{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "a20290f5-2a26-4622-9e8b-ce94cc1dcb1a"
    },
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Estimating nitrogen and phosphorus concentrations in streams and rivers  \n",
    "\n",
    "\n",
    "<center>Longzhu Shen</center> \n",
    "<p>\n",
    "<center><i>Spatial Ecology</i></center>\n",
    "<p>\n",
    "<center>May 2021</center>\n",
    "    \n",
    "Exercise base on the: \n",
    "\n",
    "**[Estimating nitrogen and phosphorus concentrations in streams and rivers, within a machine learning framework](https://www.nature.com/articles/s41597-020-0478-7)**\n",
    "\n",
    "Longzhu Q. Shen, Giuseppe Amatulli, Tushar Sethi, Peter Raymond & Sami Domisch  \n",
    "Scientific Data volume 7, Article number: 161 (2020) Cite this article "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "494d6527-1db0-4bca-a3a9-ea7d28eed532"
    },
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "**BackGround**\n",
    "- Geoenviornmental variables \n",
    "- Ground observationd : Nitrogen in US streams  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lectures: [Machine Learning Optimization](http://spatial-ecology.net/docs/source/lectures/lect_20210506_ML_OPT.pdf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "e675a434-eb22-4d12-a159-d6464ec9d916"
    },
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "**Code**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "nbpresent": {
     "id": "8f67df50-3050-47d2-a5d9-4329a61325fa"
    },
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.ensemble import RandomForestRegressor as RFReg\n",
    "from sklearn.model_selection import train_test_split,GridSearchCV\n",
    "from sklearn.pipeline import Pipeline\n",
    "from scipy.stats.stats import pearsonr\n",
    "import matplotlib.pyplot as plt\n",
    "plt.rcParams[\"figure.figsize\"] = (10,6.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpresent": {
     "id": "cf3cb8c6-45f0-48de-b171-598ebf92904f"
    },
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "read in data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sLong</th>\n",
       "      <th>sLat</th>\n",
       "      <th>mean</th>\n",
       "      <th>lu_avg_01</th>\n",
       "      <th>lu_avg_02</th>\n",
       "      <th>lu_avg_03</th>\n",
       "      <th>lu_avg_04</th>\n",
       "      <th>lu_avg_05</th>\n",
       "      <th>lu_avg_06</th>\n",
       "      <th>lu_avg_07</th>\n",
       "      <th>...</th>\n",
       "      <th>hydro_avg_13</th>\n",
       "      <th>hydro_avg_14</th>\n",
       "      <th>hydro_avg_15</th>\n",
       "      <th>hydro_avg_16</th>\n",
       "      <th>hydro_avg_17</th>\n",
       "      <th>hydro_avg_18</th>\n",
       "      <th>hydro_avg_19</th>\n",
       "      <th>dem_avg</th>\n",
       "      <th>slope_ave</th>\n",
       "      <th>lentic_lotic01</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-123.187500</td>\n",
       "      <td>46.179165</td>\n",
       "      <td>0.566667</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>45</td>\n",
       "      <td>23</td>\n",
       "      <td>22</td>\n",
       "      <td>12</td>\n",
       "      <td>...</td>\n",
       "      <td>107329616</td>\n",
       "      <td>38819396</td>\n",
       "      <td>29</td>\n",
       "      <td>310689856</td>\n",
       "      <td>136347292</td>\n",
       "      <td>310689856</td>\n",
       "      <td>136347292</td>\n",
       "      <td>1470</td>\n",
       "      <td>492</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-123.129166</td>\n",
       "      <td>45.437500</td>\n",
       "      <td>1.166667</td>\n",
       "      <td>24</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>25</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>41</td>\n",
       "      <td>...</td>\n",
       "      <td>23143</td>\n",
       "      <td>1529</td>\n",
       "      <td>68</td>\n",
       "      <td>64073</td>\n",
       "      <td>7800</td>\n",
       "      <td>64073</td>\n",
       "      <td>7800</td>\n",
       "      <td>142</td>\n",
       "      <td>291</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-123.120834</td>\n",
       "      <td>45.470833</td>\n",
       "      <td>0.628778</td>\n",
       "      <td>46</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>15</td>\n",
       "      <td>...</td>\n",
       "      <td>167329</td>\n",
       "      <td>11217</td>\n",
       "      <td>68</td>\n",
       "      <td>470310</td>\n",
       "      <td>55128</td>\n",
       "      <td>470310</td>\n",
       "      <td>55128</td>\n",
       "      <td>315</td>\n",
       "      <td>353</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-123.120834</td>\n",
       "      <td>45.504166</td>\n",
       "      <td>0.336667</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>32</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>117497</td>\n",
       "      <td>8462</td>\n",
       "      <td>67</td>\n",
       "      <td>333116</td>\n",
       "      <td>40989</td>\n",
       "      <td>333116</td>\n",
       "      <td>40989</td>\n",
       "      <td>315</td>\n",
       "      <td>441</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-123.054169</td>\n",
       "      <td>45.504166</td>\n",
       "      <td>0.595833</td>\n",
       "      <td>46</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>17</td>\n",
       "      <td>...</td>\n",
       "      <td>307267</td>\n",
       "      <td>21228</td>\n",
       "      <td>68</td>\n",
       "      <td>865826</td>\n",
       "      <td>104122</td>\n",
       "      <td>865826</td>\n",
       "      <td>104122</td>\n",
       "      <td>290</td>\n",
       "      <td>363</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-123.012497</td>\n",
       "      <td>45.520832</td>\n",
       "      <td>1.573333</td>\n",
       "      <td>30</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>32</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>29</td>\n",
       "      <td>...</td>\n",
       "      <td>233499</td>\n",
       "      <td>18441</td>\n",
       "      <td>65</td>\n",
       "      <td>662045</td>\n",
       "      <td>92675</td>\n",
       "      <td>662045</td>\n",
       "      <td>92675</td>\n",
       "      <td>208</td>\n",
       "      <td>248</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        sLong       sLat      mean  lu_avg_01  lu_avg_02  lu_avg_03  \\\n",
       "0 -123.187500  46.179165  0.566667         28          0         15   \n",
       "1 -123.129166  45.437500  1.166667         24          0          6   \n",
       "2 -123.120834  45.470833  0.628778         46          0          3   \n",
       "3 -123.120834  45.504166  0.336667         55          0          2   \n",
       "4 -123.054169  45.504166  0.595833         46          0          3   \n",
       "5 -123.012497  45.520832  1.573333         30          0          4   \n",
       "\n",
       "   lu_avg_04  lu_avg_05  lu_avg_06  lu_avg_07  ...  hydro_avg_13  \\\n",
       "0         45         23         22         12  ...     107329616   \n",
       "1         25          0          4         41  ...         23143   \n",
       "2         33          0          1         15  ...        167329   \n",
       "3         32          0          1          9  ...        117497   \n",
       "4         31          0          2         17  ...        307267   \n",
       "5         32          0          3         29  ...        233499   \n",
       "\n",
       "   hydro_avg_14  hydro_avg_15  hydro_avg_16  hydro_avg_17  hydro_avg_18  \\\n",
       "0      38819396            29     310689856     136347292     310689856   \n",
       "1          1529            68         64073          7800         64073   \n",
       "2         11217            68        470310         55128        470310   \n",
       "3          8462            67        333116         40989        333116   \n",
       "4         21228            68        865826        104122        865826   \n",
       "5         18441            65        662045         92675        662045   \n",
       "\n",
       "   hydro_avg_19  dem_avg  slope_ave  lentic_lotic01  \n",
       "0     136347292     1470        492               8  \n",
       "1          7800      142        291               2  \n",
       "2         55128      315        353               3  \n",
       "3         40989      315        441               2  \n",
       "4        104122      290        363               3  \n",
       "5         92675      208        248               3  \n",
       "\n",
       "[6 rows x 50 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dsIn = pd.read_csv(\"./txt/TN_Data_Sample.csv\")\n",
    "dsIn.head(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['sLong', 'sLat', 'mean', 'lu_avg_01', 'lu_avg_02', 'lu_avg_03',\n",
       "       'lu_avg_04', 'lu_avg_05', 'lu_avg_06', 'lu_avg_07', 'lu_avg_08',\n",
       "       'lu_avg_09', 'lu_avg_10', 'lu_avg_11', 'lu_avg_12', 'prec', 'tmin',\n",
       "       'tmax', 'soil_avg_01', 'soil_avg_02', 'soil_avg_03', 'soil_avg_04',\n",
       "       'soil_avg_05', 'soil_avg_06', 'soil_avg_07', 'soil_avg_08',\n",
       "       'soil_avg_09', 'soil_avg_10', 'hydro_avg_01', 'hydro_avg_02',\n",
       "       'hydro_avg_03', 'hydro_avg_04', 'hydro_avg_05', 'hydro_avg_06',\n",
       "       'hydro_avg_07', 'hydro_avg_08', 'hydro_avg_09', 'hydro_avg_10',\n",
       "       'hydro_avg_11', 'hydro_avg_12', 'hydro_avg_13', 'hydro_avg_14',\n",
       "       'hydro_avg_15', 'hydro_avg_16', 'hydro_avg_17', 'hydro_avg_18',\n",
       "       'hydro_avg_19', 'dem_avg', 'slope_ave', 'lentic_lotic01'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dsIn.columns.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "feats = ['mean','tmax']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x468 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(min(dsIn['mean']),max(dsIn['mean']),100)\n",
    "plt.hist((dsIn['mean']),bins,alpha=0.8);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "dsIn['log']=  np.log10(dsIn['mean'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x468 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins = np.linspace(min(dsIn['log']),max(dsIn['log']),100)\n",
    "plt.hist((dsIn['log']),bins,alpha=0.8);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = dsIn.iloc[:,3:50].values\n",
    "Y = dsIn.iloc[:,50:51].values\n",
    "feat = dsIn.iloc[:,3:50].columns.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['lu_avg_01', 'lu_avg_02', 'lu_avg_03', 'lu_avg_04', 'lu_avg_05',\n",
       "       'lu_avg_06', 'lu_avg_07', 'lu_avg_08', 'lu_avg_09', 'lu_avg_10',\n",
       "       'lu_avg_11', 'lu_avg_12', 'prec', 'tmin', 'tmax', 'soil_avg_01',\n",
       "       'soil_avg_02', 'soil_avg_03', 'soil_avg_04', 'soil_avg_05',\n",
       "       'soil_avg_06', 'soil_avg_07', 'soil_avg_08', 'soil_avg_09',\n",
       "       'soil_avg_10', 'hydro_avg_01', 'hydro_avg_02', 'hydro_avg_03',\n",
       "       'hydro_avg_04', 'hydro_avg_05', 'hydro_avg_06', 'hydro_avg_07',\n",
       "       'hydro_avg_08', 'hydro_avg_09', 'hydro_avg_10', 'hydro_avg_11',\n",
       "       'hydro_avg_12', 'hydro_avg_13', 'hydro_avg_14', 'hydro_avg_15',\n",
       "       'hydro_avg_16', 'hydro_avg_17', 'hydro_avg_18', 'hydro_avg_19',\n",
       "       'dem_avg', 'slope_ave', 'lentic_lotic01'], dtype=object)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1010, 47)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1010, 1)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y.shape"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "model building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.5, random_state=24)\n",
    "y_train = np.ravel(Y_train)\n",
    "y_test = np.ravel(Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 81 candidates, totalling 243 fits\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, estimator=Pipeline(steps=[('rf', RandomForestRegressor())]),\n",
       "             n_jobs=-1,\n",
       "             param_grid={'rf__max_depth': (50, 100, 200),\n",
       "                         'rf__max_features': ('log2', 'sqrt', 0.33),\n",
       "                         'rf__max_samples': (0.5, 0.6, 0.7),\n",
       "                         'rf__n_estimators': (500, 1000, 2000)},\n",
       "             scoring='r2', verbose=1)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipeline = Pipeline([('rf',RFReg())])\n",
    "\n",
    "parameters = {\n",
    "        'rf__max_features':(\"log2\",\"sqrt\",0.33),\n",
    "        'rf__max_samples':(0.5,0.6,0.7),\n",
    "        'rf__n_estimators':(500,1000,2000),\n",
    "        'rf__max_depth':(50,100,200)}\n",
    "\n",
    "grid_search = GridSearchCV(pipeline,parameters,n_jobs=-1,cv=3,scoring='r2',verbose=1)\n",
    "grid_search.fit(X_train,y_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6719812711835207"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Training score: 0.672\n",
      "Optimal parameters:\n",
      "\trf__max_depth: 200\n",
      "\trf__max_features: 0.33\n",
      "\trf__max_samples: 0.7\n",
      "\trf__n_estimators: 500\n"
     ]
    }
   ],
   "source": [
    "print ('Best Training score: %0.3f' % grid_search.best_score_)\n",
    "print ('Optimal parameters:')\n",
    "best_par = grid_search.best_estimator_.get_params()\n",
    "for par_name in sorted(parameters.keys()):\n",
    "    print ('\\t%s: %r' % (par_name, best_par[par_name]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9697515742517621, 0.8061496236451665]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfReg = RFReg(n_estimators=500,max_features=0.33,max_depth=200,max_samples=0.7,n_jobs=-1,random_state=24)\n",
    "rfReg.fit(X_train, y_train);\n",
    "dic_pred = {}\n",
    "dic_pred['train'] = rfReg.predict(X_train)\n",
    "dic_pred['test'] = rfReg.predict(X_test)\n",
    "[pearsonr(dic_pred['train'],y_train)[0],pearsonr(dic_pred['test'],y_test)[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x153076b2fdf0>]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x468 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y_train,dic_pred['train'])\n",
    "plt.xlabel('orig')\n",
    "plt.ylabel('pred')\n",
    "ident = [-1, 1.2]\n",
    "plt.plot(ident,ident,'r--')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x153075404250>]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x468 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y_test,dic_pred['test'])\n",
    "plt.xlabel('orig')\n",
    "plt.ylabel('pred')\n",
    "ident = [-1, 1.2]\n",
    "plt.plot(ident,ident,'r--')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "impt = [rfReg.feature_importances_, np.std([tree.feature_importances_ for tree in rfReg.estimators_],axis=1)] \n",
    "ind = np.argsort(impt[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 9,  7, 10,  1, 23,  4, 46, 12, 37, 40, 36, 42, 41, 38, 43, 20, 24,\n",
       "       27, 33, 35, 39,  8, 29, 44, 30, 22, 31, 26, 32,  0, 28,  2, 45, 18,\n",
       "       13, 14, 11, 25, 34,  5,  3,  6, 17, 21, 19, 15, 16])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = (6,12)\n",
    "plt.barh(range(len(feat)),impt[0][ind],color=\"b\", xerr=impt[1][ind], align=\"center\")\n",
    "plt.yticks(range(len(feat)),feat[ind]);"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "celltoolbar": "Slideshow",
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}