{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "from datascience import *\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plots\n",
    "plots.style.use('fivethirtyeight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multiple regression\n",
    "\n",
    "We'll illustrate multiple linear regression using a toy \"Zillow competition\" problem, using a data set of house sales in Ames, Iowa. This example parallels \n",
    "the treatment in [Chapter 17.6](https://www.inferentialthinking.com/chapters/17/6/Multiple_Regression.html) of the book.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>SalePrice</th> <th>1st Flr SF</th> <th>2nd Flr SF</th> <th>Total Bsmt SF</th> <th>Garage Area</th> <th>Wood Deck SF</th> <th>Open Porch SF</th> <th>Lot Area</th> <th>Year Built</th> <th>Yr Sold</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>35000    </td> <td>498       </td> <td>0         </td> <td>498          </td> <td>216        </td> <td>0           </td> <td>0            </td> <td>8088    </td> <td>1922      </td> <td>2006   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>39300    </td> <td>334       </td> <td>0         </td> <td>0            </td> <td>0          </td> <td>0           </td> <td>0            </td> <td>5000    </td> <td>1946      </td> <td>2007   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>40000    </td> <td>649       </td> <td>668       </td> <td>649          </td> <td>250        </td> <td>0           </td> <td>54           </td> <td>8500    </td> <td>1920      </td> <td>2008   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>45000    </td> <td>612       </td> <td>0         </td> <td>0            </td> <td>308        </td> <td>0           </td> <td>0            </td> <td>5925    </td> <td>1940      </td> <td>2009   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>52000    </td> <td>729       </td> <td>0         </td> <td>270          </td> <td>0          </td> <td>0           </td> <td>0            </td> <td>4130    </td> <td>1935      </td> <td>2008   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>52500    </td> <td>693       </td> <td>0         </td> <td>693          </td> <td>0          </td> <td>0           </td> <td>20           </td> <td>4118    </td> <td>1941      </td> <td>2006   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>55000    </td> <td>723       </td> <td>363       </td> <td>723          </td> <td>400        </td> <td>0           </td> <td>24           </td> <td>11340   </td> <td>1920      </td> <td>2008   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>55000    </td> <td>796       </td> <td>0         </td> <td>796          </td> <td>0          </td> <td>0           </td> <td>0            </td> <td>3636    </td> <td>1922      </td> <td>2008   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>57625    </td> <td>810       </td> <td>0         </td> <td>0            </td> <td>280        </td> <td>119         </td> <td>24           </td> <td>21780   </td> <td>1910      </td> <td>2009   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>58500    </td> <td>864       </td> <td>0         </td> <td>864          </td> <td>200        </td> <td>0           </td> <td>0            </td> <td>8212    </td> <td>1914      </td> <td>2010   </td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>\n",
       "<p>... (1992 rows omitted)</p>"
      ],
      "text/plain": [
       "SalePrice | 1st Flr SF | 2nd Flr SF | Total Bsmt SF | Garage Area | Wood Deck SF | Open Porch SF | Lot Area | Year Built | Yr Sold\n",
       "35000     | 498        | 0          | 498           | 216         | 0            | 0             | 8088     | 1922       | 2006\n",
       "39300     | 334        | 0          | 0             | 0           | 0            | 0             | 5000     | 1946       | 2007\n",
       "40000     | 649        | 668        | 649           | 250         | 0            | 54            | 8500     | 1920       | 2008\n",
       "45000     | 612        | 0          | 0             | 308         | 0            | 0             | 5925     | 1940       | 2009\n",
       "52000     | 729        | 0          | 270           | 0           | 0            | 0             | 4130     | 1935       | 2008\n",
       "52500     | 693        | 0          | 693           | 0           | 0            | 20            | 4118     | 1941       | 2006\n",
       "55000     | 723        | 363        | 723           | 400         | 0            | 24            | 11340    | 1920       | 2008\n",
       "55000     | 796        | 0          | 796           | 0           | 0            | 0             | 3636     | 1922       | 2008\n",
       "57625     | 810        | 0          | 0             | 280         | 119          | 24            | 21780    | 1910       | 2009\n",
       "58500     | 864        | 0          | 864           | 200         | 0            | 0             | 8212     | 1914       | 2010\n",
       "... (1992 rows omitted)"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's just get data from 1 Family houses and a subset of the columns\n",
    "all_sales = Table.read_table('house.csv')\n",
    "sales = all_sales.where('Bldg Type', '1Fam').where('Sale Condition', 'Normal').select(\n",
    "    'SalePrice', '1st Flr SF', '2nd Flr SF', \n",
    "    'Total Bsmt SF', 'Garage Area', \n",
    "    'Wood Deck SF', 'Open Porch SF', 'Lot Area', \n",
    "    'Year Built', 'Yr Sold')\n",
    "sales.sort('SalePrice')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# View a histogram of the sale prices\n",
    "sales.hist('SalePrice', bins=32, unit='$')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# View a scatter plot looking at the first floor square footage and price\n",
    "sales.scatter('1st Flr SF', 'SalePrice')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining our function to get the correlation\n",
    "def standardize(x):\n",
    "    return (x-np.mean(x))/np.std(x)\n",
    "\n",
    "def correlation(t, col1, col2):\n",
    "    x = standardize(t.column(col1))\n",
    "    y = standardize(t.column(col2))\n",
    "    return np.mean(x*y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6424662541030225"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Look at the correlation between sales price and first floor square footage\n",
    "correlation(sales, 'SalePrice', '1st Flr SF')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>Attribute</th> <th>Correlation with Sale Price</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>SalePrice    </td> <td>1                          </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>1st Flr SF   </td> <td>0.642466                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>2nd Flr SF   </td> <td>0.357522                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Total Bsmt SF</td> <td>0.652979                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Garage Area  </td> <td>0.638594                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Wood Deck SF </td> <td>0.352699                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Open Porch SF</td> <td>0.336909                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Lot Area     </td> <td>0.290823                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Year Built   </td> <td>0.565165                   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>Yr Sold      </td> <td>0.0259486                  </td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "Attribute     | Correlation with Sale Price\n",
       "SalePrice     | 1\n",
       "1st Flr SF    | 0.642466\n",
       "2nd Flr SF    | 0.357522\n",
       "Total Bsmt SF | 0.652979\n",
       "Garage Area   | 0.638594\n",
       "Wood Deck SF  | 0.352699\n",
       "Open Porch SF | 0.336909\n",
       "Lot Area      | 0.290823\n",
       "Year Built    | 0.565165\n",
       "Yr Sold       | 0.0259486"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get the correlation between all our variables and price\n",
    "r = make_array()\n",
    "for label in sales.labels:\n",
    "    r = np.append(r, correlation(sales, label, 'SalePrice'))\n",
    "\n",
    "corr_table = Table().with_columns('Attribute', sales.labels, 'Correlation with Sale Price', r)\n",
    "corr_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.2305611808841421\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Look at the correlation between first and second floor square footage\n",
    "# Any explanation for this relationship? \n",
    "sales.scatter('1st Flr SF', '2nd Flr SF')\n",
    "print(correlation(sales, '2nd Flr SF', '1st Flr SF'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7821920556134877"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Relationship between square footage of both floors and price\n",
    "both_floors = sales.column(1) + sales.column(2)\n",
    "correlation(sales.with_column('Both Floors', both_floors), 'SalePrice', 'Both Floors')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>SalePrice</th> <th>1st Flr SF</th> <th>2nd Flr SF</th> <th>Total Bsmt SF</th> <th>Garage Area</th> <th>Wood Deck SF</th> <th>Open Porch SF</th> <th>Lot Area</th> <th>Year Built</th> <th>Yr Sold</th> <th>one</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>215000   </td> <td>1656      </td> <td>0         </td> <td>1080         </td> <td>528        </td> <td>210         </td> <td>62           </td> <td>31770   </td> <td>1960      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>105000   </td> <td>896       </td> <td>0         </td> <td>882          </td> <td>730        </td> <td>140         </td> <td>0            </td> <td>11622   </td> <td>1961      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>172000   </td> <td>1329      </td> <td>0         </td> <td>1329         </td> <td>312        </td> <td>393         </td> <td>36           </td> <td>14267   </td> <td>1958      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>244000   </td> <td>2110      </td> <td>0         </td> <td>2110         </td> <td>522        </td> <td>0           </td> <td>0            </td> <td>11160   </td> <td>1968      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>189900   </td> <td>928       </td> <td>701       </td> <td>928          </td> <td>482        </td> <td>212         </td> <td>34           </td> <td>13830   </td> <td>1997      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>195500   </td> <td>926       </td> <td>678       </td> <td>926          </td> <td>470        </td> <td>360         </td> <td>36           </td> <td>9978    </td> <td>1998      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>189000   </td> <td>1028      </td> <td>776       </td> <td>994          </td> <td>442        </td> <td>140         </td> <td>60           </td> <td>7500    </td> <td>1999      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>175900   </td> <td>763       </td> <td>892       </td> <td>763          </td> <td>440        </td> <td>157         </td> <td>84           </td> <td>10000   </td> <td>1993      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>185000   </td> <td>1187      </td> <td>0         </td> <td>1168         </td> <td>420        </td> <td>483         </td> <td>21           </td> <td>7980    </td> <td>1992      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>180400   </td> <td>789       </td> <td>676       </td> <td>789          </td> <td>393        </td> <td>0           </td> <td>75           </td> <td>8402    </td> <td>1998      </td> <td>2010   </td> <td>1   </td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>\n",
       "<p>... (1992 rows omitted)</p>"
      ],
      "text/plain": [
       "SalePrice | 1st Flr SF | 2nd Flr SF | Total Bsmt SF | Garage Area | Wood Deck SF | Open Porch SF | Lot Area | Year Built | Yr Sold | one\n",
       "215000    | 1656       | 0          | 1080          | 528         | 210          | 62            | 31770    | 1960       | 2010    | 1\n",
       "105000    | 896        | 0          | 882           | 730         | 140          | 0             | 11622    | 1961       | 2010    | 1\n",
       "172000    | 1329       | 0          | 1329          | 312         | 393          | 36            | 14267    | 1958       | 2010    | 1\n",
       "244000    | 2110       | 0          | 2110          | 522         | 0            | 0             | 11160    | 1968       | 2010    | 1\n",
       "189900    | 928        | 701        | 928           | 482         | 212          | 34            | 13830    | 1997       | 2010    | 1\n",
       "195500    | 926        | 678        | 926           | 470         | 360          | 36            | 9978     | 1998       | 2010    | 1\n",
       "189000    | 1028       | 776        | 994           | 442         | 140          | 60            | 7500     | 1999       | 2010    | 1\n",
       "175900    | 763        | 892        | 763           | 440         | 157          | 84            | 10000    | 1993       | 2010    | 1\n",
       "185000    | 1187       | 0          | 1168          | 420         | 483          | 21            | 7980     | 1992       | 2010    | 1\n",
       "180400    | 789        | 676        | 789           | 393         | 0            | 75            | 8402     | 1998       | 2010    | 1\n",
       "... (1992 rows omitted)"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Add a column of all ones to the data\n",
    "# This will enable us to have an intercept in our regression model\n",
    "sales = sales.with_column('one', np.ones(sales.num_rows))\n",
    "sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1001 training and 1001 test instances.\n"
     ]
    }
   ],
   "source": [
    "# Create training and test splits of our data using the tb.split() method\n",
    "train, test = sales.split(1001)\n",
    "print(train.num_rows, 'training and', test.num_rows, 'test instances.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicting sale price for:\n",
      " Row(1st Flr SF=1592, 2nd Flr SF=0, Total Bsmt SF=1556.0, Garage Area=484.0, Wood Deck SF=120, Open Porch SF=35, Lot Area=11927, Year Built=1994, Yr Sold=2007, one=1.0)\n",
      "\n",
      "Using slopes\n",
      ": [ 8.90216092  8.90396857  8.31110986 10.99595012  9.28306723  9.3868881\n",
      "  8.12498193  9.3778265   9.56643952 10.12601652]\n",
      "\n",
      "Result: 168684.89178344968\n"
     ]
    }
   ],
   "source": [
    "# Create a function that takes a row and predicts the price based on a vector of coefficents called \"slopes\"\n",
    "def predict(slopes, row):\n",
    "    return sum(slopes * np.array(tuple(row)))\n",
    "\n",
    "# Use row 0 as an example to make predictions on\n",
    "example_row = test.drop('SalePrice').row(0)\n",
    "print('Predicting sale price for:\\n', example_row)\n",
    "\n",
    "# Create random slopes (this should lead to poor predictions)\n",
    "example_slopes = np.random.normal(10, 1, len(example_row))\n",
    "print('\\nUsing slopes\\n:', example_slopes)\n",
    "\n",
    "# The resulting predicted price based on the random slopes\n",
    "print('\\nResult:', predict(example_slopes, example_row))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Actual sale price: 228000\n",
      "Predicted sale price using random slopes: 168684.89178344968\n"
     ]
    }
   ],
   "source": [
    "# Comparing the actual price and predicted price\n",
    "print('Actual sale price:', test.column('SalePrice').item(0))\n",
    "print('Predicted sale price using random slopes:', predict(example_slopes, example_row))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a function that returns the RMSE for given slopes, features and actual prices\n",
    "def rmse(slopes, attributes, prices):\n",
    "    errors = []\n",
    "    for i in np.arange(len(prices)):\n",
    "        predicted = predict(slopes, attributes.row(i))\n",
    "        actual = prices.item(i)\n",
    "        errors.append((predicted - actual) ** 2)\n",
    "    return np.mean(errors) ** 0.5\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE of all training examples using random slopes: $79163.17\n"
     ]
    }
   ],
   "source": [
    "# Write a function to get the RMSE for the training data using our randomly generated slopes\n",
    "train_prices = train.column(0)\n",
    "train_attributes = train.drop(0)\n",
    "\n",
    "def rmse_train(slopes):\n",
    "    return rmse(slopes, train_attributes, train_prices)\n",
    "\n",
    "print('RMSE of all training examples using random slopes: $%.2f' % rmse_train(example_slopes))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best slopes for the training set:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>1st Flr SF</th> <th>2nd Flr SF</th> <th>Total Bsmt SF</th> <th>Garage Area</th> <th>Wood Deck SF</th> <th>Open Porch SF</th> <th>Lot Area</th> <th>Year Built</th> <th>Yr Sold</th> <th>one</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>68.0891   </td> <td>70.6755   </td> <td>50.5732      </td> <td>50.6218    </td> <td>37.7467     </td> <td>27.3521      </td> <td>0.619194</td> <td>502.889   </td> <td>-499.096</td> <td>9.8618</td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE of all training examples using the best slopes: $29929.64\n"
     ]
    }
   ],
   "source": [
    "# Get the \"best slopes\" by miniming the RMSE\n",
    "best_slopes = minimize(rmse_train, start=example_slopes, smooth=True, array=True)\n",
    "print('The best slopes for the training set:')\n",
    "\n",
    "# Show the best slopes found and the RMSE found on the training data\n",
    "Table(train_attributes.labels).with_row(list(best_slopes)).show()\n",
    "print('RMSE of all training examples using the best slopes: $%.2f' % rmse_train(best_slopes))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set RMSE for multiple linear regression: $32413.22\n"
     ]
    }
   ],
   "source": [
    "# Let's now evaluate the RMSE on the test data\n",
    "test_prices = test.column(0)\n",
    "test_attributes = test.drop(0)\n",
    "\n",
    "rmse_linear = rmse(best_slopes, test_attributes, test_prices)\n",
    "print('Test set RMSE for multiple linear regression: $%.2f'% rmse_linear)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# A function that produces predicted values using the \"best slopes\" found by minimizing the RMSE\n",
    "# We will use this function to make \n",
    "def fit(row):\n",
    "    return sum(best_slopes * np.array(tuple(row)))\n",
    "\n",
    "# Add the predicted values to the test set, and visualize how close the predictions are\n",
    "test.with_column('Fitted', test.drop(0).apply(fit)).scatter('Fitted', 0)\n",
    "plots.plot([0, 5e5], [0, 5e5]);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a residual plot to see if there are patterns in the data that are unexplained\n",
    "test.with_column('Residual', test_prices-test.drop(0).apply(fit)).scatter(0, 'Residual')\n",
    "plots.plot([0, 7e5], [0, 0]);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>SalePrice</th> <th>1st Flr SF</th> <th>2nd Flr SF</th> <th>Total Bsmt SF</th> <th>Garage Area</th> <th>Year Built</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>155000   </td> <td>990       </td> <td>0         </td> <td>990          </td> <td>528        </td> <td>1994      </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>168000   </td> <td>1437      </td> <td>0         </td> <td>1188         </td> <td>576        </td> <td>1956      </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>320000   </td> <td>1699      </td> <td>1523      </td> <td>1673         </td> <td>594        </td> <td>1990      </td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>\n",
       "<p>... (998 rows omitted)</p>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's try to use nearest neighbor regression rather than a linear classifier...\n",
    "train_nn = train.select(0, 1, 2, 3, 4, 8)\n",
    "test_nn = test.select(0, 1, 2, 3, 4, 8)\n",
    "train_nn.show(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the functions we need for our nearest neighbor classifier...\n",
    "\n",
    "def distance(pt1, pt2):\n",
    "    \"\"\"The distance between two points, represented as arrays.\"\"\"\n",
    "    return np.sqrt(sum((pt1 - pt2) ** 2))\n",
    "\n",
    "def row_distance(row1, row2):\n",
    "    \"\"\"The distance between two rows of a table.\"\"\"\n",
    "    return distance(np.array(tuple(row1)), np.array(tuple(row2)))\n",
    "\n",
    "def distances(training, example, output):\n",
    "    \"\"\"Compute the distance from example for each row in training.\"\"\"\n",
    "    dists = []\n",
    "    attributes = training.drop(output)\n",
    "    for row in attributes.rows:\n",
    "        dists.append(row_distance(row, example))\n",
    "    return training.with_column('Distance', dists)\n",
    "\n",
    "def closest(training, example, k, output):\n",
    "    \"\"\"Return a table of the k closest neighbors to example.\"\"\"\n",
    "    return distances(training, example, output).sort('Distance').take(np.arange(k))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Row(SalePrice=228000, 1st Flr SF=1592, 2nd Flr SF=0, Total Bsmt SF=1556.0, Garage Area=484.0, Year Built=1994)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "    <thead>\n",
       "        <tr>\n",
       "            <th>SalePrice</th> <th>1st Flr SF</th> <th>2nd Flr SF</th> <th>Total Bsmt SF</th> <th>Garage Area</th> <th>Year Built</th> <th>Distance</th>\n",
       "        </tr>\n",
       "    </thead>\n",
       "    <tbody>\n",
       "        <tr>\n",
       "            <td>214000   </td> <td>1588      </td> <td>0         </td> <td>1588         </td> <td>472        </td> <td>1999      </td> <td>34.7707 </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>167300   </td> <td>1587      </td> <td>0         </td> <td>1587         </td> <td>525        </td> <td>1971      </td> <td>56.5332 </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>205000   </td> <td>1610      </td> <td>0         </td> <td>1610         </td> <td>480        </td> <td>1977      </td> <td>59.5399 </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>213000   </td> <td>1621      </td> <td>0         </td> <td>1621         </td> <td>478        </td> <td>1975      </td> <td>73.9121 </td>\n",
       "        </tr>\n",
       "        <tr>\n",
       "            <td>214000   </td> <td>1536      </td> <td>0         </td> <td>1536         </td> <td>532        </td> <td>2002      </td> <td>76.8375 </td>\n",
       "        </tr>\n",
       "    </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "SalePrice | 1st Flr SF | 2nd Flr SF | Total Bsmt SF | Garage Area | Year Built | Distance\n",
       "214000    | 1588       | 0          | 1588          | 472         | 1999       | 34.7707\n",
       "167300    | 1587       | 0          | 1587          | 525         | 1971       | 56.5332\n",
       "205000    | 1610       | 0          | 1610          | 480         | 1977       | 59.5399\n",
       "213000    | 1621       | 0          | 1621          | 478         | 1975       | 73.9121\n",
       "214000    | 1536       | 0          | 1536          | 532         | 2002       | 76.8375"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's get the 5 closest houses to the home in the first row of the test set\n",
    "example_nn_row = test_nn.drop(0).row(0)\n",
    "print(test_nn.row(0))\n",
    "closest(train_nn, example_nn_row, 5, 'SalePrice')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "202660.0"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's create a prediction function that returns the average house value of the most similar houses\n",
    "def predict_nn(example):\n",
    "    \"\"\"Return the majority class among the k nearest neighbors.\"\"\"\n",
    "    return np.average(closest(train_nn, example, 5, 'SalePrice').column('SalePrice'))\n",
    "\n",
    "predict_nn(example_nn_row)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Actual sale price: 228000\n",
      "Predicted sale price using nearest neighbors: 202660.0\n"
     ]
    }
   ],
   "source": [
    "print('Actual sale price:', test_nn.column('SalePrice').item(0))\n",
    "print('Predicted sale price using nearest neighbors:', predict_nn(example_nn_row))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set RMSE for multiple linear regression: $32413.22\n",
      "Test set RMSE for nearest neighbor regression: $33640.39\n"
     ]
    }
   ],
   "source": [
    "# Let's calculate the RMSE on the test set for the kNN classifier\n",
    "nn_test_predictions = test_nn.drop('SalePrice').apply(predict_nn)\n",
    "rmse_nn = np.mean((test_prices - nn_test_predictions) ** 2) ** 0.5\n",
    "\n",
    "# Which is better, linear regression of kNN? \n",
    "print('Test set RMSE for multiple linear regression: $%.2f' % rmse_linear)\n",
    "print('Test set RMSE for nearest neighbor regression: $%.2f' % rmse_nn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's look at the pattern of residuals again\n",
    "test.with_column('Residual', test_prices-nn_test_predictions).scatter(0, 'Residual')\n",
    "plots.plot([0, 7e5], [0, 0]);"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}