-{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"irma_modeling.ipynb","provenance":[],"collapsed_sections":[],"toc_visible":true,"mount_file_id":"1Q4J-Du4O02VX-aMhBgnqA8CcQVFolx1x","authorship_tag":"ABX9TyPO/sXn6d3j8rTj+S49aF7/"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"dpQIA2trCnVq"},"source":["# Hurricane Irma Damage Assessment Modeling\n","\n","In this notebook we develop a multilinear and decision tree model to assess damages from hurricane Irma.\n","\n","Our dataset consists of weather, socioeconomic, and Twitter parameters from 49 counties in Florida during Irma. The dependent variable is the amount of federal aid from FEMA that a county received, `dmg`."]},{"cell_type":"code","metadata":{"id":"ShFs31gOCjKO","executionInfo":{"status":"ok","timestamp":1603605566570,"user_tz":240,"elapsed":1032,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"2c6c1ef5-6a1e-4237-e7dd-8b113aecf266","colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["%cd '/content/drive/My Drive/Colab Notebooks/disaster_assessment/irma_modeling'\n","\n","import pandas as pd\n","import numpy as np\n","from matplotlib import pyplot as plt\n","import seaborn as sns; sns.set()\n","\n","df = pd.read_csv('irma.csv')\n","df.set_index(keys='county',inplace=True)\n","df"],"execution_count":1,"outputs":[{"output_type":"stream","text":["/content/drive/My Drive/Colab Notebooks/disaster_assessment/irma_modeling\n"],"name":"stdout"},{"output_type":"execute_result","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>maxwind_mean</th>\n"," <th>maxwind_peak</th>\n"," <th>precip_total</th>\n"," <th>precip_peak</th>\n"," <th>mhi</th>\n"," <th>poverty_rate</th>\n"," <th>poverty_pop</th>\n"," <th>population</th>\n"," <th>gdp</th>\n"," <th>twt_total</th>\n"," <th>twt_peak</th>\n"," <th>dmg</th>\n"," </tr>\n"," <tr>\n"," <th>county</th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," </tr>\n"," </thead>\n"," <tbody>\n"," <tr>\n"," <th>Alachua County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>45230</td>\n"," <td>21.2</td>\n"," <td>53816</td>\n"," <td>269956</td>\n"," <td>11912080</td>\n"," <td>182</td>\n"," <td>39</td>\n"," <td>9.306765e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Baker County</th>\n"," <td>10.200000</td>\n"," <td>24.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>51856</td>\n"," <td>15.3</td>\n"," <td>3900</td>\n"," <td>28355</td>\n"," <td>472948</td>\n"," <td>1</td>\n"," <td>1</td>\n"," <td>1.360526e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Bradford County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>44997</td>\n"," <td>19.1</td>\n"," <td>4533</td>\n"," <td>27732</td>\n"," <td>532487</td>\n"," <td>4</td>\n"," <td>1</td>\n"," <td>3.246248e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Brevard County</th>\n"," <td>14.900000</td>\n"," <td>41.0</td>\n"," <td>4.69</td>\n"," <td>2.25</td>\n"," <td>52596</td>\n"," <td>12.4</td>\n"," <td>72303</td>\n"," <td>596849</td>\n"," <td>20453753</td>\n"," <td>354</td>\n"," <td>71</td>\n"," <td>3.192891e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Broward County</th>\n"," <td>15.680000</td>\n"," <td>49.0</td>\n"," <td>10.44</td>\n"," <td>4.11</td>\n"," <td>56702</td>\n"," <td>13.1</td>\n"," <td>252288</td>\n"," <td>1951260</td>\n"," <td>96591919</td>\n"," <td>2036</td>\n"," <td>382</td>\n"," <td>1.329551e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Charlotte County</th>\n"," <td>17.666667</td>\n"," <td>44.1</td>\n"," <td>1.69</td>\n"," <td>0.76</td>\n"," <td>51583</td>\n"," <td>10.8</td>\n"," <td>19300</td>\n"," <td>184998</td>\n"," <td>3966314</td>\n"," <td>72</td>\n"," <td>20</td>\n"," <td>6.625988e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Citrus County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>43147</td>\n"," <td>14.4</td>\n"," <td>20654</td>\n"," <td>147929</td>\n"," <td>3335110</td>\n"," <td>67</td>\n"," <td>13</td>\n"," <td>5.568339e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Clay County</th>\n"," <td>10.200000</td>\n"," <td>24.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>65375</td>\n"," <td>9.9</td>\n"," <td>20889</td>\n"," <td>216072</td>\n"," <td>3976007</td>\n"," <td>53</td>\n"," <td>8</td>\n"," <td>1.219590e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Collier County</th>\n"," <td>18.866667</td>\n"," <td>63.9</td>\n"," <td>5.12</td>\n"," <td>4.20</td>\n"," <td>66709</td>\n"," <td>11.7</td>\n"," <td>43075</td>\n"," <td>378488</td>\n"," <td>16124953</td>\n"," <td>231</td>\n"," <td>63</td>\n"," <td>6.194133e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Columbia County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>42097</td>\n"," <td>16.5</td>\n"," <td>10720</td>\n"," <td>70503</td>\n"," <td>1826541</td>\n"," <td>14</td>\n"," <td>4</td>\n"," <td>3.225476e+06</td>\n"," </tr>\n"," <tr>\n"," <th>DeSoto County</th>\n"," <td>17.666667</td>\n"," <td>44.1</td>\n"," <td>1.69</td>\n"," <td>0.76</td>\n"," <td>37342</td>\n"," <td>26.1</td>\n"," <td>8766</td>\n"," <td>37489</td>\n"," <td>735286</td>\n"," <td>5</td>\n"," <td>2</td>\n"," <td>4.049290e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Dixie County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>38355</td>\n"," <td>24.4</td>\n"," <td>3627</td>\n"," <td>16700</td>\n"," <td>178261</td>\n"," <td>7</td>\n"," <td>5</td>\n"," <td>9.079057e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Duval County</th>\n"," <td>15.626667</td>\n"," <td>42.9</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>52105</td>\n"," <td>15.1</td>\n"," <td>138069</td>\n"," <td>950181</td>\n"," <td>60146765</td>\n"," <td>772</td>\n"," <td>137</td>\n"," <td>4.740938e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Flagler County</th>\n"," <td>15.920000</td>\n"," <td>44.1</td>\n"," <td>8.66</td>\n"," <td>4.66</td>\n"," <td>52713</td>\n"," <td>12.0</td>\n"," <td>13137</td>\n"," <td>112067</td>\n"," <td>1809151</td>\n"," <td>63</td>\n"," <td>16</td>\n"," <td>7.243310e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Gilchrist County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>42880</td>\n"," <td>16.1</td>\n"," <td>2675</td>\n"," <td>18256</td>\n"," <td>254260</td>\n"," <td>3</td>\n"," <td>1</td>\n"," <td>6.418368e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Glades County</th>\n"," <td>14.053333</td>\n"," <td>40.0</td>\n"," <td>9.96</td>\n"," <td>7.09</td>\n"," <td>42865</td>\n"," <td>18.9</td>\n"," <td>2312</td>\n"," <td>13724</td>\n"," <td>171573</td>\n"," <td>4</td>\n"," <td>2</td>\n"," <td>1.684916e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hamilton County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>35262</td>\n"," <td>24.0</td>\n"," <td>2791</td>\n"," <td>14310</td>\n"," <td>377309</td>\n"," <td>5</td>\n"," <td>3</td>\n"," <td>7.796505e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Hardee County</th>\n"," <td>18.146667</td>\n"," <td>54.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>40056</td>\n"," <td>23.3</td>\n"," <td>6026</td>\n"," <td>27245</td>\n"," <td>893349</td>\n"," <td>8</td>\n"," <td>3</td>\n"," <td>6.791781e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hendry County</th>\n"," <td>18.866667</td>\n"," <td>63.9</td>\n"," <td>12.61</td>\n"," <td>8.84</td>\n"," <td>38361</td>\n"," <td>23.9</td>\n"," <td>9525</td>\n"," <td>41556</td>\n"," <td>1241872</td>\n"," <td>2</td>\n"," <td>1</td>\n"," <td>4.864095e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hernando County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>44710</td>\n"," <td>14.0</td>\n"," <td>25773</td>\n"," <td>190865</td>\n"," <td>3031267</td>\n"," <td>43</td>\n"," <td>12</td>\n"," <td>6.987439e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Highlands County</th>\n"," <td>18.146667</td>\n"," <td>54.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>37445</td>\n"," <td>19.8</td>\n"," <td>20051</td>\n"," <td>105424</td>\n"," <td>2088782</td>\n"," <td>46</td>\n"," <td>16</td>\n"," <td>1.104181e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Hillsborough County</th>\n"," <td>10.140000</td>\n"," <td>27.0</td>\n"," <td>5.43</td>\n"," <td>4.50</td>\n"," <td>54741</td>\n"," <td>15.5</td>\n"," <td>214442</td>\n"," <td>1436888</td>\n"," <td>77093796</td>\n"," <td>990</td>\n"," <td>234</td>\n"," <td>4.139342e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Indian River County</th>\n"," <td>16.033333</td>\n"," <td>49.9</td>\n"," <td>10.86</td>\n"," <td>8.11</td>\n"," <td>51797</td>\n"," <td>10.6</td>\n"," <td>16249</td>\n"," <td>157413</td>\n"," <td>5001702</td>\n"," <td>79</td>\n"," <td>23</td>\n"," <td>5.417251e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Lafayette County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>41549</td>\n"," <td>22.7</td>\n"," <td>1579</td>\n"," <td>8732</td>\n"," <td>160614</td>\n"," <td>2</td>\n"," <td>1</td>\n"," <td>4.387254e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Lake County</th>\n"," <td>13.760000</td>\n"," <td>36.9</td>\n"," <td>10.51</td>\n"," <td>8.05</td>\n"," <td>51429</td>\n"," <td>12.6</td>\n"," <td>43020</td>\n"," <td>356495</td>\n"," <td>7452383</td>\n"," <td>151</td>\n"," <td>47</td>\n"," <td>1.418739e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Lee County</th>\n"," <td>15.206667</td>\n"," <td>49.9</td>\n"," <td>6.08</td>\n"," <td>1.89</td>\n"," <td>54198</td>\n"," <td>11.8</td>\n"," <td>85844</td>\n"," <td>754610</td>\n"," <td>23806704</td>\n"," <td>380</td>\n"," <td>63</td>\n"," <td>6.269670e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Levy County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>37272</td>\n"," <td>20.8</td>\n"," <td>8329</td>\n"," <td>40770</td>\n"," <td>636701</td>\n"," <td>7</td>\n"," <td>2</td>\n"," <td>1.823098e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Manatee County</th>\n"," <td>15.620000</td>\n"," <td>42.0</td>\n"," <td>7.08</td>\n"," <td>5.71</td>\n"," <td>55189</td>\n"," <td>10.8</td>\n"," <td>41057</td>\n"," <td>394855</td>\n"," <td>11968028</td>\n"," <td>170</td>\n"," <td>33</td>\n"," <td>1.369299e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Marion County</th>\n"," <td>11.150000</td>\n"," <td>32.1</td>\n"," <td>0.44</td>\n"," <td>0.23</td>\n"," <td>43772</td>\n"," <td>16.2</td>\n"," <td>55880</td>\n"," <td>359977</td>\n"," <td>7956019</td>\n"," <td>121</td>\n"," <td>30</td>\n"," <td>1.934269e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Martin County</th>\n"," <td>12.593333</td>\n"," <td>22.0</td>\n"," <td>16.32</td>\n"," <td>9.43</td>\n"," <td>58344</td>\n"," <td>10.9</td>\n"," <td>17002</td>\n"," <td>160912</td>\n"," <td>6533103</td>\n"," <td>85</td>\n"," <td>22</td>\n"," <td>2.936391e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Miami-Dade County</th>\n"," <td>15.900000</td>\n"," <td>42.9</td>\n"," <td>9.33</td>\n"," <td>5.18</td>\n"," <td>49758</td>\n"," <td>16.7</td>\n"," <td>452649</td>\n"," <td>2761581</td>\n"," <td>141734334</td>\n"," <td>4063</td>\n"," <td>730</td>\n"," <td>2.417464e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Monroe County</th>\n"," <td>15.900000</td>\n"," <td>42.9</td>\n"," <td>9.33</td>\n"," <td>5.18</td>\n"," <td>63009</td>\n"," <td>11.8</td>\n"," <td>8963</td>\n"," <td>75027</td>\n"," <td>4097511</td>\n"," <td>158</td>\n"," <td>31</td>\n"," <td>1.061900e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Nassau County</th>\n"," <td>14.566667</td>\n"," <td>42.9</td>\n"," <td>10.05</td>\n"," <td>9.01</td>\n"," <td>70590</td>\n"," <td>9.1</td>\n"," <td>7484</td>\n"," <td>85832</td>\n"," <td>1886261</td>\n"," <td>49</td>\n"," <td>16</td>\n"," <td>4.184525e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Okeechobee County</th>\n"," <td>14.053333</td>\n"," <td>40.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>42524</td>\n"," <td>21.8</td>\n"," <td>8415</td>\n"," <td>41537</td>\n"," <td>1021477</td>\n"," <td>7</td>\n"," <td>2</td>\n"," <td>2.920370e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Orange County</th>\n"," <td>16.326667</td>\n"," <td>44.1</td>\n"," <td>10.36</td>\n"," <td>7.48</td>\n"," <td>54021</td>\n"," <td>15.3</td>\n"," <td>201528</td>\n"," <td>1380645</td>\n"," <td>89817807</td>\n"," <td>1794</td>\n"," <td>346</td>\n"," <td>5.216380e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Osceola County</th>\n"," <td>19.046667</td>\n"," <td>49.9</td>\n"," <td>13.43</td>\n"," <td>7.02</td>\n"," <td>49284</td>\n"," <td>14.0</td>\n"," <td>48892</td>\n"," <td>367990</td>\n"," <td>9207981</td>\n"," <td>205</td>\n"," <td>50</td>\n"," <td>1.244058e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Palm Beach County</th>\n"," <td>18.200000</td>\n"," <td>51.1</td>\n"," <td>3.96</td>\n"," <td>1.92</td>\n"," <td>60059</td>\n"," <td>11.8</td>\n"," <td>170868</td>\n"," <td>1485941</td>\n"," <td>76866505</td>\n"," <td>904</td>\n"," <td>167</td>\n"," <td>8.307278e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Pasco County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>51247</td>\n"," <td>13.0</td>\n"," <td>67635</td>\n"," <td>539630</td>\n"," <td>9330553</td>\n"," <td>206</td>\n"," <td>58</td>\n"," <td>1.824896e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Pinellas County</th>\n"," <td>15.446667</td>\n"," <td>42.9</td>\n"," <td>5.85</td>\n"," <td>4.32</td>\n"," <td>51488</td>\n"," <td>12.2</td>\n"," <td>115990</td>\n"," <td>975280</td>\n"," <td>44125945</td>\n"," <td>730</td>\n"," <td>159</td>\n"," <td>5.632790e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Polk County</th>\n"," <td>17.544444</td>\n"," <td>36.9</td>\n"," <td>8.06</td>\n"," <td>6.21</td>\n"," <td>48328</td>\n"," <td>16.1</td>\n"," <td>107844</td>\n"," <td>708009</td>\n"," <td>20779632</td>\n"," <td>524</td>\n"," <td>116</td>\n"," <td>4.351088e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Putnam County</th>\n"," <td>16.980000</td>\n"," <td>45.1</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>34390</td>\n"," <td>26.3</td>\n"," <td>18954</td>\n"," <td>74163</td>\n"," <td>1925314</td>\n"," <td>16</td>\n"," <td>4</td>\n"," <td>1.110617e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Sarasota County</th>\n"," <td>15.620000</td>\n"," <td>45.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>77022</td>\n"," <td>8.3</td>\n"," <td>20118</td>\n"," <td>254261</td>\n"," <td>7313073</td>\n"," <td>243</td>\n"," <td>68</td>\n"," <td>1.306549e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Seminole County</th>\n"," <td>16.740000</td>\n"," <td>49.9</td>\n"," <td>16.32</td>\n"," <td>9.43</td>\n"," <td>49995</td>\n"," <td>12.8</td>\n"," <td>39839</td>\n"," <td>321128</td>\n"," <td>7030683</td>\n"," <td>331</td>\n"," <td>81</td>\n"," <td>2.173058e+07</td>\n"," </tr>\n"," <tr>\n"," <th>St. Johns County</th>\n"," <td>16.980000</td>\n"," <td>42.0</td>\n"," <td>7.08</td>\n"," <td>5.71</td>\n"," <td>58423</td>\n"," <td>9.2</td>\n"," <td>38065</td>\n"," <td>426718</td>\n"," <td>15773229</td>\n"," <td>154</td>\n"," <td>34</td>\n"," <td>1.862228e+07</td>\n"," </tr>\n"," <tr>\n"," <th>St. Lucie County</th>\n"," <td>16.033333</td>\n"," <td>44.1</td>\n"," <td>11.28</td>\n"," <td>9.24</td>\n"," <td>63865</td>\n"," <td>11.2</td>\n"," <td>51321</td>\n"," <td>467832</td>\n"," <td>17902542</td>\n"," <td>112</td>\n"," <td>32</td>\n"," <td>2.350643e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Sumter County</th>\n"," <td>13.760000</td>\n"," <td>36.9</td>\n"," <td>10.51</td>\n"," <td>8.05</td>\n"," <td>57931</td>\n"," <td>9.1</td>\n"," <td>10672</td>\n"," <td>128754</td>\n"," <td>2312425</td>\n"," <td>21</td>\n"," <td>7</td>\n"," <td>3.424526e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Suwannee County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>44144</td>\n"," <td>20.3</td>\n"," <td>8299</td>\n"," <td>44191</td>\n"," <td>851332</td>\n"," <td>12</td>\n"," <td>4</td>\n"," <td>2.045323e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Union County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>47373</td>\n"," <td>22.2</td>\n"," <td>2291</td>\n"," <td>14940</td>\n"," <td>341763</td>\n"," <td>3</td>\n"," <td>3</td>\n"," <td>5.774985e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Volusia County</th>\n"," <td>16.333333</td>\n"," <td>35.0</td>\n"," <td>8.66</td>\n"," <td>4.66</td>\n"," <td>46911</td>\n"," <td>15.2</td>\n"," <td>79877</td>\n"," <td>547538</td>\n"," <td>14864234</td>\n"," <td>266</td>\n"," <td>51</td>\n"," <td>3.782572e+07</td>\n"," </tr>\n"," </tbody>\n","</table>\n","</div>"],"text/plain":[" maxwind_mean maxwind_peak ... twt_peak dmg\n","county ... \n","Alachua County 12.453333 35.0 ... 39 9.306765e+06\n","Baker County 10.200000 24.1 ... 1 1.360526e+06\n","Bradford County 12.453333 35.0 ... 1 3.246248e+06\n","Brevard County 14.900000 41.0 ... 71 3.192891e+07\n","Broward County 15.680000 49.0 ... 382 1.329551e+08\n","Charlotte County 17.666667 44.1 ... 20 6.625988e+06\n","Citrus County 14.313333 33.0 ... 13 5.568339e+06\n","Clay County 10.200000 24.1 ... 8 1.219590e+07\n","Collier County 18.866667 63.9 ... 63 6.194133e+07\n","Columbia County 12.453333 35.0 ... 4 3.225476e+06\n","DeSoto County 17.666667 44.1 ... 2 4.049290e+06\n","Dixie County 13.393333 35.9 ... 5 9.079057e+05\n","Duval County 15.626667 42.9 ... 137 4.740938e+07\n","Flagler County 15.920000 44.1 ... 16 7.243310e+06\n","Gilchrist County 13.393333 35.9 ... 1 6.418368e+05\n","Glades County 14.053333 40.0 ... 2 1.684916e+06\n","Hamilton County 11.066667 20.0 ... 3 7.796505e+05\n","Hardee County 18.146667 54.0 ... 3 6.791781e+06\n","Hendry County 18.866667 63.9 ... 1 4.864095e+06\n","Hernando County 14.313333 33.0 ... 12 6.987439e+06\n","Highlands County 18.146667 54.0 ... 16 1.104181e+07\n","Hillsborough County 10.140000 27.0 ... 234 4.139342e+07\n","Indian River County 16.033333 49.9 ... 23 5.417251e+06\n","Lafayette County 11.066667 20.0 ... 1 4.387254e+05\n","Lake County 13.760000 36.9 ... 47 1.418739e+07\n","Lee County 15.206667 49.9 ... 63 6.269670e+07\n","Levy County 13.393333 35.9 ... 2 1.823098e+06\n","Manatee County 15.620000 42.0 ... 33 1.369299e+07\n","Marion County 11.150000 32.1 ... 30 1.934269e+07\n","Martin County 12.593333 22.0 ... 22 2.936391e+06\n","Miami-Dade County 15.900000 42.9 ... 730 2.417464e+08\n","Monroe County 15.900000 42.9 ... 31 1.061900e+08\n","Nassau County 14.566667 42.9 ... 16 4.184525e+06\n","Okeechobee County 14.053333 40.0 ... 2 2.920370e+06\n","Orange County 16.326667 44.1 ... 346 5.216380e+07\n","Osceola County 19.046667 49.9 ... 50 1.244058e+07\n","Palm Beach County 18.200000 51.1 ... 167 8.307278e+07\n","Pasco County 14.313333 33.0 ... 58 1.824896e+07\n","Pinellas County 15.446667 42.9 ... 159 5.632790e+07\n","Polk County 17.544444 36.9 ... 116 4.351088e+07\n","Putnam County 16.980000 45.1 ... 4 1.110617e+07\n","Sarasota County 15.620000 45.1 ... 68 1.306549e+07\n","Seminole County 16.740000 49.9 ... 81 2.173058e+07\n","St. Johns County 16.980000 42.0 ... 34 1.862228e+07\n","St. Lucie County 16.033333 44.1 ... 32 2.350643e+07\n","Sumter County 13.760000 36.9 ... 7 3.424526e+06\n","Suwannee County 11.066667 20.0 ... 4 2.045323e+06\n","Union County 12.453333 35.0 ... 3 5.774985e+05\n","Volusia County 16.333333 35.0 ... 51 3.782572e+07\n","\n","[49 rows x 12 columns]"]},"metadata":{"tags":[]},"execution_count":1}]},{"cell_type":"markdown","metadata":{"id":"Fg-Ticrhdpih"},"source":["## Feature Selection\n","\n","We have 11 features, but do we really need them all? Especially when we only have 49 data points. As the following heat map shows, a lot of the variables correlate with each other, forming distinct clusters."]},{"cell_type":"code","metadata":{"id":"B-5M2buOeaQV","executionInfo":{"status":"ok","timestamp":1603605567553,"user_tz":240,"elapsed":1990,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"03ab3528-f9cc-4e71-f9f3-7c6e0c94d69c","colab":{"base_uri":"https://localhost:8080/","height":512}},"source":["plt.figure(figsize=(12,7))\n","sns.heatmap(df.corr(),\n"," annot=True,\n"," fmt = '.2f',\n"," cmap='coolwarm')\n","plt.show()"],"execution_count":2,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 864x504 with 2 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"SWRPvSoZQ341"},"source":["We can identify these clusters as follows:\n","\n","- Wind: `maxwind_mean`, `maxwind_peak`\n","- Precipitation: `precip_total`, `precip_peak`\n","- Socioeconomic: `mhi`, `poverty_rate`\n","- Population: `poverty_pop`, `population`, `gdp`, `twt_total`, `twt_peak`\n","\n","Our goal is to predict `dmg`. To avoid high multicollinearity among our predictors, let's simply take one variable from each cluster. We'll pick the ones with the strongest correlation to `dmg`. This gives us `maxwind_peak`, `precip_peak`, `mhi`, `twt_total`.\n","\n","Let's verify that these variables are independent of eacher using **Variance Inflation Factor (VIF)**. The VIF of an independent variable is a measure of how much its variation can be attributed to other independent variables. The higher it is, the more redundant the (not so) independent variable is. VIF is calculated as follows:\n","\n","$$ VIF = \\frac{1}{1 - R^2_i} $$\n","\n","where $R_i^2$ is the $R^2$ of the multilinear regression model of the $i^{\\text{th}}$ independent variable using the other independent variables as predictors. Why not just use $R^2_i$ to measure redundancy? One intuition is that the inverse relationship will harshly penalize smaller and smaller values of $1 - R^2_i$ (the so-called **tolerance**).\n","\n","Let's compute the VIF for each of the four predictors we're using:"]},{"cell_type":"code","metadata":{"id":"nnUg2sP6-DIk","executionInfo":{"status":"ok","timestamp":1603605568034,"user_tz":240,"elapsed":2453,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"e055b869-8298-4a60-e0fd-5f89f120ed63","colab":{"base_uri":"https://localhost:8080/","height":144}},"source":["from sklearn.preprocessing import robust_scale\n","from statsmodels.stats.outliers_influence import variance_inflation_factor as vif\n","import statsmodels.api as sm\n","\n","# preprocess data\n","features = ['maxwind_peak', 'precip_peak', 'mhi', 'twt_total']\n","X = df[features].apply(robust_scale)\n","X_vif = sm.add_constant(X.values) # for the vif func\n","\n","vif_dict = {}\n","for f in features:\n"," idx = features.index(f) # need this for vif func\n"," vif_dict[f] = vif(X_vif, idx)\n","vif_dict"],"execution_count":3,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n"," import pandas.util.testing as tm\n"],"name":"stderr"},{"output_type":"execute_result","data":{"text/plain":["{'maxwind_peak': 1.1609234512973168,\n"," 'mhi': 1.0514008951220526,\n"," 'precip_peak': 1.0447003684876797,\n"," 'twt_total': 1.097731476287398}"]},"metadata":{"tags":[]},"execution_count":3}]},{"cell_type":"markdown","metadata":{"id":"jFmHET7jZGLm"},"source":["This is good! The typical threshhold for VIF's is 10, sometimes 5. All the VIF's calculated were close to 1, indicating that the variables are very much independent of each other.\n","\n","## Modeling\n","Now that we have settled on our features, let's model our data. We will try out a linear regression and decision tree model and see which comes out best."]},{"cell_type":"code","metadata":{"id":"XHNL9NDLGKy6","executionInfo":{"status":"ok","timestamp":1603605568039,"user_tz":240,"elapsed":2442,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}}},"source":["from sklearn.model_selection import train_test_split\n","\n","Y = robust_scale(df['dmg'])\n","\n","X_train, X_test, Y_train, Y_test = train_test_split(X, Y,\n"," test_size = 0.1,\n"," random_state = 0,\n"," shuffle = True)"],"execution_count":4,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"SpztfD5hGLd1"},"source":["### Linear Regression\n","\n","First, we will run cross validation on linear regressors and find the best performing model."]},{"cell_type":"code","metadata":{"id":"0ONf_rgoZhOS","executionInfo":{"status":"ok","timestamp":1603605568041,"user_tz":240,"elapsed":2429,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"3121c275-e795-4bef-d9e8-792b1a27ddd9","colab":{"base_uri":"https://localhost:8080/","height":195}},"source":["from sklearn.model_selection import cross_validate, KFold\n","from sklearn.linear_model import LinearRegression\n","\n","reg = LinearRegression()\n","kf = KFold(5)\n","cv_scores = cross_validate(reg,\n"," X_train, Y_train,\n"," cv = kf,\n"," scoring = 'r2',\n"," return_train_score = True,\n"," return_estimator = True)\n","\n","best_score = np.max(cv_scores['test_score'])\n","best_idx = np.where(cv_scores['test_score'] == best_score)[0][0] # need the [0][0] to get the idx\n","linreg_model = cv_scores['estimator'][best_idx]\n","print('Best CV r2: %.2f' % best_score)\n","cv_scores"],"execution_count":5,"outputs":[{"output_type":"stream","text":["Best CV r2: 0.81\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["{'estimator': (LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False),\n"," LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False),\n"," LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False),\n"," LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False),\n"," LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)),\n"," 'fit_time': array([0.00223398, 0.00137377, 0.00139618, 0.00130749, 0.00127864]),\n"," 'score_time': array([0.00135279, 0.00084329, 0.00083733, 0.00079083, 0.00079751]),\n"," 'test_score': array([-0.17218448, -1.99457306, 0.57294813, 0.81115215, 0.6499881 ]),\n"," 'train_score': array([0.80465442, 0.62029971, 0.54761734, 0.42716121, 0.54831034])}"]},"metadata":{"tags":[]},"execution_count":5}]},{"cell_type":"markdown","metadata":{"id":"J4WYVukfMbh2"},"source":["### Decision Tree\n","Now let's find the optimal decision tree model."]},{"cell_type":"code","metadata":{"id":"sS8der1cMoYP","executionInfo":{"status":"ok","timestamp":1603605568050,"user_tz":240,"elapsed":2421,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"2254943f-f274-4d1b-efdc-6a11aa0ac136","colab":{"base_uri":"https://localhost:8080/","height":639}},"source":["from sklearn.tree import DecisionTreeRegressor\n","\n","tree = DecisionTreeRegressor(random_state = 0)\n","\n","cv_scores = cross_validate(tree,\n"," X_train, Y_train,\n"," cv = kf,\n"," scoring = 'r2',\n"," return_train_score = True,\n"," return_estimator = True)\n","\n","best_score = np.max(cv_scores['test_score'])\n","best_idx = np.where(cv_scores['test_score'] == best_score)[0][0]\n","dt_model = cv_scores['estimator'][best_idx]\n","print(\"Best CV r2: %.2f\" % best_score)\n","cv_scores"],"execution_count":6,"outputs":[{"output_type":"stream","text":["Best CV r2: 0.82\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["{'estimator': (DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,\n"," max_features=None, max_leaf_nodes=None,\n"," min_impurity_decrease=0.0, min_impurity_split=None,\n"," min_samples_leaf=1, min_samples_split=2,\n"," min_weight_fraction_leaf=0.0, presort='deprecated',\n"," random_state=0, splitter='best'),\n"," DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,\n"," max_features=None, max_leaf_nodes=None,\n"," min_impurity_decrease=0.0, min_impurity_split=None,\n"," min_samples_leaf=1, min_samples_split=2,\n"," min_weight_fraction_leaf=0.0, presort='deprecated',\n"," random_state=0, splitter='best'),\n"," DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,\n"," max_features=None, max_leaf_nodes=None,\n"," min_impurity_decrease=0.0, min_impurity_split=None,\n"," min_samples_leaf=1, min_samples_split=2,\n"," min_weight_fraction_leaf=0.0, presort='deprecated',\n"," random_state=0, splitter='best'),\n"," DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,\n"," max_features=None, max_leaf_nodes=None,\n"," min_impurity_decrease=0.0, min_impurity_split=None,\n"," min_samples_leaf=1, min_samples_split=2,\n"," min_weight_fraction_leaf=0.0, presort='deprecated',\n"," random_state=0, splitter='best'),\n"," DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,\n"," max_features=None, max_leaf_nodes=None,\n"," min_impurity_decrease=0.0, min_impurity_split=None,\n"," min_samples_leaf=1, min_samples_split=2,\n"," min_weight_fraction_leaf=0.0, presort='deprecated',\n"," random_state=0, splitter='best')),\n"," 'fit_time': array([0.00195765, 0.00157475, 0.00157881, 0.00152278, 0.00159955]),\n"," 'score_time': array([0.00102186, 0.00093675, 0.00092006, 0.00089002, 0.00084209]),\n"," 'test_score': array([ 0.25140641, -1.84462954, 0.17624502, 0.82366205, 0.04245883]),\n"," 'train_score': array([1., 1., 1., 1., 1.])}"]},"metadata":{"tags":[]},"execution_count":6}]},{"cell_type":"markdown","metadata":{"id":"g0Jhv1mSQ4gA"},"source":["The decision tree appears to be grossly overfitting, as the last two rows show. The train $R^2$ values are all 1! Let's see how these two models compare on the test set..."]},{"cell_type":"code","metadata":{"id":"cLZDH08SRgyS","executionInfo":{"status":"ok","timestamp":1603605568054,"user_tz":240,"elapsed":2409,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"1af108b9-1275-4073-d717-065b8ab303fa","colab":{"base_uri":"https://localhost:8080/","height":141}},"source":["linreg_train_score = linreg_model.score(X_train, Y_train)\n","linreg_test_score = linreg_model.score(X_test, Y_test)\n","tree_train_score = dt_model.score(X_train, Y_train)\n","tree_test_score = dt_model.score(X_test, Y_test)\n","\n","print(\"===== Linear Regression r2 Scores =====\")\n","print(\"Train: %.2f\" % linreg_train_score)\n","print(\"Test: %.2f\" % linreg_test_score)\n","print()\n","print(\"===== Decision Tree Regressor r2 Scores =====\")\n","print(\"Train: %.2f\" % tree_train_score)\n","print(\"Test: %.2f\" % tree_test_score)"],"execution_count":7,"outputs":[{"output_type":"stream","text":["===== Linear Regression r2 Scores =====\n","Train: 0.55\n","Test: 0.65\n","\n","===== Decision Tree Regressor r2 Scores =====\n","Train: 0.95\n","Test: 0.50\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"LRrQZPrrTY-j"},"source":["Both models overfit, although the linear regressor less so than the decision tree. The test scores are also modest. What do we do now?\n","\n","Ideas:\n","\n","- Use regularization for linear regression: ridge\n","- Look at the parameters of the dt. Is there anything that looks like it should be tweaked to prevent overfitting? Maybe eg the depth is too large.\n","- Can we use another model? Random forest regression seems like a popular alternative. It will eliminate some of the explainability form DT regression, but should alleviate some of the overfitting."]}]}
\ No newline at end of file
+{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"irma_modeling.ipynb","provenance":[],"collapsed_sections":[],"toc_visible":true,"mount_file_id":"1Q4J-Du4O02VX-aMhBgnqA8CcQVFolx1x","authorship_tag":"ABX9TyO58oGmbwmCT3JpyThA0NLp"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"dpQIA2trCnVq"},"source":["# Hurricane Irma Damage Assessment Modeling\n","\n","In this notebook we develop a multilinear and decision tree model to assess damages from hurricane Irma.\n","\n","Our dataset consists of weather, socioeconomic, and Twitter parameters from 49 counties in Florida during Irma. The dependent variable is the amount of federal aid from FEMA that a county received, `dmg`."]},{"cell_type":"code","metadata":{"id":"ShFs31gOCjKO","executionInfo":{"status":"ok","timestamp":1603699821465,"user_tz":240,"elapsed":677,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"eb454c3e-526e-4920-f137-e38c4544dce1","colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["%cd '/content/drive/My Drive/Colab Notebooks/disaster_assessment/irma_modeling'\n","\n","import pandas as pd\n","import numpy as np\n","from matplotlib import pyplot as plt\n","import seaborn as sns; sns.set()\n","\n","df = pd.read_csv('irma.csv')\n","df.set_index(keys='county',inplace=True)\n","df"],"execution_count":1,"outputs":[{"output_type":"stream","text":["/content/drive/My Drive/Colab Notebooks/disaster_assessment/irma_modeling\n"],"name":"stdout"},{"output_type":"execute_result","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>maxwind_mean</th>\n"," <th>maxwind_peak</th>\n"," <th>precip_total</th>\n"," <th>precip_peak</th>\n"," <th>mhi</th>\n"," <th>poverty_rate</th>\n"," <th>poverty_pop</th>\n"," <th>population</th>\n"," <th>gdp</th>\n"," <th>twt_total</th>\n"," <th>twt_peak</th>\n"," <th>dmg</th>\n"," </tr>\n"," <tr>\n"," <th>county</th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," <th></th>\n"," </tr>\n"," </thead>\n"," <tbody>\n"," <tr>\n"," <th>Alachua County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>45230</td>\n"," <td>21.2</td>\n"," <td>53816</td>\n"," <td>269956</td>\n"," <td>11912080</td>\n"," <td>182</td>\n"," <td>39</td>\n"," <td>9.306765e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Baker County</th>\n"," <td>10.200000</td>\n"," <td>24.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>51856</td>\n"," <td>15.3</td>\n"," <td>3900</td>\n"," <td>28355</td>\n"," <td>472948</td>\n"," <td>1</td>\n"," <td>1</td>\n"," <td>1.360526e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Bradford County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>44997</td>\n"," <td>19.1</td>\n"," <td>4533</td>\n"," <td>27732</td>\n"," <td>532487</td>\n"," <td>4</td>\n"," <td>1</td>\n"," <td>3.246248e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Brevard County</th>\n"," <td>14.900000</td>\n"," <td>41.0</td>\n"," <td>4.69</td>\n"," <td>2.25</td>\n"," <td>52596</td>\n"," <td>12.4</td>\n"," <td>72303</td>\n"," <td>596849</td>\n"," <td>20453753</td>\n"," <td>354</td>\n"," <td>71</td>\n"," <td>3.192891e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Broward County</th>\n"," <td>15.680000</td>\n"," <td>49.0</td>\n"," <td>10.44</td>\n"," <td>4.11</td>\n"," <td>56702</td>\n"," <td>13.1</td>\n"," <td>252288</td>\n"," <td>1951260</td>\n"," <td>96591919</td>\n"," <td>2036</td>\n"," <td>382</td>\n"," <td>1.329551e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Charlotte County</th>\n"," <td>17.666667</td>\n"," <td>44.1</td>\n"," <td>1.69</td>\n"," <td>0.76</td>\n"," <td>51583</td>\n"," <td>10.8</td>\n"," <td>19300</td>\n"," <td>184998</td>\n"," <td>3966314</td>\n"," <td>72</td>\n"," <td>20</td>\n"," <td>6.625988e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Citrus County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>43147</td>\n"," <td>14.4</td>\n"," <td>20654</td>\n"," <td>147929</td>\n"," <td>3335110</td>\n"," <td>67</td>\n"," <td>13</td>\n"," <td>5.568339e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Clay County</th>\n"," <td>10.200000</td>\n"," <td>24.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>65375</td>\n"," <td>9.9</td>\n"," <td>20889</td>\n"," <td>216072</td>\n"," <td>3976007</td>\n"," <td>53</td>\n"," <td>8</td>\n"," <td>1.219590e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Collier County</th>\n"," <td>18.866667</td>\n"," <td>63.9</td>\n"," <td>5.12</td>\n"," <td>4.20</td>\n"," <td>66709</td>\n"," <td>11.7</td>\n"," <td>43075</td>\n"," <td>378488</td>\n"," <td>16124953</td>\n"," <td>231</td>\n"," <td>63</td>\n"," <td>6.194133e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Columbia County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>42097</td>\n"," <td>16.5</td>\n"," <td>10720</td>\n"," <td>70503</td>\n"," <td>1826541</td>\n"," <td>14</td>\n"," <td>4</td>\n"," <td>3.225476e+06</td>\n"," </tr>\n"," <tr>\n"," <th>DeSoto County</th>\n"," <td>17.666667</td>\n"," <td>44.1</td>\n"," <td>1.69</td>\n"," <td>0.76</td>\n"," <td>37342</td>\n"," <td>26.1</td>\n"," <td>8766</td>\n"," <td>37489</td>\n"," <td>735286</td>\n"," <td>5</td>\n"," <td>2</td>\n"," <td>4.049290e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Dixie County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>38355</td>\n"," <td>24.4</td>\n"," <td>3627</td>\n"," <td>16700</td>\n"," <td>178261</td>\n"," <td>7</td>\n"," <td>5</td>\n"," <td>9.079057e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Duval County</th>\n"," <td>15.626667</td>\n"," <td>42.9</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>52105</td>\n"," <td>15.1</td>\n"," <td>138069</td>\n"," <td>950181</td>\n"," <td>60146765</td>\n"," <td>772</td>\n"," <td>137</td>\n"," <td>4.740938e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Flagler County</th>\n"," <td>15.920000</td>\n"," <td>44.1</td>\n"," <td>8.66</td>\n"," <td>4.66</td>\n"," <td>52713</td>\n"," <td>12.0</td>\n"," <td>13137</td>\n"," <td>112067</td>\n"," <td>1809151</td>\n"," <td>63</td>\n"," <td>16</td>\n"," <td>7.243310e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Gilchrist County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>42880</td>\n"," <td>16.1</td>\n"," <td>2675</td>\n"," <td>18256</td>\n"," <td>254260</td>\n"," <td>3</td>\n"," <td>1</td>\n"," <td>6.418368e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Glades County</th>\n"," <td>14.053333</td>\n"," <td>40.0</td>\n"," <td>9.96</td>\n"," <td>7.09</td>\n"," <td>42865</td>\n"," <td>18.9</td>\n"," <td>2312</td>\n"," <td>13724</td>\n"," <td>171573</td>\n"," <td>4</td>\n"," <td>2</td>\n"," <td>1.684916e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hamilton County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>35262</td>\n"," <td>24.0</td>\n"," <td>2791</td>\n"," <td>14310</td>\n"," <td>377309</td>\n"," <td>5</td>\n"," <td>3</td>\n"," <td>7.796505e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Hardee County</th>\n"," <td>18.146667</td>\n"," <td>54.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>40056</td>\n"," <td>23.3</td>\n"," <td>6026</td>\n"," <td>27245</td>\n"," <td>893349</td>\n"," <td>8</td>\n"," <td>3</td>\n"," <td>6.791781e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hendry County</th>\n"," <td>18.866667</td>\n"," <td>63.9</td>\n"," <td>12.61</td>\n"," <td>8.84</td>\n"," <td>38361</td>\n"," <td>23.9</td>\n"," <td>9525</td>\n"," <td>41556</td>\n"," <td>1241872</td>\n"," <td>2</td>\n"," <td>1</td>\n"," <td>4.864095e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Hernando County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>44710</td>\n"," <td>14.0</td>\n"," <td>25773</td>\n"," <td>190865</td>\n"," <td>3031267</td>\n"," <td>43</td>\n"," <td>12</td>\n"," <td>6.987439e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Highlands County</th>\n"," <td>18.146667</td>\n"," <td>54.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>37445</td>\n"," <td>19.8</td>\n"," <td>20051</td>\n"," <td>105424</td>\n"," <td>2088782</td>\n"," <td>46</td>\n"," <td>16</td>\n"," <td>1.104181e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Hillsborough County</th>\n"," <td>10.140000</td>\n"," <td>27.0</td>\n"," <td>5.43</td>\n"," <td>4.50</td>\n"," <td>54741</td>\n"," <td>15.5</td>\n"," <td>214442</td>\n"," <td>1436888</td>\n"," <td>77093796</td>\n"," <td>990</td>\n"," <td>234</td>\n"," <td>4.139342e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Indian River County</th>\n"," <td>16.033333</td>\n"," <td>49.9</td>\n"," <td>10.86</td>\n"," <td>8.11</td>\n"," <td>51797</td>\n"," <td>10.6</td>\n"," <td>16249</td>\n"," <td>157413</td>\n"," <td>5001702</td>\n"," <td>79</td>\n"," <td>23</td>\n"," <td>5.417251e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Lafayette County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>41549</td>\n"," <td>22.7</td>\n"," <td>1579</td>\n"," <td>8732</td>\n"," <td>160614</td>\n"," <td>2</td>\n"," <td>1</td>\n"," <td>4.387254e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Lake County</th>\n"," <td>13.760000</td>\n"," <td>36.9</td>\n"," <td>10.51</td>\n"," <td>8.05</td>\n"," <td>51429</td>\n"," <td>12.6</td>\n"," <td>43020</td>\n"," <td>356495</td>\n"," <td>7452383</td>\n"," <td>151</td>\n"," <td>47</td>\n"," <td>1.418739e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Lee County</th>\n"," <td>15.206667</td>\n"," <td>49.9</td>\n"," <td>6.08</td>\n"," <td>1.89</td>\n"," <td>54198</td>\n"," <td>11.8</td>\n"," <td>85844</td>\n"," <td>754610</td>\n"," <td>23806704</td>\n"," <td>380</td>\n"," <td>63</td>\n"," <td>6.269670e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Levy County</th>\n"," <td>13.393333</td>\n"," <td>35.9</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>37272</td>\n"," <td>20.8</td>\n"," <td>8329</td>\n"," <td>40770</td>\n"," <td>636701</td>\n"," <td>7</td>\n"," <td>2</td>\n"," <td>1.823098e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Manatee County</th>\n"," <td>15.620000</td>\n"," <td>42.0</td>\n"," <td>7.08</td>\n"," <td>5.71</td>\n"," <td>55189</td>\n"," <td>10.8</td>\n"," <td>41057</td>\n"," <td>394855</td>\n"," <td>11968028</td>\n"," <td>170</td>\n"," <td>33</td>\n"," <td>1.369299e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Marion County</th>\n"," <td>11.150000</td>\n"," <td>32.1</td>\n"," <td>0.44</td>\n"," <td>0.23</td>\n"," <td>43772</td>\n"," <td>16.2</td>\n"," <td>55880</td>\n"," <td>359977</td>\n"," <td>7956019</td>\n"," <td>121</td>\n"," <td>30</td>\n"," <td>1.934269e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Martin County</th>\n"," <td>12.593333</td>\n"," <td>22.0</td>\n"," <td>16.32</td>\n"," <td>9.43</td>\n"," <td>58344</td>\n"," <td>10.9</td>\n"," <td>17002</td>\n"," <td>160912</td>\n"," <td>6533103</td>\n"," <td>85</td>\n"," <td>22</td>\n"," <td>2.936391e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Miami-Dade County</th>\n"," <td>15.900000</td>\n"," <td>42.9</td>\n"," <td>9.33</td>\n"," <td>5.18</td>\n"," <td>49758</td>\n"," <td>16.7</td>\n"," <td>452649</td>\n"," <td>2761581</td>\n"," <td>141734334</td>\n"," <td>4063</td>\n"," <td>730</td>\n"," <td>2.417464e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Monroe County</th>\n"," <td>15.900000</td>\n"," <td>42.9</td>\n"," <td>9.33</td>\n"," <td>5.18</td>\n"," <td>63009</td>\n"," <td>11.8</td>\n"," <td>8963</td>\n"," <td>75027</td>\n"," <td>4097511</td>\n"," <td>158</td>\n"," <td>31</td>\n"," <td>1.061900e+08</td>\n"," </tr>\n"," <tr>\n"," <th>Nassau County</th>\n"," <td>14.566667</td>\n"," <td>42.9</td>\n"," <td>10.05</td>\n"," <td>9.01</td>\n"," <td>70590</td>\n"," <td>9.1</td>\n"," <td>7484</td>\n"," <td>85832</td>\n"," <td>1886261</td>\n"," <td>49</td>\n"," <td>16</td>\n"," <td>4.184525e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Okeechobee County</th>\n"," <td>14.053333</td>\n"," <td>40.0</td>\n"," <td>2.63</td>\n"," <td>1.36</td>\n"," <td>42524</td>\n"," <td>21.8</td>\n"," <td>8415</td>\n"," <td>41537</td>\n"," <td>1021477</td>\n"," <td>7</td>\n"," <td>2</td>\n"," <td>2.920370e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Orange County</th>\n"," <td>16.326667</td>\n"," <td>44.1</td>\n"," <td>10.36</td>\n"," <td>7.48</td>\n"," <td>54021</td>\n"," <td>15.3</td>\n"," <td>201528</td>\n"," <td>1380645</td>\n"," <td>89817807</td>\n"," <td>1794</td>\n"," <td>346</td>\n"," <td>5.216380e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Osceola County</th>\n"," <td>19.046667</td>\n"," <td>49.9</td>\n"," <td>13.43</td>\n"," <td>7.02</td>\n"," <td>49284</td>\n"," <td>14.0</td>\n"," <td>48892</td>\n"," <td>367990</td>\n"," <td>9207981</td>\n"," <td>205</td>\n"," <td>50</td>\n"," <td>1.244058e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Palm Beach County</th>\n"," <td>18.200000</td>\n"," <td>51.1</td>\n"," <td>3.96</td>\n"," <td>1.92</td>\n"," <td>60059</td>\n"," <td>11.8</td>\n"," <td>170868</td>\n"," <td>1485941</td>\n"," <td>76866505</td>\n"," <td>904</td>\n"," <td>167</td>\n"," <td>8.307278e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Pasco County</th>\n"," <td>14.313333</td>\n"," <td>33.0</td>\n"," <td>13.18</td>\n"," <td>5.97</td>\n"," <td>51247</td>\n"," <td>13.0</td>\n"," <td>67635</td>\n"," <td>539630</td>\n"," <td>9330553</td>\n"," <td>206</td>\n"," <td>58</td>\n"," <td>1.824896e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Pinellas County</th>\n"," <td>15.446667</td>\n"," <td>42.9</td>\n"," <td>5.85</td>\n"," <td>4.32</td>\n"," <td>51488</td>\n"," <td>12.2</td>\n"," <td>115990</td>\n"," <td>975280</td>\n"," <td>44125945</td>\n"," <td>730</td>\n"," <td>159</td>\n"," <td>5.632790e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Polk County</th>\n"," <td>17.544444</td>\n"," <td>36.9</td>\n"," <td>8.06</td>\n"," <td>6.21</td>\n"," <td>48328</td>\n"," <td>16.1</td>\n"," <td>107844</td>\n"," <td>708009</td>\n"," <td>20779632</td>\n"," <td>524</td>\n"," <td>116</td>\n"," <td>4.351088e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Putnam County</th>\n"," <td>16.980000</td>\n"," <td>45.1</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>34390</td>\n"," <td>26.3</td>\n"," <td>18954</td>\n"," <td>74163</td>\n"," <td>1925314</td>\n"," <td>16</td>\n"," <td>4</td>\n"," <td>1.110617e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Sarasota County</th>\n"," <td>15.620000</td>\n"," <td>45.1</td>\n"," <td>9.21</td>\n"," <td>7.06</td>\n"," <td>77022</td>\n"," <td>8.3</td>\n"," <td>20118</td>\n"," <td>254261</td>\n"," <td>7313073</td>\n"," <td>243</td>\n"," <td>68</td>\n"," <td>1.306549e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Seminole County</th>\n"," <td>16.740000</td>\n"," <td>49.9</td>\n"," <td>16.32</td>\n"," <td>9.43</td>\n"," <td>49995</td>\n"," <td>12.8</td>\n"," <td>39839</td>\n"," <td>321128</td>\n"," <td>7030683</td>\n"," <td>331</td>\n"," <td>81</td>\n"," <td>2.173058e+07</td>\n"," </tr>\n"," <tr>\n"," <th>St. Johns County</th>\n"," <td>16.980000</td>\n"," <td>42.0</td>\n"," <td>7.08</td>\n"," <td>5.71</td>\n"," <td>58423</td>\n"," <td>9.2</td>\n"," <td>38065</td>\n"," <td>426718</td>\n"," <td>15773229</td>\n"," <td>154</td>\n"," <td>34</td>\n"," <td>1.862228e+07</td>\n"," </tr>\n"," <tr>\n"," <th>St. Lucie County</th>\n"," <td>16.033333</td>\n"," <td>44.1</td>\n"," <td>11.28</td>\n"," <td>9.24</td>\n"," <td>63865</td>\n"," <td>11.2</td>\n"," <td>51321</td>\n"," <td>467832</td>\n"," <td>17902542</td>\n"," <td>112</td>\n"," <td>32</td>\n"," <td>2.350643e+07</td>\n"," </tr>\n"," <tr>\n"," <th>Sumter County</th>\n"," <td>13.760000</td>\n"," <td>36.9</td>\n"," <td>10.51</td>\n"," <td>8.05</td>\n"," <td>57931</td>\n"," <td>9.1</td>\n"," <td>10672</td>\n"," <td>128754</td>\n"," <td>2312425</td>\n"," <td>21</td>\n"," <td>7</td>\n"," <td>3.424526e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Suwannee County</th>\n"," <td>11.066667</td>\n"," <td>20.0</td>\n"," <td>4.39</td>\n"," <td>2.38</td>\n"," <td>44144</td>\n"," <td>20.3</td>\n"," <td>8299</td>\n"," <td>44191</td>\n"," <td>851332</td>\n"," <td>12</td>\n"," <td>4</td>\n"," <td>2.045323e+06</td>\n"," </tr>\n"," <tr>\n"," <th>Union County</th>\n"," <td>12.453333</td>\n"," <td>35.0</td>\n"," <td>18.25</td>\n"," <td>11.39</td>\n"," <td>47373</td>\n"," <td>22.2</td>\n"," <td>2291</td>\n"," <td>14940</td>\n"," <td>341763</td>\n"," <td>3</td>\n"," <td>3</td>\n"," <td>5.774985e+05</td>\n"," </tr>\n"," <tr>\n"," <th>Volusia County</th>\n"," <td>16.333333</td>\n"," <td>35.0</td>\n"," <td>8.66</td>\n"," <td>4.66</td>\n"," <td>46911</td>\n"," <td>15.2</td>\n"," <td>79877</td>\n"," <td>547538</td>\n"," <td>14864234</td>\n"," <td>266</td>\n"," <td>51</td>\n"," <td>3.782572e+07</td>\n"," </tr>\n"," </tbody>\n","</table>\n","</div>"],"text/plain":[" maxwind_mean maxwind_peak ... twt_peak dmg\n","county ... \n","Alachua County 12.453333 35.0 ... 39 9.306765e+06\n","Baker County 10.200000 24.1 ... 1 1.360526e+06\n","Bradford County 12.453333 35.0 ... 1 3.246248e+06\n","Brevard County 14.900000 41.0 ... 71 3.192891e+07\n","Broward County 15.680000 49.0 ... 382 1.329551e+08\n","Charlotte County 17.666667 44.1 ... 20 6.625988e+06\n","Citrus County 14.313333 33.0 ... 13 5.568339e+06\n","Clay County 10.200000 24.1 ... 8 1.219590e+07\n","Collier County 18.866667 63.9 ... 63 6.194133e+07\n","Columbia County 12.453333 35.0 ... 4 3.225476e+06\n","DeSoto County 17.666667 44.1 ... 2 4.049290e+06\n","Dixie County 13.393333 35.9 ... 5 9.079057e+05\n","Duval County 15.626667 42.9 ... 137 4.740938e+07\n","Flagler County 15.920000 44.1 ... 16 7.243310e+06\n","Gilchrist County 13.393333 35.9 ... 1 6.418368e+05\n","Glades County 14.053333 40.0 ... 2 1.684916e+06\n","Hamilton County 11.066667 20.0 ... 3 7.796505e+05\n","Hardee County 18.146667 54.0 ... 3 6.791781e+06\n","Hendry County 18.866667 63.9 ... 1 4.864095e+06\n","Hernando County 14.313333 33.0 ... 12 6.987439e+06\n","Highlands County 18.146667 54.0 ... 16 1.104181e+07\n","Hillsborough County 10.140000 27.0 ... 234 4.139342e+07\n","Indian River County 16.033333 49.9 ... 23 5.417251e+06\n","Lafayette County 11.066667 20.0 ... 1 4.387254e+05\n","Lake County 13.760000 36.9 ... 47 1.418739e+07\n","Lee County 15.206667 49.9 ... 63 6.269670e+07\n","Levy County 13.393333 35.9 ... 2 1.823098e+06\n","Manatee County 15.620000 42.0 ... 33 1.369299e+07\n","Marion County 11.150000 32.1 ... 30 1.934269e+07\n","Martin County 12.593333 22.0 ... 22 2.936391e+06\n","Miami-Dade County 15.900000 42.9 ... 730 2.417464e+08\n","Monroe County 15.900000 42.9 ... 31 1.061900e+08\n","Nassau County 14.566667 42.9 ... 16 4.184525e+06\n","Okeechobee County 14.053333 40.0 ... 2 2.920370e+06\n","Orange County 16.326667 44.1 ... 346 5.216380e+07\n","Osceola County 19.046667 49.9 ... 50 1.244058e+07\n","Palm Beach County 18.200000 51.1 ... 167 8.307278e+07\n","Pasco County 14.313333 33.0 ... 58 1.824896e+07\n","Pinellas County 15.446667 42.9 ... 159 5.632790e+07\n","Polk County 17.544444 36.9 ... 116 4.351088e+07\n","Putnam County 16.980000 45.1 ... 4 1.110617e+07\n","Sarasota County 15.620000 45.1 ... 68 1.306549e+07\n","Seminole County 16.740000 49.9 ... 81 2.173058e+07\n","St. Johns County 16.980000 42.0 ... 34 1.862228e+07\n","St. Lucie County 16.033333 44.1 ... 32 2.350643e+07\n","Sumter County 13.760000 36.9 ... 7 3.424526e+06\n","Suwannee County 11.066667 20.0 ... 4 2.045323e+06\n","Union County 12.453333 35.0 ... 3 5.774985e+05\n","Volusia County 16.333333 35.0 ... 51 3.782572e+07\n","\n","[49 rows x 12 columns]"]},"metadata":{"tags":[]},"execution_count":1}]},{"cell_type":"markdown","metadata":{"id":"Fg-Ticrhdpih"},"source":["## Feature Selection\n","\n","We have 11 features, but do we really need them all? Especially when we only have 49 data points. As the following heat map shows, a lot of the variables correlate with each other, forming distinct clusters."]},{"cell_type":"code","metadata":{"id":"B-5M2buOeaQV","executionInfo":{"status":"ok","timestamp":1603699822936,"user_tz":240,"elapsed":2122,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"50f1168f-a4f5-4274-e7e2-06b21a75dc74","colab":{"base_uri":"https://localhost:8080/","height":512}},"source":["plt.figure(figsize=(12,7))\n","sns.heatmap(df.corr(),\n"," annot=True,\n"," fmt = '.2f',\n"," cmap='coolwarm')\n","plt.show()"],"execution_count":2,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 864x504 with 2 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"SWRPvSoZQ341"},"source":["We can identify these clusters as follows:\n","\n","- Wind: `maxwind_mean`, `maxwind_peak`\n","- Precipitation: `precip_total`, `precip_peak`\n","- Socioeconomic: `mhi`, `poverty_rate`\n","- Population: `poverty_pop`, `population`, `gdp`, `twt_total`, `twt_peak`\n","\n","Our goal is to predict `dmg`. To avoid high multicollinearity among our predictors, let's simply take one variable from each cluster. We'll pick the ones with the strongest correlation to `dmg`. This gives us `maxwind_peak`, `precip_peak`, `mhi`, `twt_total`.\n","\n","Let's verify that these variables are independent of eacher using **Variance Inflation Factor (VIF)**. The VIF of an independent variable is a measure of how much its variation can be attributed to other independent variables. The higher it is, the more redundant the (not so) independent variable is. VIF is calculated as follows:\n","\n","$$ VIF = \\frac{1}{1 - R^2_i} $$\n","\n","where $R_i^2$ is the $R^2$ of the multilinear regression model of the $i^{\\text{th}}$ independent variable using the other independent variables as predictors. Why not just use $R^2_i$ to measure redundancy? One intuition is that the inverse relationship will harshly penalize smaller and smaller values of $1 - R^2_i$ (the so-called **tolerance**).\n","\n","Let's compute the VIF for each of the four predictors we're using:"]},{"cell_type":"code","metadata":{"id":"nnUg2sP6-DIk","executionInfo":{"status":"ok","timestamp":1603699822942,"user_tz":240,"elapsed":2105,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"08b73a92-ccb8-47ac-8066-2147602e4f91","colab":{"base_uri":"https://localhost:8080/","height":144}},"source":["from sklearn.preprocessing import robust_scale\n","from statsmodels.stats.outliers_influence import variance_inflation_factor as vif\n","import statsmodels.api as sm\n","\n","# preprocess data\n","features = ['maxwind_peak', 'precip_peak', 'mhi', 'twt_total']\n","df = df[features+['dmg']]\n","\n","X = df[features].apply(robust_scale)\n","X_vif = sm.add_constant(X.values) # for the vif func\n","\n","vif_dict = {}\n","for f in features:\n"," idx = features.index(f) # need this for vif func\n"," vif_dict[f] = vif(X_vif, idx)\n","vif_dict"],"execution_count":3,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n"," import pandas.util.testing as tm\n"],"name":"stderr"},{"output_type":"execute_result","data":{"text/plain":["{'maxwind_peak': 1.1609234512973168,\n"," 'mhi': 1.0514008951220526,\n"," 'precip_peak': 1.0447003684876797,\n"," 'twt_total': 1.097731476287398}"]},"metadata":{"tags":[]},"execution_count":3}]},{"cell_type":"markdown","metadata":{"id":"jFmHET7jZGLm"},"source":["This is good! The typical threshhold for VIF's is 10, sometimes 5. All the VIF's calculated were close to 1, indicating that the variables are very much independent of each other.\n","\n","## Modeling\n","Now that we have settled on our features, let's model our data. We will try out a linear regression and decision tree model and see which comes out best."]},{"cell_type":"code","metadata":{"id":"XHNL9NDLGKy6","executionInfo":{"status":"ok","timestamp":1603699822946,"user_tz":240,"elapsed":2092,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}}},"source":["from sklearn.model_selection import train_test_split\n","\n","Y = robust_scale(df['dmg'])\n","\n","X_train, X_test, Y_train, Y_test = train_test_split(X, Y,\n"," test_size = 0.1,\n"," random_state = 0,\n"," shuffle = True)"],"execution_count":4,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"SpztfD5hGLd1"},"source":["### Linear Regression\n"]},{"cell_type":"code","metadata":{"id":"0ONf_rgoZhOS","executionInfo":{"status":"ok","timestamp":1603699822949,"user_tz":240,"elapsed":2082,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}}},"source":["from sklearn.linear_model import LinearRegression\n","\n","linreg = LinearRegression()\n","\n","linreg_model = linreg.fit(X_train, Y_train)"],"execution_count":5,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"J4WYVukfMbh2"},"source":["### Decision Tree\n","Now let's find the optimal decision tree model."]},{"cell_type":"code","metadata":{"id":"sS8der1cMoYP","executionInfo":{"status":"ok","timestamp":1603699823902,"user_tz":240,"elapsed":3023,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"e0264778-4298-4af5-82ff-ff84db063a64","colab":{"base_uri":"https://localhost:8080/","height":35}},"source":["from sklearn.model_selection import GridSearchCV\n","from sklearn.tree import DecisionTreeRegressor\n","\n","tree = DecisionTreeRegressor(random_state = 0)\n","param_grid = {'min_impurity_decrease': [1e-2, 1e-1, 0.2, 0.3],\n"," 'max_features': [None, 'sqrt'],\n"," 'max_depth': [1, 2, 3, 4, 5, 6]}\n","\n","dt_model = GridSearchCV(tree, param_grid, scoring='r2')\n","dt_model.fit(X_train, Y_train)\n","dt_model.best_params_"],"execution_count":6,"outputs":[{"output_type":"execute_result","data":{"text/plain":["{'max_depth': 2, 'max_features': 'sqrt', 'min_impurity_decrease': 0.01}"]},"metadata":{"tags":[]},"execution_count":6}]},{"cell_type":"markdown","metadata":{"id":"g0Jhv1mSQ4gA"},"source":["### Model Selection\n","Let's see how the models stack up on their train/test results"]},{"cell_type":"code","metadata":{"id":"yCbkNzvnmocN","executionInfo":{"status":"ok","timestamp":1603699823910,"user_tz":240,"elapsed":3014,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"9d0b8df9-e1dd-4a1d-f992-e4082fc8e907","colab":{"base_uri":"https://localhost:8080/","height":141}},"source":["linreg_train_score = linreg_model.score(X_train, Y_train)\n","linreg_test_score = linreg_model.score(X_test, Y_test)\n","dt_train_score = dt_model.score(X_train, Y_train)\n","dt_test_score = dt_model.score(X_test, Y_test)\n","\n","print(\"===== Linear Regression r2 Scores =====\")\n","print(\"Train: %.2f\" % linreg_train_score)\n","print(\"Test: %.2f\" % linreg_test_score)\n","print()\n","print(\"===== Decision Tree r2 Scores =====\")\n","print(\"Train: %.2f\" % dt_train_score)\n","print(\"Test: %.2f\" % dt_test_score)"],"execution_count":7,"outputs":[{"output_type":"stream","text":["===== Linear Regression r2 Scores =====\n","Train: 0.56\n","Test: 0.79\n","\n","===== Decision Tree r2 Scores =====\n","Train: 0.73\n","Test: 0.18\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"nFw2immG5Opb"},"source":["The linear regression model seems to generalize well, the test score is *higher* than the train score. The decision tree model on the other hand, is severely overfit.\n","\n","Let's plot the predictions from each model against the true values of the dependent variable and see if this gives any insight."]},{"cell_type":"code","metadata":{"id":"nX7t6gdi8a7u","executionInfo":{"status":"ok","timestamp":1603699824139,"user_tz":240,"elapsed":3230,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"1948cbb2-a769-4bc2-b3ba-12cb1589ba9a","colab":{"base_uri":"https://localhost:8080/","height":589}},"source":["from sklearn.metrics import r2_score\n","\n","Y_linreg = linreg_model.predict(X)\n","Y_dt = dt_model.predict(X)\n","\n","\n","Y_arr = np.array([Y, Y_linreg, Y_dt]).T\n","Y_df = pd.DataFrame(data=Y_arr,\n"," columns = ['Y_true', 'Y_linreg', 'Y_dt'])\n","\n","sns.scatterplot(data = Y_df,\n"," x = 'Y_true',\n"," y = 'Y_linreg')\n","plt.plot(Y, Y, 'k-')\n","\n","plt.figure()\n","sns.scatterplot(data = Y_df,\n"," x = 'Y_true',\n"," y = 'Y_dt')\n","plt.plot(Y, Y, 'k-')\n","\n","print(\"Linear regression r2 = %.2f\" % r2_score(Y, Y_linreg))\n","print(\"Decision Tree r2 = %.2f\" % r2_score(Y, Y_dt))"],"execution_count":8,"outputs":[{"output_type":"stream","text":["Linear regression r2 = 0.76\n","Decision Tree r2 = 0.48\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"LRrQZPrrTY-j"},"source":["Both models appear to struggle to accurately predict large values of Y (i.e. `dmg`), especially the decision tree model.\n","\n","How good are the models for \"low\" damage? Let's cut off the Y_true past 2 and see how the models compare."]},{"cell_type":"code","metadata":{"id":"w7g4B2RA6Psm","executionInfo":{"status":"ok","timestamp":1603699824836,"user_tz":240,"elapsed":3914,"user":{"displayName":"Angel Umana","photoUrl":"","userId":"02017183028986324110"}},"outputId":"69890709-e70a-4366-9e53-528bd5fc5081","colab":{"base_uri":"https://localhost:8080/","height":589}},"source":["Y_low = Y[Y < 2]\n","X_low = X[Y < 2]\n","\n","Y_low_linreg = linreg_model.predict(X_low)\n","Y_low_dt = dt_model.predict(X_low)\n","\n","\n","Y_low_arr = np.array([Y_low, Y_low_linreg, Y_low_dt]).T\n","Y_low_df = pd.DataFrame(data=Y_low_arr,\n"," columns = ['Y_low_true', 'Y_low_linreg', 'Y_low_dt'])\n","\n","sns.scatterplot(data = Y_low_df,\n"," x = 'Y_low_true',\n"," y = 'Y_low_linreg')\n","plt.plot(Y_low, Y_low, 'k-')\n","\n","plt.figure()\n","sns.scatterplot(data = Y_low_df,\n"," x = 'Y_low_true',\n"," y = 'Y_low_dt')\n","plt.plot(Y_low, Y_low, 'k-')\n","\n","print(\"Linear regression r2 = %.2f\" % r2_score(Y_low, Y_low_linreg))\n","print(\"Decision Tree r2 = %.2f\" % r2_score(Y_low, Y_low_dt))"],"execution_count":9,"outputs":[{"output_type":"stream","text":["Linear regression r2 = 0.57\n","Decision Tree r2 = 0.70\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"t66hzINe7h7V"},"source":["Suddenly, the predictive capabilities of the two models flipped! This seems to solidify the idea that the linear regression model was able to adapt to outliers better than the decision tree model. At the same time, the decision tree model can generalize better to lower values of `dmg`. In a context such as ours (natural disaster assessment) it is crucial to account for outliers (counties with a lot of federal aid needed), thus the linear regression model is better suited for future predictions.\n","\n","## Final Remarks\n","\n","Where should we go from here?\n","\n","Ideas:\n","\n","- Investigate feature_importances_ from the decision tree model. While it may not be better suited for generalization, maybe we can learn something from how it weighed the predictors. Depending on what is found, we can try to modify the linear regression model.\n","- Incorporate data from hurricanes Harvey and Michael"]},{"cell_type":"code","metadata":{"id":"wpNUo0XL9wsZ"},"source":[""],"execution_count":null,"outputs":[]}]}
\ No newline at end of file
projects / irma-modeling / .git / commitdiff