code/main-result.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook produces the main result for Gemello. This notebook has already computed the result for Latent bayesian melding and FHMM and stored in a Pandas.Panel. We just pull the data from that Panel. Additionally, this notebook computes the main result. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "from create_df import read_df\n",
    "df, dfc, all_homes, appliance_min, national_average = read_df()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "run all_functions.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "run features.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r--  1 nipunbatra  staff   436K Feb  9 18:57 accuracy.h5\r\n"
     ]
    }
   ],
   "source": [
    "!ls -lah accuracy.h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "accuracy_store = pd.HDFStore(\"accuracy.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<class 'pandas.io.pytables.HDFStore'>\n",
       "File path: accuracy.h5\n",
       "/FHMM                        wide_table   (typ->appendable,nrows->510,ncols->6,indexers->[major_axis,minor_axis])\n",
       "/LBM                         wide_table   (typ->appendable,nrows->510,ncols->6,indexers->[major_axis,minor_axis])\n",
       "/N-NILM                      wide_table   (typ->appendable,nrows->445,ncols->6,indexers->[major_axis,minor_axis])\n",
       "/Regional average            wide_table   (typ->appendable,nrows->445,ncols->6,indexers->[major_axis,minor_axis])"
      ]
     },
     "execution_count": 39,
     "output_type": "execute_result",
     "metadata": {}
    }
   ],
   "source": [
    "accuracy_store"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "appliance_name_map = {\"fridge\":\"Fridge\",\"hvac\":\"HVAC\",\"wm\":\"Washing\\nmachine\",\n",
    "                      \"dr\":\"Dryer\",\"dw\":\"Dish\\nwasher\",\"light\":\"Lights\"}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, producing the main result for Gemello. The main computation had been done by FILL THIS UP."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### ACCURACY on the train set for Gemello"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "appliances = [\"dr\",\"hvac\",\"wm\",\"dw\",\"fridge\",\"light\"]\n",
    "features = [\"Static\", \"Monthly\", \"Monthly+Static\"]\n",
    "\n",
    "out={}\n",
    "means = {}\n",
    "sems = {}\n",
    "for feature in features:\n",
    "    out[feature]={}\n",
    "    means[feature]={}\n",
    "    sems[feature]={}\n",
    "    for appliance in appliances:\n",
    "    \n",
    "        out[feature][appliance] = {}\n",
    "        for home in all_homes[appliance]:\n",
    "            try:\n",
    "                j = json.load(open(\"../main-out-new/%s_%s_%d.json\" %(appliance, feature, home),'r'))\n",
    "                out[feature][appliance][int(home)] = j['accuracy']\n",
    "            except:\n",
    "                pass\n",
    "                                     \n",
    "train_df = pd.DataFrame({k:pd.DataFrame(out[k]).mean() for k in features})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Monthly</th>\n",
       "      <th>Monthly+Static</th>\n",
       "      <th>Static</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>dr</th>\n",
       "      <td>27.399581</td>\n",
       "      <td>28.697681</td>\n",
       "      <td>26.140318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dw</th>\n",
       "      <td>21.222768</td>\n",
       "      <td>22.525817</td>\n",
       "      <td>19.941430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fridge</th>\n",
       "      <td>66.048773</td>\n",
       "      <td>68.082701</td>\n",
       "      <td>63.642268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hvac</th>\n",
       "      <td>76.502214</td>\n",
       "      <td>76.081346</td>\n",
       "      <td>68.980325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>light</th>\n",
       "      <td>36.843681</td>\n",
       "      <td>36.792903</td>\n",
       "      <td>37.165298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>wm</th>\n",
       "      <td>46.280977</td>\n",
       "      <td>47.294136</td>\n",
       "      <td>48.233818</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ]
     },
     "output_type": "execute_result",
     "metadata": {}
    }
   ],
   "source": [
    "train_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each appliance, we will choose the feature giving best accuracy on CV-train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "to_use_feature = {}\n",
    "for appliance in [\"dr\",\"hvac\",\"wm\",\"dw\",\"fridge\",\"light\"]:\n",
    "    to_use_feature[appliance] = train_df.ix[appliance].argmax()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dr': 'Monthly+Static',\n",
       " 'dw': 'Monthly+Static',\n",
       " 'fridge': 'Monthly+Static',\n",
       " 'hvac': 'Monthly',\n",
       " 'light': 'Static',\n",
       " 'wm': 'Static'}"
      ]
     },
     "execution_count": 48,
     "output_type": "execute_result",
     "metadata": {}
    }
   ],
   "source": [
    "to_use_feature"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TEST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "from constants import APPROACH as APPROACH_NAME"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "appliances = [\"dr\",\"hvac\",\"wm\",\"dw\",\"fridge\",\"light\"]\n",
    "\n",
    "out={}\n",
    "means = {APPROACH_NAME:{}}\n",
    "sems = {APPROACH_NAME:{}}\n",
    "\n",
    "for appliance in appliances:\n",
    "    feature=to_use_feature[appliance]\n",
    "    \n",
    "\n",
    "    out[appliance] = {}\n",
    "    for home in all_homes[appliance]:\n",
    "        try:\n",
    "\n",
    "            out[appliance][int(home)] = pd.read_csv(\"../main-out-new/%s_%s_%d.csv\" %(appliance, feature, home),\n",
    "                                 header=None, index_col=0, names=[\"accuracy\"]).squeeze()\n",
    "        except Exception, e:\n",
    "            print e\n",
    "\n",
    "\n",
    "    means[APPROACH_NAME][appliance_name_map[appliance]]=pd.DataFrame(out[appliance]).mean().mean()\n",
    "    sems[APPROACH_NAME][appliance_name_map[appliance]]=pd.DataFrame(out[appliance]).mean().sem()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Gemello': {'Dish\\nwasher': 27.333401871684234,\n",
       "  'Dryer': 33.052222078385938,\n",
       "  'Fridge': 68.541552745826976,\n",
       "  'HVAC': 76.239355252614146,\n",
       "  'Lights': 42.757293980395275,\n",
       "  'Washing\\nmachine': 57.272795296939734}}"
      ]
     },
     "execution_count": 53,
     "output_type": "execute_result",
     "metadata": {}
    }
   ],
   "source": [
    "means"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, getting data for LBM and FHMM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "for approach in [\"FHMM\",\"LBM\"]:\n",
    "\n",
    "    fr = accuracy_store[approach]\n",
    "    means[approach]={}\n",
    "    sems[approach] = {}\n",
    "\n",
    "    for appliance in ['hvac','fridge','dw','dr','light','wm']:\n",
    "        if appliance==\"hvac\":\n",
    "            start, stop=5, 11\n",
    "        else:\n",
    "            start, stop=1, 13\n",
    "        \n",
    "        temp = fr[appliance]\n",
    "        if approach=='N-NILM':\n",
    "            temp=temp.T\n",
    "        means[approach][appliance_name_map[appliance]] = temp.ix[range(start,stop)].mean().mean()\n",
    "        sems[approach][appliance_name_map[appliance]] = temp.ix[range(start,stop)].mean().sem()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "main_df = pd.DataFrame(means)\n",
    "se = pd.DataFrame(sems)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>FHMM</th>\n",
       "      <th>Gemello</th>\n",
       "      <th>LBM</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Dish\\nwasher</th>\n",
       "      <td>0.430750</td>\n",
       "      <td>27.333402</td>\n",
       "      <td>16.495110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dryer</th>\n",
       "      <td>10.529382</td>\n",
       "      <td>33.052222</td>\n",
       "      <td>5.940615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fridge</th>\n",
       "      <td>38.975759</td>\n",
       "      <td>68.541553</td>\n",
       "      <td>71.822239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HVAC</th>\n",
       "      <td>61.039103</td>\n",
       "      <td>76.239355</td>\n",
       "      <td>55.722782</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Lights</th>\n",
       "      <td>33.414945</td>\n",
       "      <td>42.757294</td>\n",
       "      <td>11.290612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Washing\\nmachine</th>\n",
       "      <td>0.130659</td>\n",
       "      <td>57.272795</td>\n",
       "      <td>29.972193</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ]
     },
     "output_type": "execute_result",
     "metadata": {}
    }
   ],
   "source": [
    "main_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "main_df = main_df.ix[['HVAC','Fridge','Washing\\nmachine','Lights','Dryer','Dish\\nwasher']]\n",
    "se_df =se.ix[['HVAC','Fridge','Washing\\nmachine','Lights','Dryer','Dish\\nwasher']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "from common_functions import latexify, format_axes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "main_df = main_df[[\"FHMM\",\"LBM\",APPROACH]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPEAAACTCAYAAABf9/9YAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n"
     },
     "output_type": "display_data",
     "metadata": {}
    }
   ],
   "source": [
    "#latexify(columns=2, fig_height=2.6)\n",
    "latexify(columns=1)\n",
    "\n",
    "tableau20blind =  [(31, 119, 180), (174, 199, 232), (255, 127, 14), (255, 187, 120),  \n",
    "             (44, 160, 44), (152, 223, 138), (214, 39, 40), (255, 152, 150),  \n",
    "             (148, 103, 189), (197, 176, 213), (140, 86, 75), (196, 156, 148),  \n",
    "             (227, 119, 194), (247, 182, 210), (127, 127, 127), (199, 199, 199),  \n",
    "             (188, 189, 34), (219, 219, 141), (23, 190, 207), (158, 218, 229)]\n",
    "  \n",
    "# Rescale to values between 0 and 1\n",
    "for i in range(len(tableau20blind)):  \n",
    "    r, g, b = tableau20blind[i]  \n",
    "    tableau20blind[i] = (r / 255., g / 255., b / 255.)\n",
    "\n",
    "main_df.plot(kind=\"bar\",yerr=se_df,rot=0, colors=tableau20blind[::2])\n",
    "plt.ylim((0,100))\n",
    "plt.ylabel(\"Energy Accuracy(\\%)\\n(Higher is better)\");\n",
    "format_axes(plt.gca())\n",
    "\"\"\"\n",
    "ax = plt.gca()\n",
    "for p in ax.patches:\n",
    "    if p.get_height()<20:\n",
    "        color=\"black\"\n",
    "        offset=5\n",
    "    else:\n",
    "        color=\"black\"\n",
    "        offset=-15\n",
    "    ax.annotate(str(int(round(p.get_height(),0))), (p.get_x(), p.get_height() +offset), color=color, fontsize=6)\n",
    "\"\"\"\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"/Users/nipunbatra/git/monthly-disaggregation/results/main_new.pdf\",bbox_inches=\"tight\")\n",
    "plt.savefig(\"/Users/nipunbatra/git/monthly-disaggregation/results/main_new.png\",bbox_inches=\"tight\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2.0
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}