notebooks/sensitivity_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# License \n",
    "***\n",
    "Copyright 2017 J. Patrick Hall, phall@h2o.ai & Navdeep Gill, navdeep@h2o.ai\n",
    "\n",
    "Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:\n",
    "\n",
    "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.\n",
    "\n",
    "THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sensitivity Analysis\n",
    "***"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preliminaries: imports, start h2o, load and clean data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# imports\n",
    "import h2o \n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from h2o.estimators.gbm import H2OGradientBoostingEstimator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Checking whether there is an H2O instance running at http://localhost:54321. connected.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div style=\"overflow:auto\"><table style=\"width:50%\"><tr><td>H2O cluster uptime:</td>\n",
       "<td>2 mins 07 secs</td></tr>\n",
       "<tr><td>H2O cluster version:</td>\n",
       "<td>3.14.0.2</td></tr>\n",
       "<tr><td>H2O cluster version age:</td>\n",
       "<td>2 months and 11 days </td></tr>\n",
       "<tr><td>H2O cluster name:</td>\n",
       "<td>H2O_from_python_phall_vocuai</td></tr>\n",
       "<tr><td>H2O cluster total nodes:</td>\n",
       "<td>1</td></tr>\n",
       "<tr><td>H2O cluster free memory:</td>\n",
       "<td>3.545 Gb</td></tr>\n",
       "<tr><td>H2O cluster total cores:</td>\n",
       "<td>8</td></tr>\n",
       "<tr><td>H2O cluster allowed cores:</td>\n",
       "<td>8</td></tr>\n",
       "<tr><td>H2O cluster status:</td>\n",
       "<td>locked, healthy</td></tr>\n",
       "<tr><td>H2O connection url:</td>\n",
       "<td>http://localhost:54321</td></tr>\n",
       "<tr><td>H2O connection proxy:</td>\n",
       "<td>None</td></tr>\n",
       "<tr><td>H2O internal security:</td>\n",
       "<td>False</td></tr>\n",
       "<tr><td>H2O API Extensions:</td>\n",
       "<td>XGBoost, Algos, AutoML, Core V3, Core V4</td></tr>\n",
       "<tr><td>Python version:</td>\n",
       "<td>3.5.2 final</td></tr></table></div>"
      ],
      "text/plain": [
       "--------------------------  ----------------------------------------\n",
       "H2O cluster uptime:         2 mins 07 secs\n",
       "H2O cluster version:        3.14.0.2\n",
       "H2O cluster version age:    2 months and 11 days\n",
       "H2O cluster name:           H2O_from_python_phall_vocuai\n",
       "H2O cluster total nodes:    1\n",
       "H2O cluster free memory:    3.545 Gb\n",
       "H2O cluster total cores:    8\n",
       "H2O cluster allowed cores:  8\n",
       "H2O cluster status:         locked, healthy\n",
       "H2O connection url:         http://localhost:54321\n",
       "H2O connection proxy:\n",
       "H2O internal security:      False\n",
       "H2O API Extensions:         XGBoost, Algos, AutoML, Core V3, Core V4\n",
       "Python version:             3.5.2 final\n",
       "--------------------------  ----------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# start h2o\n",
    "h2o.init()\n",
    "h2o.remove_all()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load and prepare data for modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parse progress: |█████████████████████████████████████████████████████████| 100%\n"
     ]
    }
   ],
   "source": [
    "# load data\n",
    "path = '../data/train.csv'\n",
    "frame = h2o.import_file(path=path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# assign target and inputs\n",
    "y = 'SalePrice'\n",
    "X = [name for name in frame.columns if name not in [y, 'Id']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Impute missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# determine column types\n",
    "# impute\n",
    "reals, enums = [], []\n",
    "for key, val in frame.types.items():\n",
    "    if key in X:\n",
    "        if val == 'enum':\n",
    "            enums.append(key)\n",
    "        else: \n",
    "            reals.append(key)\n",
    "            \n",
    "_ = frame[reals].impute(method='median')\n",
    "_ = frame[enums].impute(method='mode')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# split into training and validation\n",
    "train, valid = frame.split_frame([0.7], seed=12345)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train a predictive model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gbm Model Build progress: |███████████████████████████████████████████████| 100%\n",
      "gbm prediction progress: |████████████████████████████████████████████████| 100%\n"
     ]
    }
   ],
   "source": [
    "# train GBM model\n",
    "model = H2OGradientBoostingEstimator(ntrees=100,\n",
    "                                     max_depth=10,\n",
    "                                     distribution='huber',\n",
    "                                     learn_rate=0.1,\n",
    "                                     stopping_rounds=5,\n",
    "                                     seed=12345)\n",
    "\n",
    "model.train(y=y, x=X, training_frame=train, validation_frame=valid)\n",
    "\n",
    "preds = valid.cbind(model.predict(valid))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Determine important variables for use in sensitivity analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABMMAAANLCAYAAABMrVoAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x118c3db00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.varimp_plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Helper function for finding quantile indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SalePrice quantiles:\n",
      " {0: 534, 80: 148, 50: 348, 99: 770, 20: 1137, 70: 1441, 40: 1369, 10: 1136, 60: 488, 90: 818, 30: 1114}\n",
      "prediction quantiles:\n",
      " {0: 534, 80: 471, 50: 270, 99: 225, 20: 873, 70: 488, 40: 1351, 10: 69, 60: 392, 90: 641, 30: 1130}\n"
     ]
    }
   ],
   "source": [
    "def get_quantile_dict(y, id_, frame):\n",
    "\n",
    "    \"\"\" Returns the percentiles of a column y as the indices for another column id_.\n",
    "    \n",
    "    Args:\n",
    "        y: Column in which to find percentiles.\n",
    "        id_: Id column that stores indices for percentiles of y.\n",
    "        frame: H2OFrame containing y and id_. \n",
    "    \n",
    "    Returns:\n",
    "        Dictionary of percentile values and index column values.\n",
    "    \n",
    "    \"\"\"\n",
    "    \n",
    "    quantiles_df = frame.as_data_frame()\n",
    "    quantiles_df.sort_values(y, inplace=True)\n",
    "    quantiles_df.reset_index(inplace=True)\n",
    "    \n",
    "    percentiles_dict = {}\n",
    "    percentiles_dict[0] = quantiles_df.loc[0, id_]\n",
    "    percentiles_dict[99] = quantiles_df.loc[quantiles_df.shape[0]-1, id_]\n",
    "    inc = quantiles_df.shape[0]//10\n",
    "    \n",
    "    for i in range(1, 10):\n",
    "        percentiles_dict[i * 10] = quantiles_df.loc[i * inc,  id_]\n",
    "\n",
    "    return percentiles_dict\n",
    "\n",
    "sale_quantile_dict = get_quantile_dict('SalePrice', 'Id', preds)\n",
    "pred_quantile_dict = get_quantile_dict('predict', 'Id', preds)\n",
    "\n",
    "print('SalePrice quantiles:\\n', sale_quantile_dict)\n",
    "print('prediction quantiles:\\n',pred_quantile_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Get validation data ranges"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lowest SalePrice:\n",
      " "
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  SalePrice</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\">      39300</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "lowest prediction:\n",
      " "
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\">  65629.6</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "highest SalePrice:\n",
      " "
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  SalePrice</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\">     538000</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "highest prediction:\n",
      " "
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\">   440843</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "print('lowest SalePrice:\\n', preds[preds['Id'] == int(sale_quantile_dict[0])]['SalePrice'])\n",
    "print('lowest prediction:\\n', preds[preds['Id'] == int(pred_quantile_dict[0])]['predict'])\n",
    "print('highest SalePrice:\\n', preds[preds['Id'] == int(sale_quantile_dict[99])]['SalePrice'])\n",
    "print('highest prediction:\\n', preds[preds['Id'] == int(pred_quantile_dict[99])]['predict'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This result alone is interesting. The model appears to be struggling to accurately predict low and high values for SalePrice. This behavior should be corrected to increase the accuracy of predictions. A strategy for improving predictions for these homes with extreme values might be to weight them higher during training using observation weights, or they may need their own models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Now use trained model to test predictions for interesting situations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### How will the model handle making the home with the lowest predicted price even less desirable?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  Id</th><th style=\"text-align: right;\">  MSSubClass</th><th>MSZoning  </th><th style=\"text-align: right;\">  LotFrontage</th><th style=\"text-align: right;\">  LotArea</th><th>Street  </th><th>Alley  </th><th>LotShape  </th><th>LandContour  </th><th>Utilities  </th><th>LotConfig  </th><th>LandSlope  </th><th>Neighborhood  </th><th>Condition1  </th><th>Condition2  </th><th>BldgType  </th><th>HouseStyle  </th><th style=\"text-align: right;\">  OverallQual</th><th style=\"text-align: right;\">  OverallCond</th><th style=\"text-align: right;\">  YearBuilt</th><th style=\"text-align: right;\">  YearRemodAdd</th><th>RoofStyle  </th><th>RoofMatl  </th><th>Exterior1st  </th><th>Exterior2nd  </th><th>MasVnrType  </th><th style=\"text-align: right;\">  MasVnrArea</th><th>ExterQual  </th><th>ExterCond  </th><th>Foundation  </th><th>BsmtQual  </th><th>BsmtCond  </th><th>BsmtExposure  </th><th>BsmtFinType1  </th><th style=\"text-align: right;\">  BsmtFinSF1</th><th>BsmtFinType2  </th><th style=\"text-align: right;\">  BsmtFinSF2</th><th style=\"text-align: right;\">  BsmtUnfSF</th><th style=\"text-align: right;\">  TotalBsmtSF</th><th>Heating  </th><th>HeatingQC  </th><th>CentralAir  </th><th>Electrical  </th><th style=\"text-align: right;\">  1stFlrSF</th><th style=\"text-align: right;\">  2ndFlrSF</th><th style=\"text-align: right;\">  LowQualFinSF</th><th style=\"text-align: right;\">  GrLivArea</th><th style=\"text-align: right;\">  BsmtFullBath</th><th style=\"text-align: right;\">  BsmtHalfBath</th><th style=\"text-align: right;\">  FullBath</th><th style=\"text-align: right;\">  HalfBath</th><th style=\"text-align: right;\">  BedroomAbvGr</th><th style=\"text-align: right;\">  KitchenAbvGr</th><th>KitchenQual  </th><th style=\"text-align: right;\">  TotRmsAbvGrd</th><th>Functional  </th><th style=\"text-align: right;\">  Fireplaces</th><th>FireplaceQu  </th><th>GarageType  </th><th style=\"text-align: right;\">  GarageYrBlt</th><th>GarageFinish  </th><th style=\"text-align: right;\">  GarageCars</th><th style=\"text-align: right;\">  GarageArea</th><th>GarageQual  </th><th>GarageCond  </th><th>PavedDrive  </th><th style=\"text-align: right;\">  WoodDeckSF</th><th style=\"text-align: right;\">  OpenPorchSF</th><th style=\"text-align: right;\">  EnclosedPorch</th><th style=\"text-align: right;\">  3SsnPorch</th><th style=\"text-align: right;\">  ScreenPorch</th><th style=\"text-align: right;\">  PoolArea</th><th>PoolQC  </th><th>Fence  </th><th>MiscFeature  </th><th style=\"text-align: right;\">  MiscVal</th><th style=\"text-align: right;\">  MoSold</th><th style=\"text-align: right;\">  YrSold</th><th>SaleType  </th><th>SaleCondition  </th><th style=\"text-align: right;\">  SalePrice</th><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\"> 534</td><td style=\"text-align: right;\">          20</td><td>RL        </td><td style=\"text-align: right;\">           50</td><td style=\"text-align: right;\">     5000</td><td>Pave    </td><td>NA     </td><td>Reg       </td><td>Low          </td><td>AllPub     </td><td>Inside     </td><td>Mod        </td><td>BrkSide       </td><td>Norm        </td><td>Norm        </td><td>1Fam      </td><td>1Story      </td><td style=\"text-align: right;\">            1</td><td style=\"text-align: right;\">            3</td><td style=\"text-align: right;\">       1946</td><td style=\"text-align: right;\">          1950</td><td>Gable      </td><td>CompShg   </td><td>VinylSd      </td><td>VinylSd      </td><td>None        </td><td style=\"text-align: right;\">           0</td><td>Fa         </td><td>Fa         </td><td>Slab        </td><td>NA        </td><td>NA        </td><td>NA            </td><td>NA            </td><td style=\"text-align: right;\">           0</td><td>NA            </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td>GasA     </td><td>Fa         </td><td>N           </td><td>FuseF       </td><td style=\"text-align: right;\">       334</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">        334</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">         1</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             1</td><td style=\"text-align: right;\">             1</td><td>Fa           </td><td style=\"text-align: right;\">             2</td><td>Typ         </td><td style=\"text-align: right;\">           0</td><td>NA           </td><td>NA          </td><td style=\"text-align: right;\">      1978.51</td><td>NA            </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">           0</td><td>NA          </td><td>NA          </td><td>N           </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">              0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">         0</td><td>NA      </td><td>NA     </td><td>NA           </td><td style=\"text-align: right;\">        0</td><td style=\"text-align: right;\">       1</td><td style=\"text-align: right;\">    2007</td><td>WD        </td><td>Normal         </td><td style=\"text-align: right;\">      39300</td><td style=\"text-align: right;\">  65629.6</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# look at current row\n",
    "print(preds[preds['Id'] == int(pred_quantile_dict[0])])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error: 67.00%\n"
     ]
    }
   ],
   "source": [
    "# find current error\n",
    "observed = preds[preds['Id'] == int(pred_quantile_dict[0])]['SalePrice'][0,0]\n",
    "predicted = preds[preds['Id'] == int(pred_quantile_dict[0])]['predict'][0,0]\n",
    "print('Error: %.2f%%' % (100*(abs(observed - predicted)/observed)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gbm prediction progress: |████████████████████████████████████████████████| 100%\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  Id</th><th style=\"text-align: right;\">  MSSubClass</th><th>MSZoning  </th><th style=\"text-align: right;\">  LotFrontage</th><th style=\"text-align: right;\">  LotArea</th><th>Street  </th><th>Alley  </th><th>LotShape  </th><th>LandContour  </th><th>Utilities  </th><th>LotConfig  </th><th>LandSlope  </th><th>Neighborhood  </th><th>Condition1  </th><th>Condition2  </th><th>BldgType  </th><th>HouseStyle  </th><th style=\"text-align: right;\">  OverallQual</th><th style=\"text-align: right;\">  OverallCond</th><th style=\"text-align: right;\">  YearBuilt</th><th style=\"text-align: right;\">  YearRemodAdd</th><th>RoofStyle  </th><th>RoofMatl  </th><th>Exterior1st  </th><th>Exterior2nd  </th><th>MasVnrType  </th><th style=\"text-align: right;\">  MasVnrArea</th><th>ExterQual  </th><th>ExterCond  </th><th>Foundation  </th><th>BsmtQual  </th><th>BsmtCond  </th><th>BsmtExposure  </th><th>BsmtFinType1  </th><th style=\"text-align: right;\">  BsmtFinSF1</th><th>BsmtFinType2  </th><th style=\"text-align: right;\">  BsmtFinSF2</th><th style=\"text-align: right;\">  BsmtUnfSF</th><th style=\"text-align: right;\">  TotalBsmtSF</th><th>Heating  </th><th>HeatingQC  </th><th>CentralAir  </th><th>Electrical  </th><th style=\"text-align: right;\">  1stFlrSF</th><th style=\"text-align: right;\">  2ndFlrSF</th><th style=\"text-align: right;\">  LowQualFinSF</th><th style=\"text-align: right;\">  GrLivArea</th><th style=\"text-align: right;\">  BsmtFullBath</th><th style=\"text-align: right;\">  BsmtHalfBath</th><th style=\"text-align: right;\">  FullBath</th><th style=\"text-align: right;\">  HalfBath</th><th style=\"text-align: right;\">  BedroomAbvGr</th><th style=\"text-align: right;\">  KitchenAbvGr</th><th>KitchenQual  </th><th style=\"text-align: right;\">  TotRmsAbvGrd</th><th>Functional  </th><th style=\"text-align: right;\">  Fireplaces</th><th>FireplaceQu  </th><th>GarageType  </th><th style=\"text-align: right;\">  GarageYrBlt</th><th>GarageFinish  </th><th style=\"text-align: right;\">  GarageCars</th><th style=\"text-align: right;\">  GarageArea</th><th>GarageQual  </th><th>GarageCond  </th><th>PavedDrive  </th><th style=\"text-align: right;\">  WoodDeckSF</th><th style=\"text-align: right;\">  OpenPorchSF</th><th style=\"text-align: right;\">  EnclosedPorch</th><th style=\"text-align: right;\">  3SsnPorch</th><th style=\"text-align: right;\">  ScreenPorch</th><th style=\"text-align: right;\">  PoolArea</th><th>PoolQC  </th><th>Fence  </th><th>MiscFeature  </th><th style=\"text-align: right;\">  MiscVal</th><th style=\"text-align: right;\">  MoSold</th><th style=\"text-align: right;\">  YrSold</th><th>SaleType  </th><th>SaleCondition  </th><th style=\"text-align: right;\">  SalePrice</th><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\"> 534</td><td style=\"text-align: right;\">          20</td><td>RL        </td><td style=\"text-align: right;\">           50</td><td style=\"text-align: right;\">     5000</td><td>Pave    </td><td>NA     </td><td>Reg       </td><td>Low          </td><td>AllPub     </td><td>Inside     </td><td>Mod        </td><td>IDOTRR        </td><td>Norm        </td><td>Norm        </td><td>1Fam      </td><td>1Story      </td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">            3</td><td style=\"text-align: right;\">       1946</td><td style=\"text-align: right;\">          1950</td><td>Gable      </td><td>CompShg   </td><td>VinylSd      </td><td>VinylSd      </td><td>None        </td><td style=\"text-align: right;\">           0</td><td>Fa         </td><td>Fa         </td><td>Slab        </td><td>NA        </td><td>NA        </td><td>NA            </td><td>NA            </td><td style=\"text-align: right;\">           0</td><td>NA            </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td>GasA     </td><td>Fa         </td><td>N           </td><td>FuseF       </td><td style=\"text-align: right;\">       334</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">        500</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">         1</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             1</td><td style=\"text-align: right;\">             1</td><td>Fa           </td><td style=\"text-align: right;\">             2</td><td>Typ         </td><td style=\"text-align: right;\">           0</td><td>NA           </td><td>NA          </td><td style=\"text-align: right;\">      1978.51</td><td>NA            </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">           0</td><td>NA          </td><td>NA          </td><td>N           </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">              0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">         0</td><td>NA      </td><td>NA     </td><td>NA           </td><td style=\"text-align: right;\">        0</td><td style=\"text-align: right;\">       1</td><td style=\"text-align: right;\">    2007</td><td>WD        </td><td>Normal         </td><td style=\"text-align: right;\">      39300</td><td style=\"text-align: right;\">    50466</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Error: 28.41%\n"
     ]
    }
   ],
   "source": [
    "# change value of important variables\n",
    "test_case = preds[preds['Id'] == int(pred_quantile_dict[0])]\n",
    "test_case = test_case.drop('predict')\n",
    "test_case['OverallQual'] = 0\n",
    "test_case['Neighborhood'] = 'IDOTRR'\n",
    "test_case['GrLivArea'] = 500\n",
    "test_case = test_case.cbind(model.predict(test_case))\n",
    "print(test_case)\n",
    "\n",
    "# recalculate error\n",
    "observed = test_case['SalePrice'][0,0]\n",
    "predicted = test_case['predict'][0,0]\n",
    "print('Error: %.2f%%' % (100*(abs(observed - predicted)/observed)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While the model does not seem to handle low-valued homes very well, making the home with the lowest predicted price less appealling does not seem to make the model's predictions any worse. While this prediction behavior appears somewhat stable, which would normally be desirable, this is not particularly good news as the underlying prediction is so inaccurate. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### How will the model handle making the home with the highest predicted price even more desirable?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  Id</th><th style=\"text-align: right;\">  MSSubClass</th><th>MSZoning  </th><th style=\"text-align: right;\">  LotFrontage</th><th style=\"text-align: right;\">  LotArea</th><th>Street  </th><th>Alley  </th><th>LotShape  </th><th>LandContour  </th><th>Utilities  </th><th>LotConfig  </th><th>LandSlope  </th><th>Neighborhood  </th><th>Condition1  </th><th>Condition2  </th><th>BldgType  </th><th>HouseStyle  </th><th style=\"text-align: right;\">  OverallQual</th><th style=\"text-align: right;\">  OverallCond</th><th style=\"text-align: right;\">  YearBuilt</th><th style=\"text-align: right;\">  YearRemodAdd</th><th>RoofStyle  </th><th>RoofMatl  </th><th>Exterior1st  </th><th>Exterior2nd  </th><th>MasVnrType  </th><th style=\"text-align: right;\">  MasVnrArea</th><th>ExterQual  </th><th>ExterCond  </th><th>Foundation  </th><th>BsmtQual  </th><th>BsmtCond  </th><th>BsmtExposure  </th><th>BsmtFinType1  </th><th style=\"text-align: right;\">  BsmtFinSF1</th><th>BsmtFinType2  </th><th style=\"text-align: right;\">  BsmtFinSF2</th><th style=\"text-align: right;\">  BsmtUnfSF</th><th style=\"text-align: right;\">  TotalBsmtSF</th><th>Heating  </th><th>HeatingQC  </th><th>CentralAir  </th><th>Electrical  </th><th style=\"text-align: right;\">  1stFlrSF</th><th style=\"text-align: right;\">  2ndFlrSF</th><th style=\"text-align: right;\">  LowQualFinSF</th><th style=\"text-align: right;\">  GrLivArea</th><th style=\"text-align: right;\">  BsmtFullBath</th><th style=\"text-align: right;\">  BsmtHalfBath</th><th style=\"text-align: right;\">  FullBath</th><th style=\"text-align: right;\">  HalfBath</th><th style=\"text-align: right;\">  BedroomAbvGr</th><th style=\"text-align: right;\">  KitchenAbvGr</th><th>KitchenQual  </th><th style=\"text-align: right;\">  TotRmsAbvGrd</th><th>Functional  </th><th style=\"text-align: right;\">  Fireplaces</th><th>FireplaceQu  </th><th>GarageType  </th><th style=\"text-align: right;\">  GarageYrBlt</th><th>GarageFinish  </th><th style=\"text-align: right;\">  GarageCars</th><th style=\"text-align: right;\">  GarageArea</th><th>GarageQual  </th><th>GarageCond  </th><th>PavedDrive  </th><th style=\"text-align: right;\">  WoodDeckSF</th><th style=\"text-align: right;\">  OpenPorchSF</th><th style=\"text-align: right;\">  EnclosedPorch</th><th style=\"text-align: right;\">  3SsnPorch</th><th style=\"text-align: right;\">  ScreenPorch</th><th style=\"text-align: right;\">  PoolArea</th><th>PoolQC  </th><th>Fence  </th><th>MiscFeature  </th><th style=\"text-align: right;\">  MiscVal</th><th style=\"text-align: right;\">  MoSold</th><th style=\"text-align: right;\">  YrSold</th><th>SaleType  </th><th>SaleCondition  </th><th style=\"text-align: right;\">  SalePrice</th><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\"> 225</td><td style=\"text-align: right;\">          20</td><td>RL        </td><td style=\"text-align: right;\">          103</td><td style=\"text-align: right;\">    13472</td><td>Pave    </td><td>NA     </td><td>Reg       </td><td>Lvl          </td><td>AllPub     </td><td>Inside     </td><td>Gtl        </td><td>NridgHt       </td><td>Norm        </td><td>Norm        </td><td>1Fam      </td><td>1Story      </td><td style=\"text-align: right;\">           10</td><td style=\"text-align: right;\">            5</td><td style=\"text-align: right;\">       2003</td><td style=\"text-align: right;\">          2003</td><td>Hip        </td><td>CompShg   </td><td>VinylSd      </td><td>VinylSd      </td><td>BrkFace     </td><td style=\"text-align: right;\">         922</td><td>Ex         </td><td>TA         </td><td>PConc       </td><td>Ex        </td><td>TA        </td><td>Gd            </td><td>GLQ           </td><td style=\"text-align: right;\">          56</td><td>Unf           </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">       2336</td><td style=\"text-align: right;\">         2392</td><td>GasA     </td><td>Ex         </td><td>Y           </td><td>SBrkr       </td><td style=\"text-align: right;\">      2392</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">       2392</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">         2</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             3</td><td style=\"text-align: right;\">             1</td><td>Ex           </td><td style=\"text-align: right;\">             8</td><td>Typ         </td><td style=\"text-align: right;\">           1</td><td>Ex           </td><td>Attchd      </td><td style=\"text-align: right;\">         2003</td><td>Fin           </td><td style=\"text-align: right;\">           3</td><td style=\"text-align: right;\">         968</td><td>TA          </td><td>TA          </td><td>Y           </td><td style=\"text-align: right;\">         248</td><td style=\"text-align: right;\">          105</td><td style=\"text-align: right;\">              0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">         0</td><td>NA      </td><td>NA     </td><td>NA           </td><td style=\"text-align: right;\">        0</td><td style=\"text-align: right;\">       6</td><td style=\"text-align: right;\">    2009</td><td>WD        </td><td>Normal         </td><td style=\"text-align: right;\">     386250</td><td style=\"text-align: right;\">   440843</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# look at current row\n",
    "print(preds[preds['Id'] == int(pred_quantile_dict[99])])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error: 14.13%\n"
     ]
    }
   ],
   "source": [
    "# find current error\n",
    "observed = preds[preds['Id'] == int(pred_quantile_dict[99])]['SalePrice'][0,0]\n",
    "predicted = preds[preds['Id'] == int(pred_quantile_dict[99])]['predict'][0,0]\n",
    "print('Error: %.2f%%' % (100*(abs(observed - predicted)/observed)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gbm prediction progress: |████████████████████████████████████████████████| 100%\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead>\n",
       "<tr><th style=\"text-align: right;\">  Id</th><th style=\"text-align: right;\">  MSSubClass</th><th>MSZoning  </th><th style=\"text-align: right;\">  LotFrontage</th><th style=\"text-align: right;\">  LotArea</th><th>Street  </th><th>Alley  </th><th>LotShape  </th><th>LandContour  </th><th>Utilities  </th><th>LotConfig  </th><th>LandSlope  </th><th>Neighborhood  </th><th>Condition1  </th><th>Condition2  </th><th>BldgType  </th><th>HouseStyle  </th><th style=\"text-align: right;\">  OverallQual</th><th style=\"text-align: right;\">  OverallCond</th><th style=\"text-align: right;\">  YearBuilt</th><th style=\"text-align: right;\">  YearRemodAdd</th><th>RoofStyle  </th><th>RoofMatl  </th><th>Exterior1st  </th><th>Exterior2nd  </th><th>MasVnrType  </th><th style=\"text-align: right;\">  MasVnrArea</th><th>ExterQual  </th><th>ExterCond  </th><th>Foundation  </th><th>BsmtQual  </th><th>BsmtCond  </th><th>BsmtExposure  </th><th>BsmtFinType1  </th><th style=\"text-align: right;\">  BsmtFinSF1</th><th>BsmtFinType2  </th><th style=\"text-align: right;\">  BsmtFinSF2</th><th style=\"text-align: right;\">  BsmtUnfSF</th><th style=\"text-align: right;\">  TotalBsmtSF</th><th>Heating  </th><th>HeatingQC  </th><th>CentralAir  </th><th>Electrical  </th><th style=\"text-align: right;\">  1stFlrSF</th><th style=\"text-align: right;\">  2ndFlrSF</th><th style=\"text-align: right;\">  LowQualFinSF</th><th style=\"text-align: right;\">  GrLivArea</th><th style=\"text-align: right;\">  BsmtFullBath</th><th style=\"text-align: right;\">  BsmtHalfBath</th><th style=\"text-align: right;\">  FullBath</th><th style=\"text-align: right;\">  HalfBath</th><th style=\"text-align: right;\">  BedroomAbvGr</th><th style=\"text-align: right;\">  KitchenAbvGr</th><th>KitchenQual  </th><th style=\"text-align: right;\">  TotRmsAbvGrd</th><th>Functional  </th><th style=\"text-align: right;\">  Fireplaces</th><th>FireplaceQu  </th><th>GarageType  </th><th style=\"text-align: right;\">  GarageYrBlt</th><th>GarageFinish  </th><th style=\"text-align: right;\">  GarageCars</th><th style=\"text-align: right;\">  GarageArea</th><th>GarageQual  </th><th>GarageCond  </th><th>PavedDrive  </th><th style=\"text-align: right;\">  WoodDeckSF</th><th style=\"text-align: right;\">  OpenPorchSF</th><th style=\"text-align: right;\">  EnclosedPorch</th><th style=\"text-align: right;\">  3SsnPorch</th><th style=\"text-align: right;\">  ScreenPorch</th><th style=\"text-align: right;\">  PoolArea</th><th>PoolQC  </th><th>Fence  </th><th>MiscFeature  </th><th style=\"text-align: right;\">  MiscVal</th><th style=\"text-align: right;\">  MoSold</th><th style=\"text-align: right;\">  YrSold</th><th>SaleType  </th><th>SaleCondition  </th><th style=\"text-align: right;\">  SalePrice</th><th style=\"text-align: right;\">  predict</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "<tr><td style=\"text-align: right;\"> 225</td><td style=\"text-align: right;\">          20</td><td>RL        </td><td style=\"text-align: right;\">          103</td><td style=\"text-align: right;\">    13472</td><td>Pave    </td><td>NA     </td><td>Reg       </td><td>Lvl          </td><td>AllPub     </td><td>Inside     </td><td>Gtl        </td><td>StoneBr       </td><td>Norm        </td><td>Norm        </td><td>1Fam      </td><td>1Story      </td><td style=\"text-align: right;\">           10</td><td style=\"text-align: right;\">            5</td><td style=\"text-align: right;\">       2003</td><td style=\"text-align: right;\">          2003</td><td>Hip        </td><td>CompShg   </td><td>VinylSd      </td><td>VinylSd      </td><td>BrkFace     </td><td style=\"text-align: right;\">         922</td><td>Ex         </td><td>TA         </td><td>PConc       </td><td>Ex        </td><td>TA        </td><td>Gd            </td><td>GLQ           </td><td style=\"text-align: right;\">          56</td><td>Unf           </td><td style=\"text-align: right;\">           0</td><td style=\"text-align: right;\">       2336</td><td style=\"text-align: right;\">         2392</td><td>GasA     </td><td>Ex         </td><td>Y           </td><td>SBrkr       </td><td style=\"text-align: right;\">      2392</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">       5000</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">             0</td><td style=\"text-align: right;\">         2</td><td style=\"text-align: right;\">         0</td><td style=\"text-align: right;\">             3</td><td style=\"text-align: right;\">             1</td><td>Ex           </td><td style=\"text-align: right;\">             8</td><td>Typ         </td><td style=\"text-align: right;\">           1</td><td>Ex           </td><td>Attchd      </td><td style=\"text-align: right;\">         2003</td><td>Fin           </td><td style=\"text-align: right;\">           3</td><td style=\"text-align: right;\">         968</td><td>TA          </td><td>TA          </td><td>Y           </td><td style=\"text-align: right;\">         248</td><td style=\"text-align: right;\">          105</td><td style=\"text-align: right;\">              0</td><td style=\"text-align: right;\">          0</td><td style=\"text-align: right;\">            0</td><td style=\"text-align: right;\">         0</td><td>NA      </td><td>NA     </td><td>NA           </td><td style=\"text-align: right;\">        0</td><td style=\"text-align: right;\">       6</td><td style=\"text-align: right;\">    2009</td><td>WD        </td><td>Normal         </td><td style=\"text-align: right;\">     386250</td><td style=\"text-align: right;\">   478120</td></tr>\n",
       "</tbody>\n",
       "</table>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Error: 23.78%\n"
     ]
    }
   ],
   "source": [
    "# change value of important variables\n",
    "test_case = preds[preds['Id'] == int(pred_quantile_dict[99])]\n",
    "test_case = test_case.drop('predict')\n",
    "test_case['Neighborhood'] = 'StoneBr'\n",
    "test_case['GrLivArea'] = 5000\n",
    "test_case = test_case.cbind(model.predict(test_case))\n",
    "print(test_case)\n",
    "\n",
    "# recalculate error\n",
    "observed = test_case['SalePrice'][0,0]\n",
    "predicted = test_case['predict'][0,0]\n",
    "print('Error: %.2f%%' % (100*(abs(observed - predicted)/observed)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This result may point to unstable predictions for the higher end of SalesPrice."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Shutdown H2O"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Are you sure you want to shutdown the H2O instance running at http://localhost:54321 (Y/N)? y\n",
      "H2O session _sid_b43f closed.\n"
     ]
    }
   ],
   "source": [
    "h2o.cluster().shutdown(prompt=True)"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}