scratch_work_bennet.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from core.preprocessing import make_small_train, make_small_dev, StatisticalClearSky, DataManager, make_batch, CenterScale\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "df = pd.read_pickle('data/master_dataset.pkl').fillna(0)\n",
    "df = df.loc['2015-07-15':'2017-07-14']\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "sns.set(style='darkgrid', palette='colorblind', context='talk')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_trends_and_total(df, start=None, stop=None):\n",
    "    if start is None:\n",
    "        start = df.index[0]\n",
    "    if stop is None:\n",
    "        stop = df.index[-1]\n",
    "    cols = df.columns\n",
    "    fig, axes = plt.subplots(nrows=2, figsize=(12,10), sharex=True)\n",
    "    df.loc[start:stop].plot(y=cols[:-1], legend=False, linewidth=1, ax=axes[0])\n",
    "    df.loc[start:stop].plot(y=cols[-1], linewidth=1, ax=axes[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "mixed = make_small_train(df, 'mixed')\n",
    "plot_trends_and_total(mixed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "sunny = make_small_train(df, 'sunny')\n",
    "plot_trends_and_total(sunny)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "cloudy = make_small_train(df, 'cloudy')\n",
    "plot_trends_and_total(cloudy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "comb = make_small_train(df, 'combined')\n",
    "plot_trends_and_total(comb)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "dev = make_small_dev(df)\n",
    "plot_trends_and_total(dev)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "from core.problems import ManyToOneRegression\n",
    "\n",
    "# Problem parameters\n",
    "window = 5*12 # window size (to be determined from correlation analysis)\n",
    "future = 3*12  # prediction window into the future\n",
    "\n",
    "# Optimization parameters\n",
    "nepochs = 50\n",
    "batchsize = 1000\n",
    "\n",
    "problem = ManyToOneRegression(mixed, window=window, future=future)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "feat = problem.featurize(100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "%time np.array([mixed.iloc[i:i+window, :-1].values.ravel() for i in xrange(len(mixed) - problem.window - problem.future)]).shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "mixed.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "i = 75\n",
    "mixed.iloc[i:i+window, :-1].values.ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dm = DataManager()\n",
    "dm.load_all_and_split(reindex=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scale = CenterScale(y_data=-1)\n",
    "scale.fit(dm.original_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "test = scale.apply(dm.original_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test['S03'].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_trends_and_total(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dm.swap_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "present = 3*12\n",
    "future = 3*12\n",
    "\n",
    "dm = DataManager()\n",
    "dm.load_all_and_split(kind='all')\n",
    "dftrain = dm.original_train\n",
    "dfdev = dm.original_dev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X, Y = make_batch(dftrain, 100, present, future, exo=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X[4,-5:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for s in xrange(20):\n",
    "    plt.plot(X[52,s*present:(s+1)*present])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "[plt.plot(Y[52])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dftrain.index.time[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from scipy.signal import fftconvolve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def std_window_1(signal, window):\n",
    "    s = fftconvolve(signal, np.ones(window), mode='same')\n",
    "    q = fftconvolve(np.power(signal, 2), np.ones(window), mode='same')\n",
    "    var = (q - np.power(s, 2) *1./ window) / (window - 1)\n",
    "    stdev = np.sqrt(var)\n",
    "    return stdev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def std_window_2(signal, window):\n",
    "    if isinstance(signal, pd.core.series.Series):\n",
    "        stdev = signal.rolling(window=window, center=True).std()\n",
    "    else:\n",
    "        series = pd.Series(data=signal)\n",
    "        stdev = series.rolling(window=window, center=True).std()\n",
    "    return stdev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def totalvar(arr):\n",
    "    diff = arr[:-1] - arr[1:]\n",
    "    return np.sum(np.abs(diff))\n",
    "\n",
    "def totalvar_window(signal, window):\n",
    "    if isinstance(signal, pd.core.series.Series):\n",
    "        stdev = signal.rolling(window=window, center=True).apply(totalvar)\n",
    "    else:\n",
    "        series = pd.Series(data=signal)\n",
    "        stdev = series.rolling(window=window, center=True).std()\n",
    "    return stdev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%timeit std_window_1(dm.original_train['total_power'], 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%timeit std_window_2(dm.original_train['total_power'], 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.plot(std_window_2(dm.original_full['total_power'].iloc[100*3*2:100*3*5], window=24), linewidth=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.plot(totalvar_window(dm.detrended_full['total_power'].iloc[100*3*2:100*3*5], window=24), linewidth=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plot_trends_and_total(dm.original_full.iloc[100*3*2:100*3*5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dm.make_error_target()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "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>error_target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>210528 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                    error_target\n",
       "2015-07-15 00:00:00          NaN\n",
       "2015-07-15 00:05:00          NaN\n",
       "2015-07-15 00:10:00          NaN\n",
       "2015-07-15 00:15:00          NaN\n",
       "2015-07-15 00:20:00          NaN\n",
       "2015-07-15 00:25:00          NaN\n",
       "2015-07-15 00:30:00          NaN\n",
       "2015-07-15 00:35:00          NaN\n",
       "2015-07-15 00:40:00          NaN\n",
       "2015-07-15 00:45:00          NaN\n",
       "2015-07-15 00:50:00          NaN\n",
       "2015-07-15 00:55:00          NaN\n",
       "2015-07-15 01:00:00          NaN\n",
       "2015-07-15 01:05:00          NaN\n",
       "2015-07-15 01:10:00          NaN\n",
       "2015-07-15 01:15:00          NaN\n",
       "2015-07-15 01:20:00          NaN\n",
       "2015-07-15 01:25:00          NaN\n",
       "2015-07-15 01:30:00          NaN\n",
       "2015-07-15 01:35:00          NaN\n",
       "2015-07-15 01:40:00          NaN\n",
       "2015-07-15 01:45:00          NaN\n",
       "2015-07-15 01:50:00          NaN\n",
       "2015-07-15 01:55:00          NaN\n",
       "2015-07-15 02:00:00          NaN\n",
       "2015-07-15 02:05:00          NaN\n",
       "2015-07-15 02:10:00          NaN\n",
       "2015-07-15 02:15:00          NaN\n",
       "2015-07-15 02:20:00          NaN\n",
       "2015-07-15 02:25:00          NaN\n",
       "...                          ...\n",
       "2017-07-14 21:30:00          NaN\n",
       "2017-07-14 21:35:00          NaN\n",
       "2017-07-14 21:40:00          NaN\n",
       "2017-07-14 21:45:00          NaN\n",
       "2017-07-14 21:50:00          NaN\n",
       "2017-07-14 21:55:00          NaN\n",
       "2017-07-14 22:00:00          NaN\n",
       "2017-07-14 22:05:00          NaN\n",
       "2017-07-14 22:10:00          NaN\n",
       "2017-07-14 22:15:00          NaN\n",
       "2017-07-14 22:20:00          NaN\n",
       "2017-07-14 22:25:00          NaN\n",
       "2017-07-14 22:30:00          NaN\n",
       "2017-07-14 22:35:00          NaN\n",
       "2017-07-14 22:40:00          NaN\n",
       "2017-07-14 22:45:00          NaN\n",
       "2017-07-14 22:50:00          NaN\n",
       "2017-07-14 22:55:00          NaN\n",
       "2017-07-14 23:00:00          NaN\n",
       "2017-07-14 23:05:00          NaN\n",
       "2017-07-14 23:10:00          NaN\n",
       "2017-07-14 23:15:00          NaN\n",
       "2017-07-14 23:20:00          NaN\n",
       "2017-07-14 23:25:00          NaN\n",
       "2017-07-14 23:30:00          NaN\n",
       "2017-07-14 23:35:00          NaN\n",
       "2017-07-14 23:40:00          NaN\n",
       "2017-07-14 23:45:00          NaN\n",
       "2017-07-14 23:50:00          NaN\n",
       "2017-07-14 23:55:00          NaN\n",
       "\n",
       "[210528 rows x 1 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dm.error_full"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x114849d50>]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAHHCAYAAAAoFvU6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nn+QZ09XhbuIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11472d690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(dm.detrended_full['total_power'][:1200])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x114783610>]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAogAAAG4CAYAAADG2xaKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfxTFT3R3htX6AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110e9e250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(dm.error_full[:1200])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from core.utilities import totalvar_window"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>error_target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:00:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 00:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:00:00</th>\n",
       "      <td>3.751478e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:05:00</th>\n",
       "      <td>2.295213e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:10:00</th>\n",
       "      <td>2.018971e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:15:00</th>\n",
       "      <td>1.015011e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:20:00</th>\n",
       "      <td>1.105136e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:25:00</th>\n",
       "      <td>1.105136e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:30:00</th>\n",
       "      <td>5.525679e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:35:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:40:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:45:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:50:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 01:55:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:00:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:05:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:10:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:15:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:20:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-15 02:25:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:30:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:35:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:40:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:45:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:50:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 21:55:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:00:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:05:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:10:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:15:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:20:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:25:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:30:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:35:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:40:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:45:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:50:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 22:55:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:00:00</th>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:05:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:10:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:15:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:20:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:25:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:30:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:35:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:40:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:45:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:50:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-07-14 23:55:00</th>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>210528 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     error_target\n",
       "2015-07-15 00:00:00           NaN\n",
       "2015-07-15 00:05:00           NaN\n",
       "2015-07-15 00:10:00           NaN\n",
       "2015-07-15 00:15:00           NaN\n",
       "2015-07-15 00:20:00           NaN\n",
       "2015-07-15 00:25:00           NaN\n",
       "2015-07-15 00:30:00           NaN\n",
       "2015-07-15 00:35:00           NaN\n",
       "2015-07-15 00:40:00           NaN\n",
       "2015-07-15 00:45:00           NaN\n",
       "2015-07-15 00:50:00           NaN\n",
       "2015-07-15 00:55:00           NaN\n",
       "2015-07-15 01:00:00  3.751478e-13\n",
       "2015-07-15 01:05:00  2.295213e-13\n",
       "2015-07-15 01:10:00  2.018971e-13\n",
       "2015-07-15 01:15:00  1.015011e-13\n",
       "2015-07-15 01:20:00  1.105136e-15\n",
       "2015-07-15 01:25:00  1.105136e-15\n",
       "2015-07-15 01:30:00  5.525679e-16\n",
       "2015-07-15 01:35:00  0.000000e+00\n",
       "2015-07-15 01:40:00  0.000000e+00\n",
       "2015-07-15 01:45:00  0.000000e+00\n",
       "2015-07-15 01:50:00  0.000000e+00\n",
       "2015-07-15 01:55:00  0.000000e+00\n",
       "2015-07-15 02:00:00  0.000000e+00\n",
       "2015-07-15 02:05:00  0.000000e+00\n",
       "2015-07-15 02:10:00  0.000000e+00\n",
       "2015-07-15 02:15:00  0.000000e+00\n",
       "2015-07-15 02:20:00  0.000000e+00\n",
       "2015-07-15 02:25:00  0.000000e+00\n",
       "...                           ...\n",
       "2017-07-14 21:30:00  0.000000e+00\n",
       "2017-07-14 21:35:00  0.000000e+00\n",
       "2017-07-14 21:40:00  0.000000e+00\n",
       "2017-07-14 21:45:00  0.000000e+00\n",
       "2017-07-14 21:50:00  0.000000e+00\n",
       "2017-07-14 21:55:00  0.000000e+00\n",
       "2017-07-14 22:00:00  0.000000e+00\n",
       "2017-07-14 22:05:00  0.000000e+00\n",
       "2017-07-14 22:10:00  0.000000e+00\n",
       "2017-07-14 22:15:00  0.000000e+00\n",
       "2017-07-14 22:20:00  0.000000e+00\n",
       "2017-07-14 22:25:00  0.000000e+00\n",
       "2017-07-14 22:30:00  0.000000e+00\n",
       "2017-07-14 22:35:00  0.000000e+00\n",
       "2017-07-14 22:40:00  0.000000e+00\n",
       "2017-07-14 22:45:00  0.000000e+00\n",
       "2017-07-14 22:50:00  0.000000e+00\n",
       "2017-07-14 22:55:00  0.000000e+00\n",
       "2017-07-14 23:00:00  0.000000e+00\n",
       "2017-07-14 23:05:00           NaN\n",
       "2017-07-14 23:10:00           NaN\n",
       "2017-07-14 23:15:00           NaN\n",
       "2017-07-14 23:20:00           NaN\n",
       "2017-07-14 23:25:00           NaN\n",
       "2017-07-14 23:30:00           NaN\n",
       "2017-07-14 23:35:00           NaN\n",
       "2017-07-14 23:40:00           NaN\n",
       "2017-07-14 23:45:00           NaN\n",
       "2017-07-14 23:50:00           NaN\n",
       "2017-07-14 23:55:00           NaN\n",
       "\n",
       "[210528 rows x 1 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = pd.DataFrame(data=totalvar_window(dm.detrended_full['total_power'], 24))\n",
    "df1.columns= ['error_target']\n",
    "df1"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}