB2SzfWUNdlPlsmvwBGnwBahpb+f6r68jL8DDv7DEAXDG1mKU7Snjig+2cM24A540fyCPvbKUw28v1M5N3sE1zu+ifk07/nHQe/fwMvjV/FZf96n2e/+pspkZZEjdSuPYSnqDxWHLSPTEfiHvSBRMG8kH5Qf77svF8vL2ae19bT1FuBldM7X6T1jNLdrF8Zw0//NQkbpw1oq0j9Gvnn8Q/Nx3gR6+v59xxRW19HOv21LV9p3rKyP7Z/PiaKWw/2MDv39vG52ePaFcTBvjyU8vYV2f1c80Y0Y8zT0p+gnO7pK0PZNnOav68eCe5GR7e33KQ7185MenJA6zmrUlD8tuu/YpkjGHV7hrOGz8wacPhow37TYVedyHhieJX15/CvZ+aRF5mWtQayFUPf8DM+9/mkl++x5o9tXz3iont+jzuvuxkBuam89cVe2jyB3hn0wGuKx2e1OtKIn3i5EG8eec5BI057plTDzb46JeVlrAx9L3FTbNHsvx7FzHvnDE8+vlTmT68H3e/tIbd3ZzWv9lvTelx+uj+zD2jpN0oGrdL+PE1U6hrbuX+hRvayssqa5kyNC8lZ5y3n38S++paeGnFnnbl9S2tbLNnOD7U6OebF47r8fhuPnMUB+p93PvaemaOKow6v1kyTBmaz7o9dUctFXCg3sfBBj+To1xrcqI5sf7Le5HTx/RnQnEeeRke6pqPnjaitrmVzfsbuHTSYO6bM4nF91xwVH9GRpqb6cP7Ubanlg176wkZjmtIZCIV52cyfVg/lu6oOa7tq+p9KRuplEwi0jYqLs3t4tc3nII/EOKJD7ZHfY0xhj9+sJ1dh44kmflLd3GwwcedF43r9DUnD87j1nPH8OLyCl5bXUlLa5AtBxqYPPTY/UnJcNZJA5g0JI9nlrQf6bhpnzV9zP9dN42/feOso9ZW6Qnnjivii2eU8D9XTea5r8zu0SGtk4fm09waZFuHufDC84yl6u/V0zSBJFm0Gkj4H/CzM4dz0+klUavek4fms+1gY9tKcpOH9vyZTWlJAev21NLsP/ZysQcbfAzI6T1XbCfL8MIsZowoYOmO6NN/f7j1EPe+tp4f//1IbeL1NXuZPDSPmV1Msf/NC8dy6sgC7nlpLa+v2UswZDqdu6kniAhzpg9h3Z66drWt8JopU4f1S1lsLpdw76cm8fnZI485PU2iTbETxNoOzVjrKmsRIe5BFn2FJpAky8tI67QPJPwPOGFw11+0cMJ4cXkF/bO9DE7SFe1dOa2kkEDIHNc6Egcb/BTl9nyMqXBaSQHrK+uiTkz423e3AvBm2T4qapo41OBjxa6aoyY47CjN7eKhG04hI83Ff724GuC4RrQly2WTrX6eN9btayvbsK+e/Mw0ipMwBUpfMKYom8w099EJZE8dY4pyog7vPdFoAkmyvEwPdS1HH2A27qujX1Yag/K6bu6ZbJ/dbT/YyKShyZsjqCszRhQgYo0CO5aD9c6ogQCUlhQSMrBy19HNe2WVtby/5SCfnz0CEeHPH+3knxsPYAzHTCAAQ/tl8syXZ1OQ5aV/tpchKTxQDy/MYtKQPBZGzJi7YW8dJw9O7PxjfYnH7WLikLyjpsYvq6x1TP8HaAJJuo41kCZ/gNZgiA1765kw+NgdowPzMtou8kvVFzM/K43xg3JZurPrfpCW1iD1vsAJ2QfSmVNG9MMlHNU/FL4aPzfdw39dcjKXTBrEnxfv5PF/b2dwXkbUyfw6Gj84l1e+dia/n1ua8gP15VOKWbnrMJWHmwmFDJv21TummSaaCcW5bNpX3zZb88EGH3trWxzT/wGaQJKuYx/IZx79iK/+eTmb9tVzcvHxTRURThxTUvjFnD26P0u2Hepyic3wNSBFDkkguRlpTCjOO6pm9r1X1rG6opafXTuN/Mw0vn/lJEr6Z7NxXz2fmDCwW8lgRP+sbk3clyxX2Ffgv76mkl3VTTT5g0w4zu/viWrcoFzqfQH22dPQh5ulJzoosWoCSbLcdA8trSF8gSBN/gAb9tXxz40HaG4NHrP/Iyx8RpPKM5trZgzFFwh1OZw3fA1IUQ9er5BqpSMLWLnrcNuaDb5AkL8sr+Cm2SO51J71dXB+Bi/edjr/cdE4bjt3TCrDjVnJgGymDe/HKysreXdzFdD5okxOMnaglUA32ytihn92NfPwiUYTSJLl2UM/61sClB9owBjIteevOd4ayE2nj+Rnn5masmnWwar9nDw4l78s2x11m7ar0B1SAwEY0i+T5tYgTa3WCLUmn/VzdFH7K/ezvB6+fsHYlP4N4zVn2hDW763jp29sZPbowuNuijtRjRtkLRW9xV4Rc8v++raFzZxCE0iS5WVayaLOvu4D4JHPzeBbF4497jO4gbkZXJviqT5EhOtKh7O6opaN++o63Sa8ouHxTmNyIsixTwYa7ZFY4USSmcJFuZLlymnFuARaAiHu/dSklPfLpFr/nHT6Z3vZ0lYDqU/4wma9nSaQJMvLOFID2by/Hq/Hxemj+/OtC8f1+Nj1eF000Ro9tHJX58N5uzuNyYkgPK1EvT3Srtlv/cz0nngJZGBuBjefOYpvXzSOk4+z+fVEN3ZQDpsPWB3pW/Y3tNVKnMIZg5VTKNyEVdfSyub99ZxUlNPlIjK9WXialcYo1z1U1Z+Y05h0JZxAwp9Jsz8E0GPTzfS07145MdUh9CrjBuXy8oo97K1tod4XYJyD+j9AayBJF66B1DUH2Lyvvk+foWTbB8VoF84dajwxpzHpSviCsfBn0mTXQLJOwBqIOtpYeyTW+1usgQXhjnWn0ASSZOE+kMrDzVTWtjCuG6u89TZul5CZ5o5aA6lu9PfIbKi9SU7HBBLuA9EE4gjj7RrHkx/sAOjTJ4ix0ASSZOEayLKd1rUC4/r4GUp2uocGX+dzYlU3+imMY4GbvqgtgbT1gVifjdZAnOHUkQVcOGEgG/fV0z/bm9SFrHqjE7OhthfJ8rpxu4R3N1chApNSMBliIuWku6M2YVU3tnLqSIclkPAoLH+4CctOIGn6r+UEbpfwu5tK+dVbm8lw4EmDfsuTTETIy/BQ09TKreeOoTg/M9UhxSU73dNpE5YxhpomP4XZ8a/P3pdEG4WV4dXKvVO4XcK3Lx6f6jBSQhNIDxiUl8GgvAzuvGhsqkOJm9WEdXQCqWsJEAwZChzWhJXuceFxyZFRWK3hJiz911InvqSdJonIvSKyR0RW2bfLI567R0TKRWSTiFwSUX6pXVYuIncnK7ae9oe5pcyfN/uoJUH7opwoNZCaRusiQqd1ootIu6QabsI6ES8kVKqjZJ8mPWiM+XlkgYhMBK4HJgFDgLdEJLw828PARUAFsFREFhhj1ic5xqQbVtB3p6/oKFoTVnWTlUAKHJZAwEqqDb4jnejpHlefu0hUqVikop49B5hvjPEB20WkHJhpP1dujNkGICLz7W37fAI5keREGYXVVgNxWBMW2J9Jy5EaiI7AUk6R7J6+O0RkjYg8ISLhOamHApEz8lXYZdHKVS+Sk975dSDVDm3CAmskVuQoLO3/UE4RVwIRkbdEZF0ntznAo8AYYDqwF/hF+GWd7Mp0Ud7Z+84TkWUisqyqqiqeX0F1U3a6h+bWIMFQ+z9NjYObsLIjaiDNrQEy0nQElnKGuE6VjDEXHs92IvJ74HX7YQUQObXsMCC8yES08o7v+xjwGEBpaWmnSUYlR9vcT/5A20WSAIca/XjdLrId2HyTm+5hT00TYPWBaA1EOUUyR2EVRzy8Glhn318AXC8i6SIyChgLfAwsBcaKyCgR8WJ1tC9IVnwqNtkdrrwOq2n0U5Cd5qiprMOyIy6ubPIHdRoT5RjJPFX6qYhMx2qG2gF8FcAYUyYiL2B1jgeA240xQQARuQN4E3ADTxhjypIYn4pBdofZZ8OqG1sddw1IWE56Go32wILm1qAj+4GUMyUtgRhjburiufuB+zspXwgsTFZMKn456dbZdceLCa2r0J154AxP7xIKGZr8QYYVaA1EOYP29qluCU/p3thhKK/VhOXQBGLPh9XUGqTZHyRT58FSDqEJRHVLx/UvwqqbnDcTb1hkv1CTP6DXgSjH0ASiuiU34+g+kEAwRG1zq3NrIBFJVTvRlZNoAlHdkp3efvpygNrmVoyBwixnzcQbFk4gdS2t+AIhnQdLOYYmENUtHVfgA6hpagWceREhHPlMDtb7AF1MSjmHJhDVLeGJAiOvAzmyCp8zO4/DtbKqBk0gylk0gahuERGyve3nw2qym7OceuAM9wsdqLMSSKZDE6lyHk0gqts6zsgbXkQpw6Ft/1oDUU6lCUR1W8c1QZodvohSuA+krQbi0M9BOY8mENVt2emedqOwjizj6swDZ7rHxYAcL2WVtQA6jFc5hiYQ1W05HdZFDycQpx44RYRTRxawt7YFcG4iVc6jCUR1W8d10cNNWE7tAwEoHVnYdl8TiHIKTSCq27K87nZzYR0ZxuvcA2dpSUHbfR2FpZxCE4jqtvQ0F/5gqO1xc2sQj0tIczv36zRpSD7pHuv3z3JwTUw5i3P/41XMvG4X/sCRBNLkDzp+5JHX42L68H6Ac/uClPNoAlHd5vW0TyAtrTqBIMA544oozPa21USUOtFpY63qNq/n6CYsTSDw1XNGc+PMEY5c1lc5k54qqW7zut0EQ4ZgyADahBXmcbscO6Gkcqa4EoiIXCsiZSISEpHSDs/dIyLlIrJJRC6JKL/ULisXkbsjykeJyBIR2SIiz4uI/if2Ul67iSbcjKVNWEo5U7w1kHXANcB7kYUiMhG4HpgEXAo8IiJuEXEDDwOXAROBG+xtAR4AHjTGjAVqgFvijE0lSccE0qw1EKUcKa4EYozZYIzZ1MlTc4D5xhifMWY7UA7MtG/lxphtxhg/MB+YI1aj8SeAF+3XPwVcFU9sKnnCCcQXsK7/aPIHHX0NiFJOlaw+kKHA7ojHFXZZtPL+wGFjTKBDueqF0t3hBHKkCcvJV6Er5VTHHIUlIm8Bgzt56jvGmFejvayTMkPnCct0sX20mOYB8wBGjBgRbTOVJOlpdhOWPRJLO9GVcqZjJhBjzIUx7LcCGB7xeBhQad/vrPwg0E9EPHYtJHL7zmJ6DHgMoLS0NGqiUcnhdXfoA2nVJiylnChZTVgLgOtFJF1ERgFjgY+BpcBYe8SVF6ujfYExxgD/Aj5jv34uEK12o1LsqE701iAZmkCUcpx4h/FeLSIVwOnA30TkTQBjTBnwArAeeAO43RgTtGsXdwBvAhuAF+xtAe4Cvi0i5Vh9Io/HE5tKnrYEEgwRDBn8gZA2YSnlQHFdiW6MeRl4Ocpz9wP3d1K+EFjYSfk2rFFaqpeLbMJy+mJSSjmZXomuui2yCcvpy9kq5WSaQFS3HbkOJCKB6BoYSjmOJhDVbekRfSBty9lqDUQpx9EEorrN67aSha81GLEeun6VlHIa/a9X3RZ5IWGT35o8IDNNm7CUchpNIKrbIkdhtbTVQLQJSymn0QSiui1yFFaTjsJSyrE0gahu62wYr14HopTzaAJR3eZxCSJWH0i4CUtn41XKeTSBqG4TEbxuV/smLK2BKOU4mkBUTLwel3UhoV4HopRjaQJRMUn3uNouJPR6XLhdnS3popQ6kWkCUTEJN2HpeuhKOZde/aVikp7mxhcIIegILKWcShOIiolVAwkSChntQFfKoTSBqJh4PfYorFCQ3Iy0VIejlEoBTSAqJl67E73RFyQvQ79GSjmRdqKrmIQ70etaWsnTGohSjhTvmujXikiZiIREpDSivEREmkVklX37bcRzp4rIWhEpF5GHRETs8kIRWSQiW+yfBfHEppIr3IRV3xIgL1NrIEo5Ubw1kHXANcB7nTy31Rgz3b7dGlH+KDAPGGvfLrXL7wbeNsaMBd62H6teKnwhYV1zq/aBKOVQcSUQY8wGY8ym491eRIqBPGPMR8YYA/wJuMp+eg7wlH3/qYhy1Qt5PS4a/QF8gZD2gSjlUMnsAxklIitF5F0ROdsuGwpURGxTYZcBDDLG7AWwfw5MYmwqTuluF4ca/ABaA1HKoY556igibwGDO3nqO8aYV6O8bC8wwhhzSEROBV4RkUlAZ/NdmOOO9khM87CawRgxYkR3X64SID3N1TaRovaBKOVMx/zPN8Zc2N2dGmN8gM++v1xEtgLjsGocwyI2HQZU2vf3i0ixMWav3dR1oIv9PwY8BlBaWtrtBKTiF16VECA3XWsgSjlRUpqwRKRIRNz2/dFYneXb7KapehGZbY+++gIQrsUsAOba9+dGlKteKLyoFEBepiYQpZwo3mG8V4tIBXA68DcRedN+6hxgjYisBl4EbjXGVNvP3Qb8ASgHtgJ/t8t/AlwkIluAi+zHqpeKTCC52omulCPF9Z9vjHkZeLmT8r8Cf43ymmXA5E7KDwEXxBOP6jle95H5r7QGopQz6ZXoKiZaA1FKaQJRMQknEBHI8WoCUcqJNIGomIQTSE66B5euRqiUI2kCUTFJt4fx6kSKSjmXJhAVk/Q066uj/R9KOZcmEBWT8IWEOgJLKefSBKJiEu4D0YkUlXIuTSAqJkcSiNZAlHIqTSAqJuEmLO0DUcq5NIGomLTVQLQPRCnH0gSiYhJOIFoDUcq5NIGomKRrH4hSjqcJRMVkRGE2nzl1GGeNHZDqUJRSKaLtDyomXo+Ln187LdVhKKVSSGsgSimlYqIJRCmlVEw0gSillIqJJhCllFIxiXdN9J+JyEYRWSMiL4tIv4jn7hGRchHZJCKXRJRfapeVi8jdEeWjRGSJiGwRkedFxBtPbEoppZIr3hrIImCyMWYqsBm4B0BEJgLXA5OAS4FHRMQtIm7gYeAyYCJwg70twAPAg8aYsUANcEucsSmllEqiuBKIMeYfxpiA/XAxMMy+PweYb4zxGWO2A+XATPtWbozZZozxA/OBOSIiwCeAF+3XPwVcFU9sSimlkiuRfSA3A3+37w8Fdkc8V2GXRSvvDxyOSEbhcqWUUr3UMS8kFJG3gMGdPPUdY8yr9jbfAQLAM+GXdbK9ofOEZbrYPlpM84B59sMGEdkUbdsORgC7jnPbaPKB2jj3kahYQOM5Fo2na4mIpzfFAhrPsYTjGRn3nowxcd2AucBHQFZE2T3APRGP3wROt29vdtwOK4EcBDx2ebvtEnUDqhKwj8d6Sywaj8bTG+LpTbFoPD0XjzEm7lFYlwJ3AZ8yxjRFPLUAuF5E0kVkFDAW+BhYCoy1R1x5sTraFxjrt/oX8Bn79XOBV+OJLYrDCdjHawnYByQmFtB4jkXj6Voi4ulNsYDGcyyJiifuubB+A6QDi6x+cBYbY241xpSJyAvAeqymrduNMUEAEbkDq0biBp4wxpTZ+7oLmC8i/wOsBB6PM7bOxF39M8Yk6o+YiKqoxnMMGk/XEhRPb4oFNJ5jSUg8EGcCMcac1MVz9wP3d1K+EFjYSfk2rFFayfRYkvffHb0pFtB4jkXjia43xQIaz7EkLB6x28SUUkqpbtGpTJRSSsWkTycQEXlCRA6IyLqIsmki8pGIrBWR10Qkzy73isiTdvlqETkv4jX3i8huEWnoJfG8YZeVichv7Sv4UxnPO/b0M6vs28BUxCIiuRExrBKRgyLyyxR/Np+1p/IpE5GfxhKLvZ/hIvIvEdlg7+ubdnmhiCyyp/hZJCIFdrmIyENiTQm0RkRmROzrDRE5LCKvpzoeERkpIsvtv1eZiNya4s8mGPH9WZDiz+b8Dt/nFhHp9gXUCf58HhCRdfbts8d880QN50rFDTgHmAGsiyhbCpxr378ZuM++fzvwpH1/ILAccNmPZwPFQEMviSfP/inAX4HrUxzPO0Bpb/hsOuxzOXBOquLBugB2F1BkP/cUcEGM8RQDM+z7uVhTA00EfgrcbZffDTxg378c68Jdsb+/SyL2dQHwSeD1OP5eCYkH8ALp9v0cYAcwJIWfTVz/44mOJ2KfhUA1EZdDpOBvdQXW9FQeIBtYhn0sivre8X6Yqb4BJbQ/CNRxpG9nOLDevv8w8PmI7d4GZnbYVyK+XImMJw1r6N5nUxkPCUggSfhsxmLNaiCpigc4DXgrovwm4JEEfa9fBS4CNgHFdlkxsMm+/zvghojt27azH59HHAkk0fHYZeGE260EkshYSMD/eJI+m3nAM6mMB/gv4LsR5Y8D13X1Xn26CSuKdcCn7PvXYh0IAFZjzbvlEevalFMjnut18YjIm8ABoJ4jc4SlLB7gSbua/T0R6WzmgJ6MBeAG4Hljf9NTFE85cLKIlIiIB2v+tri/UyJSApwCLAEGGWP2Atg/w82H0aYFSrh447GbWNbYzz9gjKlMVSxAhogsE5HFsTQXJSGesOuB51Icz2rgMhHJEpEBwPkc4/t8IiaQm4HbRWQ5VnXOb5c/gfVBLQN+CXyIdY1Kr4zHGHMJ1llBOtZEk6mM53PGmCnA2fbtphTGEpaQf7h44jHG1AC3Ac8D72M1z8T1nRKRHKxmy28ZY+q62rSTsoQPqUxEPMaY3caasfskYK6IDEpVLMAIY0wpcCPwSxEZE0ssCYwHESkGpmBdHxezeOMxxvwD6xKLD7H+tz7iGN/neC8k7HWMMRuBiwFEZBxWux7GmqjxzvB2IvIhsKU3x2OMabE7+uZgtU2mJB5jzB77Z72IPIvVfPOnVMRiP56GNe3N8nhjiDceY13c9ZpdPg8Ixvr+IpKGdQB4xhjzkl28X0SKjTF77QPNAbu8gvZnh8OAmM/seyIeY0yliJRhnYR0q1adqFjCtR9jzDYReQfrbH1rd2JJZDy264CXjTGt3Y0j0fGYiOv37P/1Lo+RJ1wNROwRQiLiAr4L/NZ+nCUi2fb9i7DOINf3tnhEJMf+Y2M3i1wObExhPB67Ohv+kl6J1dTT47FEvPQGEl/7iCmeiNcUAF8D/hDjewtWm/MGY8z/RTy1AGtqH2g/xc8C4Av2iJrZQG24uSIREhWPiAwTkUx7nwXAmVht7qmIpUBE0u19DrBj6fYxIAl/q7i+zwn8fNwi0t/e51RgKvCPLt880R1KPXmzP/S9QCtWVr0F+CbWKITNwE840ilagvXF3QC8BYyM2M9P7deH7J/3pioeYBDWaKA1QBnwa+xJJlMUTzbWqKNwPL8C3Kn6W9nPbwNO7iXfneewDkLriXG0nL2fs7CaNdYAq+zb5Vgdz29jnQm+DRTa2wtW5/5WYC0RgxywmtOqgGb7d7skVfFgdeauwWpfXwPMS2EsZ9iPV9s/b+kFf6sSYA+djDJMweeTEfFdXgxMP9Z765XoSimlYnLCNWEppZTqGZpAlFJKxUQTiFJKqZhoAlFKKRUTTSBKKaVioglEKaVUTDSBKGUTa9r60jhe/0URGRLD664SkYnd2L5EIqah72KbG7sbi1LdoQlEqcT5ItCtBCJHJmE87gRynEqw5ntSKmk0gag+S0T+W0S+Yd9/UET+ad+/QESeFpGLxVogaoWI/EWsyeYQkVNF5F2xFjp6Mzx1TMR+XSLylIj8T5T3dYvIH8VadGetiNwpIp8BSoFnxJq1OFNEvi8iS+3tHrOnnAjXdP5XRN4F7sKaAfhn9us6ndzPjnm1iHyEtT5JuLxERN63f8cVInKG/dRPgLPtfd5px/wzO541IvLV2D95pWzxTAehN72l8oa1GM5f7PvvAx9jraHyA6wD83tAtv38XcD37ec/5MgiUJ8FnrDvv2Pv8zngO12876nAoojH/SJeHzlNRWHE/T8Dn4zY7pGI5/4IfOYYv+sajix29TPsdUyALCDDvj8WWGbfP4+I9UCw1pv4rn0/HWtm4VGp/hvqrW/fTrjZeJWjLAdOFZFcwAeswKoFnI01YdxE4AP7xN+LNT31eGAysMgud2PNiRX2O+AFY81KGs02YLSI/Br4G9EnnDtfRP4b6yBfiDWX2Gv2c88f7y8pIvlYSepdu+jPwGX2/TTgNyIyHWsm4HFRdnMxMNWuKQHkYyWc7ccbh1IdaQJRfZYxplVEdgBfwqpVrMFaBGcM1oFxkTHmhsjXiMgUoMwYc3qU3X6IdeD/hTGmJcr71og1pfwlWM1J12GtJRL5PhnAI1g1kt0ici/WZHVhjd34VYXoa33cCewHpmE1SXcas72Prxtj4lpzQqlI2gei+rr3gP+0f74P3Io1G+li4Hi/iLUAAAFVSURBVEwROQnapmQfhzWrbpGInG6Xp4nIpIj9PY61qM5f7A7uo9hTgbuMMX8Fvoe1tjpYq0fm2vfDyeKg3ffyGaKLfN1RjDGHgVoROcsu+lzE0/nAXmNMCGuhL3eUfb4J3CbWlPyIyDixp6hXKlaaQFRf9z7Wyo0fGWP2Y52Bv2+MqcIaFfWcWMupLsaaAt6PdTB/QERWYyWbMyJ3aKw1FVYAfxZrbZCOhgLviMgqrP6Le+zyPwK/tct9wO+xpst+BWuK/mjmA/8lIiujdaJj1bIetjvRmyPKH8Fa5W8xVvNVuGazBgjYHe93Yq1Tsh5YYQ8B/h3aAqHipNO5K6WUionWQJRSSsVEq7BKdUFElmANe410kzFmbZLe72GspVYj/coY82Qy3k+peGgTllJKqZhoE5ZSSqmYaAJRSikVE00gSimlYqIJRCmlVEw0gSillIrJ/wcNIOgcAU3g3AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1284bb4ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sj_monthly.diff(12).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f128412f630>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1284148438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "iq_monthly.diff(12).plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These plots clearly have a pattern to them.\n",
    "**Next Step** shall be to utilise this pattern in creating a powerful arima model."
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}