code/model-features.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "sns.set_palette([\"#999999\", \"#E69F00\", \"#56B4E9\", \"#009E73\", \"#F0E442\", \"#0072B2\", \"#D55E00\", \"#CC79A7\"])\n",
    "ncolours = len(plt.rcParams['axes.prop_cycle'])\n",
    "colours = [list(plt.rcParams['axes.prop_cycle'])[i]['color'] for i in range(ncolours)]\n",
    "plt.rcParams['xtick.labelsize'] = 15\n",
    "plt.rcParams['ytick.labelsize'] = 15\n",
    "plt.rcParams['axes.labelsize'] = 22\n",
    "plt.rcParams['font.size'] = 24\n",
    "plt.rcParams['font.family'] = 'serif'\n",
    "plt.rcParams['text.usetex'] = True\n",
    "\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_int = pd.read_csv('../data/data_int.csv', index_col=0)\n",
    "data_auc = pd.read_csv('../data/data_auc.csv', index_col=0)\n",
    "data_zauc = pd.read_csv('../data/data_zauc.csv', index_col=0)\n",
    "\n",
    "metadata = pd.read_csv('../data/metadata_clean.csv', index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>44826</th>\n",
       "      <th>44827</th>\n",
       "      <th>44828</th>\n",
       "      <th>44829</th>\n",
       "      <th>44830</th>\n",
       "      <th>44831</th>\n",
       "      <th>44832</th>\n",
       "      <th>44833</th>\n",
       "      <th>44834</th>\n",
       "      <th>44835</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19777</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19778</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>933.0</td>\n",
       "      <td>907.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3784.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19779</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>622.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>92.0</td>\n",
       "      <td>2590.0</td>\n",
       "      <td>2349.0</td>\n",
       "      <td>2326.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>1785.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19780</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3596.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>2018.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19781</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>775.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 44836 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         0    1    2     3      4    5    6    7    8    9   ...    44826  \\\n",
       "19777  0.0  0.0  0.0   0.0    0.0  0.0  0.0  0.0  0.0  0.0   ...      0.0   \n",
       "19778  0.0  0.0  0.0   0.0    0.0  0.0  0.0  0.0  0.0  0.0   ...    933.0   \n",
       "19779  0.0  0.0  0.0   0.0  622.0  0.0  0.0  0.0  0.0  0.0   ...      0.0   \n",
       "19780  0.0  0.0  0.0   0.0    0.0  0.0  0.0  0.0  0.0  0.0   ...      0.0   \n",
       "19781  0.0  0.0  0.0  33.0  775.0  0.0  0.0  0.0  0.0  0.0   ...      0.0   \n",
       "\n",
       "       44827   44828   44829   44830  44831  44832  44833  44834   44835  \n",
       "19777    0.0     0.0     0.0     0.0    0.0    0.0    4.0    0.0     0.0  \n",
       "19778  907.0     0.0     0.0  3784.0    0.0    0.0  101.0   81.0     0.0  \n",
       "19779   92.0  2590.0  2349.0  2326.0    0.0    0.0   64.0   36.0  1785.0  \n",
       "19780    0.0     0.0  3596.0     0.0    0.0    0.0    9.0   35.0  2018.0  \n",
       "19781    0.0     0.0     0.0    46.0    0.0    0.0    1.0    0.0     0.0  \n",
       "\n",
       "[5 rows x 44836 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_int.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "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>strain</th>\n",
       "      <th>genus</th>\n",
       "      <th>species</th>\n",
       "      <th>elevation</th>\n",
       "      <th>province</th>\n",
       "      <th>pH</th>\n",
       "      <th>temperature</th>\n",
       "      <th>moisture</th>\n",
       "      <th>soiltype</th>\n",
       "      <th>media</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19777</th>\n",
       "      <td>409</td>\n",
       "      <td>Xenorhabdus</td>\n",
       "      <td>stockiae</td>\n",
       "      <td>186</td>\n",
       "      <td>phitsanulok</td>\n",
       "      <td>6.8</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>loam</td>\n",
       "      <td>LB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19789</th>\n",
       "      <td>409</td>\n",
       "      <td>Xenorhabdus</td>\n",
       "      <td>stockiae</td>\n",
       "      <td>186</td>\n",
       "      <td>phitsanulok</td>\n",
       "      <td>6.8</td>\n",
       "      <td>28</td>\n",
       "      <td>1.0</td>\n",
       "      <td>loam</td>\n",
       "      <td>SF900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19817</th>\n",
       "      <td>432</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>243</td>\n",
       "      <td>phitsanulok</td>\n",
       "      <td>6.6</td>\n",
       "      <td>26</td>\n",
       "      <td>1.5</td>\n",
       "      <td>sandy loam</td>\n",
       "      <td>LB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19818</th>\n",
       "      <td>433</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>243</td>\n",
       "      <td>phitsanulok</td>\n",
       "      <td>6.8</td>\n",
       "      <td>27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>sandy loam</td>\n",
       "      <td>LB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19819</th>\n",
       "      <td>434</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>255</td>\n",
       "      <td>phitsanulok</td>\n",
       "      <td>6.8</td>\n",
       "      <td>27</td>\n",
       "      <td>1.0</td>\n",
       "      <td>sandy loam</td>\n",
       "      <td>LB</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       strain         genus      species  elevation     province   pH  \\\n",
       "19777     409   Xenorhabdus     stockiae        186  phitsanulok  6.8   \n",
       "19789     409   Xenorhabdus     stockiae        186  phitsanulok  6.8   \n",
       "19817     432  Photorhabdus  luminescens        243  phitsanulok  6.6   \n",
       "19818     433  Photorhabdus  luminescens        243  phitsanulok  6.8   \n",
       "19819     434  Photorhabdus  luminescens        255  phitsanulok  6.8   \n",
       "\n",
       "       temperature  moisture    soiltype  media  \n",
       "19777           28       1.0        loam     LB  \n",
       "19789           28       1.0        loam  SF900  \n",
       "19817           26       1.5  sandy loam     LB  \n",
       "19818           27       1.0  sandy loam     LB  \n",
       "19819           27       1.0  sandy loam     LB  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualising the distribution of genera w/r/t metadata"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's first explore the distribution of numerical metadata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "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>pH</th>\n",
       "      <th>temperature</th>\n",
       "      <th>moisture</th>\n",
       "      <th>elevation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>228.000000</td>\n",
       "      <td>228.000000</td>\n",
       "      <td>228.000000</td>\n",
       "      <td>228.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>6.728070</td>\n",
       "      <td>26.929825</td>\n",
       "      <td>1.635965</td>\n",
       "      <td>371.938596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.333057</td>\n",
       "      <td>3.272417</td>\n",
       "      <td>1.178525</td>\n",
       "      <td>385.358196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>5.200000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>6.600000</td>\n",
       "      <td>26.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>114.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>6.800000</td>\n",
       "      <td>28.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>226.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>7.000000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>348.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>7.000000</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>1291.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               pH  temperature    moisture    elevation\n",
       "count  228.000000   228.000000  228.000000   228.000000\n",
       "mean     6.728070    26.929825    1.635965   371.938596\n",
       "std      0.333057     3.272417    1.178525   385.358196\n",
       "min      5.200000    18.000000    0.500000     0.000000\n",
       "25%      6.600000    26.000000    1.000000   114.000000\n",
       "50%      6.800000    28.000000    1.000000   226.000000\n",
       "75%      7.000000    29.000000    2.000000   348.000000\n",
       "max      7.000000    32.000000    8.000000  1291.000000"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata[['pH', 'temperature', 'moisture', 'elevation']].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model we will be using is based on decision trees and therefore it's robust to different feature scalings and to outliers. For ease of visualisation, however, we rescale the numerical data so that pH, temperature and moisture lie between 0 and 1, and the elevation&mdash;which originally spans three orders of magnitude&mdash;is log-transformed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "metadata[['pH', 'temperature', 'moisture']] = scaler.fit_transform(metadata[['pH', 'temperature', 'moisture']])\n",
    "\n",
    "metadata.elevation = metadata.elevation.apply(np.log1p)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we plot the distribution of the genera, and see that it appears to be more or less evenly distributed with respect to these variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19de095d710>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 817.381x720 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.pairplot(metadata, vars=['pH', 'temperature', 'moisture', 'elevation'], hue='genus', palette=colours[5::-4], kind='scatter',\n",
    "             diag_kind='hist', diag_kws=dict(alpha=0.8, bins=15), plot_kws=dict(alpha=0.8, s=80))\n",
    "ax._legend.remove()\n",
    "plt.legend(fontsize=16, bbox_to_anchor=(1.04,1), loc=\"upper left\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the rest of the geographical metadata, the situation is similar."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19de03fef28>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "sns.countplot(y='province', data=metadata, hue='genus', palette=colours[5::-4])\n",
    "plt.legend(fontsize=18)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19de0cde860>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "sns.countplot(y='media', data=metadata, hue='genus', palette=colours[5::-4])\n",
    "plt.xlim(0,100)\n",
    "plt.legend(fontsize=18)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0,0.5,'soil type')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "sns.countplot(y='soiltype', data=metadata, hue='genus', palette=colours[5::-4])\n",
    "plt.legend(fontsize=18)\n",
    "plt.ylabel('soil type')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As a last step, we need to turn these categorical features into numerical ones. Since there is no specific ordering among the feature values of any of these columns, we apply one-hot encoding."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "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>strain</th>\n",
       "      <th>genus</th>\n",
       "      <th>species</th>\n",
       "      <th>elevation</th>\n",
       "      <th>pH</th>\n",
       "      <th>temperature</th>\n",
       "      <th>moisture</th>\n",
       "      <th>media_LB</th>\n",
       "      <th>media_SF900</th>\n",
       "      <th>province_chaiyaphum</th>\n",
       "      <th>...</th>\n",
       "      <th>province_mahasarakham</th>\n",
       "      <th>province_nakhonratchasima</th>\n",
       "      <th>province_phetchabuns</th>\n",
       "      <th>province_phitsanulok</th>\n",
       "      <th>province_saraburi</th>\n",
       "      <th>province_udonthani</th>\n",
       "      <th>soiltype_clay loam</th>\n",
       "      <th>soiltype_loam</th>\n",
       "      <th>soiltype_sand</th>\n",
       "      <th>soiltype_sandy loam</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19777</th>\n",
       "      <td>409</td>\n",
       "      <td>Xenorhabdus</td>\n",
       "      <td>stockiae</td>\n",
       "      <td>5.231109</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.714286</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19789</th>\n",
       "      <td>409</td>\n",
       "      <td>Xenorhabdus</td>\n",
       "      <td>stockiae</td>\n",
       "      <td>5.231109</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.714286</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19817</th>\n",
       "      <td>432</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>5.497168</td>\n",
       "      <td>0.777778</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0.133333</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19818</th>\n",
       "      <td>433</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>5.497168</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.642857</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19819</th>\n",
       "      <td>434</td>\n",
       "      <td>Photorhabdus</td>\n",
       "      <td>luminescens</td>\n",
       "      <td>5.545177</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.642857</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       strain         genus      species  elevation        pH  temperature  \\\n",
       "19777     409   Xenorhabdus     stockiae   5.231109  0.888889     0.714286   \n",
       "19789     409   Xenorhabdus     stockiae   5.231109  0.888889     0.714286   \n",
       "19817     432  Photorhabdus  luminescens   5.497168  0.777778     0.571429   \n",
       "19818     433  Photorhabdus  luminescens   5.497168  0.888889     0.642857   \n",
       "19819     434  Photorhabdus  luminescens   5.545177  0.888889     0.642857   \n",
       "\n",
       "       moisture  media_LB  media_SF900  province_chaiyaphum  \\\n",
       "19777  0.066667         1            0                    0   \n",
       "19789  0.066667         0            1                    0   \n",
       "19817  0.133333         1            0                    0   \n",
       "19818  0.066667         1            0                    0   \n",
       "19819  0.066667         1            0                    0   \n",
       "\n",
       "              ...           province_mahasarakham  province_nakhonratchasima  \\\n",
       "19777         ...                               0                          0   \n",
       "19789         ...                               0                          0   \n",
       "19817         ...                               0                          0   \n",
       "19818         ...                               0                          0   \n",
       "19819         ...                               0                          0   \n",
       "\n",
       "       province_phetchabuns  province_phitsanulok  province_saraburi  \\\n",
       "19777                     0                     1                  0   \n",
       "19789                     0                     1                  0   \n",
       "19817                     0                     1                  0   \n",
       "19818                     0                     1                  0   \n",
       "19819                     0                     1                  0   \n",
       "\n",
       "       province_udonthani  soiltype_clay loam  soiltype_loam  soiltype_sand  \\\n",
       "19777                   0                   0              1              0   \n",
       "19789                   0                   0              1              0   \n",
       "19817                   0                   0              0              0   \n",
       "19818                   0                   0              0              0   \n",
       "19819                   0                   0              0              0   \n",
       "\n",
       "       soiltype_sandy loam  \n",
       "19777                    0  \n",
       "19789                    0  \n",
       "19817                    1  \n",
       "19818                    1  \n",
       "19819                    1  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metadata = pd.get_dummies(metadata, columns=['media', 'province', 'soiltype'])\n",
    "metadata.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We save this for future use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "metadata.to_csv('../data/model/metadata.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Metabolite data\n",
    "\n",
    "### Rescaling and removal of poor predictors\n",
    "\n",
    "We start by peeking at a few columns of the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>48.076923</td>\n",
       "      <td>196.560440</td>\n",
       "      <td>18.129121</td>\n",
       "      <td>44.395604</td>\n",
       "      <td>235.387363</td>\n",
       "      <td>180.442308</td>\n",
       "      <td>28.711538</td>\n",
       "      <td>382.513736</td>\n",
       "      <td>1221.489011</td>\n",
       "      <td>272.335165</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>896.676271</td>\n",
       "      <td>998.021017</td>\n",
       "      <td>345.881582</td>\n",
       "      <td>644.709812</td>\n",
       "      <td>414.779321</td>\n",
       "      <td>2760.100777</td>\n",
       "      <td>521.092904</td>\n",
       "      <td>1005.280248</td>\n",
       "      <td>3892.201639</td>\n",
       "      <td>1389.553042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>93.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>377.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>29.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>17104.000000</td>\n",
       "      <td>9861.000000</td>\n",
       "      <td>6599.000000</td>\n",
       "      <td>12299.000000</td>\n",
       "      <td>5638.000000</td>\n",
       "      <td>51257.000000</td>\n",
       "      <td>9942.000000</td>\n",
       "      <td>5773.000000</td>\n",
       "      <td>30253.000000</td>\n",
       "      <td>10452.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  0            1            2             3            4  \\\n",
       "count    364.000000   364.000000   364.000000    364.000000   364.000000   \n",
       "mean      48.076923   196.560440    18.129121     44.395604   235.387363   \n",
       "std      896.676271   998.021017   345.881582    644.709812   414.779321   \n",
       "min        0.000000     0.000000     0.000000      0.000000     0.000000   \n",
       "25%        0.000000     0.000000     0.000000      0.000000     0.000000   \n",
       "50%        0.000000     0.000000     0.000000      0.000000    93.000000   \n",
       "75%        0.000000     0.000000     0.000000      0.000000   377.250000   \n",
       "max    17104.000000  9861.000000  6599.000000  12299.000000  5638.000000   \n",
       "\n",
       "                  5            6            7             8             9  \n",
       "count    364.000000   364.000000   364.000000    364.000000    364.000000  \n",
       "mean     180.442308    28.711538   382.513736   1221.489011    272.335165  \n",
       "std     2760.100777   521.092904  1005.280248   3892.201639   1389.553042  \n",
       "min        0.000000     0.000000     0.000000      0.000000      0.000000  \n",
       "25%        0.000000     0.000000     0.000000      0.000000      0.000000  \n",
       "50%        0.000000     0.000000     0.000000      0.000000      0.000000  \n",
       "75%        0.000000     0.000000     0.000000     31.000000     29.000000  \n",
       "max    51257.000000  9942.000000  5773.000000  30253.000000  10452.000000  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_int[data_int.columns[:10]].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>140.554011</td>\n",
       "      <td>706.054003</td>\n",
       "      <td>59.092418</td>\n",
       "      <td>150.725187</td>\n",
       "      <td>670.925295</td>\n",
       "      <td>521.973140</td>\n",
       "      <td>95.691795</td>\n",
       "      <td>116.704959</td>\n",
       "      <td>1456.295861</td>\n",
       "      <td>665.876896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2609.647928</td>\n",
       "      <td>3497.228104</td>\n",
       "      <td>1127.411473</td>\n",
       "      <td>2472.620855</td>\n",
       "      <td>1464.161092</td>\n",
       "      <td>7735.987197</td>\n",
       "      <td>1752.560518</td>\n",
       "      <td>307.936574</td>\n",
       "      <td>4598.554307</td>\n",
       "      <td>3548.824653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>62.445250</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>836.579500</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.105500</td>\n",
       "      <td>42.487250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>49773.375000</td>\n",
       "      <td>32889.396500</td>\n",
       "      <td>21509.640000</td>\n",
       "      <td>47184.761000</td>\n",
       "      <td>20737.809000</td>\n",
       "      <td>142852.257000</td>\n",
       "      <td>33438.115000</td>\n",
       "      <td>1610.047000</td>\n",
       "      <td>34814.323000</td>\n",
       "      <td>30485.268000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  0             1             2             3             4  \\\n",
       "count    364.000000    364.000000    364.000000    364.000000    364.000000   \n",
       "mean     140.554011    706.054003     59.092418    150.725187    670.925295   \n",
       "std     2609.647928   3497.228104   1127.411473   2472.620855   1464.161092   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000     62.445250   \n",
       "75%        0.000000      0.000000      0.000000      0.000000    836.579500   \n",
       "max    49773.375000  32889.396500  21509.640000  47184.761000  20737.809000   \n",
       "\n",
       "                   5             6            7             8             9  \n",
       "count     364.000000    364.000000   364.000000    364.000000    364.000000  \n",
       "mean      521.973140     95.691795   116.704959   1456.295861    665.876896  \n",
       "std      7735.987197   1752.560518   307.936574   4598.554307   3548.824653  \n",
       "min         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "25%         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "50%         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "75%         0.000000      0.000000     0.000000     11.105500     42.487250  \n",
       "max    142852.257000  33438.115000  1610.047000  34814.323000  30485.268000  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_auc[data_auc.columns[:10]].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "      <td>364.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>140.554011</td>\n",
       "      <td>706.054003</td>\n",
       "      <td>59.092418</td>\n",
       "      <td>150.520782</td>\n",
       "      <td>670.925295</td>\n",
       "      <td>521.973140</td>\n",
       "      <td>95.231952</td>\n",
       "      <td>116.510342</td>\n",
       "      <td>1421.750907</td>\n",
       "      <td>665.173688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2609.647928</td>\n",
       "      <td>3497.228104</td>\n",
       "      <td>1127.411473</td>\n",
       "      <td>2472.630257</td>\n",
       "      <td>1464.161092</td>\n",
       "      <td>7735.987197</td>\n",
       "      <td>1752.581532</td>\n",
       "      <td>308.002694</td>\n",
       "      <td>4593.569615</td>\n",
       "      <td>3548.953034</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>62.445250</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>836.579500</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.141250</td>\n",
       "      <td>35.909750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>49773.375000</td>\n",
       "      <td>32889.396500</td>\n",
       "      <td>21509.640000</td>\n",
       "      <td>47184.761000</td>\n",
       "      <td>20737.809000</td>\n",
       "      <td>142852.257000</td>\n",
       "      <td>33438.115000</td>\n",
       "      <td>1610.047000</td>\n",
       "      <td>34814.323000</td>\n",
       "      <td>30485.268000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  0             1             2             3             4  \\\n",
       "count    364.000000    364.000000    364.000000    364.000000    364.000000   \n",
       "mean     140.554011    706.054003     59.092418    150.520782    670.925295   \n",
       "std     2609.647928   3497.228104   1127.411473   2472.630257   1464.161092   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000     62.445250   \n",
       "75%        0.000000      0.000000      0.000000      0.000000    836.579500   \n",
       "max    49773.375000  32889.396500  21509.640000  47184.761000  20737.809000   \n",
       "\n",
       "                   5             6            7             8             9  \n",
       "count     364.000000    364.000000   364.000000    364.000000    364.000000  \n",
       "mean      521.973140     95.231952   116.510342   1421.750907    665.173688  \n",
       "std      7735.987197   1752.581532   308.002694   4593.569615   3548.953034  \n",
       "min         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "25%         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "50%         0.000000      0.000000     0.000000      0.000000      0.000000  \n",
       "75%         0.000000      0.000000     0.000000      8.141250     35.909750  \n",
       "max    142852.257000  33438.115000  1610.047000  34814.323000  30485.268000  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_zauc[data_zauc.columns[:10]].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We observe that the metabolite intensity and AUC are extremely skewed and span several orders of magnitude. We transform them to log scale for ease of visualisation&mdash;recall that tree-based algorithms are robust to feature scalings and outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_int = np.log1p(data_int)\n",
    "data_auc = np.log1p(data_auc)\n",
    "data_zauc = np.log1p(data_zauc)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualise what a single sample from the intensity dataset looks like"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(6,figsize=(20,12), sharex=True)\n",
    "for i, ax in enumerate(axes):\n",
    "    ax.plot(data_int.iloc[0].values[7500*i:min(7500*(i+1), data_int.shape[1])], c=colours[5])\n",
    "    ax.set_yticks([])\n",
    "    ax.set_xticks([])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And one single peak across samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'log intensity peak 3')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "sns.distplot(data_int[data_int.columns[3]], kde=False, color=colours[5])\n",
    "plt.ylabel('count')\n",
    "plt.xlabel('log intensity peak '+data_int.columns[3])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we remove any constant columns (zero variance) from the metabolite data, since they will not contribute to the classification. In this case, the constant will be equal to zero, indicating that a certain metabolite is not present in *any* of the samples. Furthermore, we can remove near-zero-variance features, *i.e.*, metabolites that are only present in very few of the samples, anticipating that they will also be poor class predictors. We set the threshold at 20 values, which is ca. 10% of the samples.\n",
    "\n",
    "We also recall that we do not have metadata for all samples, and therefore some of them will not contribute to training the model. For this reason we apply the column removal step only considering the labelled samples; in other words, if a certain compound has (near-) zero variance, we remove its corresponding column even if some unlabelled samples have (many) non-zero values. The removal applies to all samples, however."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# only consider samples with metadata\n",
    "\n",
    "data_int2 = data_int.loc[metadata.index]\n",
    "data_auc2 = data_auc.loc[metadata.index]\n",
    "data_zauc2 = data_zauc.loc[metadata.index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of constant columns in intensity data: 2667\n",
      "Number of constant columns in AUC data: 2626\n",
      "Number of constant columns in zeroed AUC data: 2716\n"
     ]
    }
   ],
   "source": [
    "nothing_int = [col for col in data_int.columns if len(data_int2[col].nonzero()[0]) == 0]\n",
    "nothing_auc = [col for col in data_auc.columns if len(data_auc2[col].nonzero()[0]) == 0]\n",
    "nothing_zauc = [col for col in data_zauc.columns if len(data_zauc2[col].nonzero()[0]) == 0]\n",
    "\n",
    "print('Number of constant columns in intensity data: %i' % len(nothing_int))\n",
    "print('Number of constant columns in AUC data: %i' % len(nothing_auc))\n",
    "print('Number of constant columns in zeroed AUC data: %i' % len(nothing_zauc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_int_Z = data_int.drop(nothing_int, axis=1)\n",
    "data_auc_Z = data_auc.drop(nothing_auc, axis=1)\n",
    "data_zauc_Z = data_zauc.drop(nothing_zauc, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of near-zero-variance columns in intensity data: 14449\n",
      "Number of near-zero-variance columns in AUC data: 14163\n",
      "Number of near-zero-variance columns in zeroed AUC data: 14752\n"
     ]
    }
   ],
   "source": [
    "nothing_int = [col for col in data_int.columns if len(data_int2[col].nonzero()[0]) <= 20]\n",
    "nothing_auc = [col for col in data_auc.columns if len(data_auc2[col].nonzero()[0]) <= 20]\n",
    "nothing_zauc = [col for col in data_zauc.columns if len(data_zauc2[col].nonzero()[0]) <= 20]\n",
    "\n",
    "print('Number of near-zero-variance columns in intensity data: %i' % len(nothing_int))\n",
    "print('Number of near-zero-variance columns in AUC data: %i' % len(nothing_auc))\n",
    "print('Number of near-zero-variance columns in zeroed AUC data: %i' % len(nothing_zauc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_int_NZ = data_int.drop(nothing_int, axis=1)\n",
    "data_auc_NZ = data_auc.drop(nothing_auc, axis=1)\n",
    "data_zauc_NZ = data_zauc.drop(nothing_zauc, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We save both the dataset with removal of zero-variance features and the one with removal of near-zero-variance features for each type of data, for future comparison."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_int_Z.to_csv('../data/model/data_int_Z.csv')\n",
    "data_auc_Z.to_csv('../data/model/data_auc_Z.csv')\n",
    "data_zauc_Z.to_csv('../data/model/data_zauc_Z.csv')\n",
    "\n",
    "data_int_NZ.to_csv('../data/model/data_int_NZ.csv')\n",
    "data_auc_NZ.to_csv('../data/model/data_auc_NZ.csv')\n",
    "data_zauc_NZ.to_csv('../data/model/data_zauc_NZ.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Clustering of correlated features\n",
    "\n",
    "Tree-based methods are also robust to highly correlated features: for a given split, they choose one of them ignoring the others. This robustness is good from a performance point of view, but the ignored features will not show up in an importance ranking, so that we may lose valuable information when interpreting the model. For this reason, we decide to cluster highly correlated features, removing from the dataset all but one member of any one cluster, but saving the cluster themselves, in order to go back to this data after determining the feature importances.\n",
    "\n",
    "We do this for the datasets with removal of near-zero-variance features, for two correlation thresholds: 0.99 and 0.95.\n",
    "\n",
    "**WARNING:** This is a memory-intensive process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def searchcluster(column, links):\n",
    "    cluster = [column]\n",
    "    \n",
    "    found = links[links.col1 == column].col2.values.tolist()\n",
    "    \n",
    "    cluster.extend(found)\n",
    "    \n",
    "    while len(found)>0:\n",
    "        found2 = links[links.col1 == found[0]].col2.values\n",
    "        found2 = np.setdiff1d(found2,cluster).tolist()\n",
    "        found.extend(found2)\n",
    "        cluster.extend(found2)\n",
    "        found.remove(found[0])\n",
    "    return cluster"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Intensity data\n",
    "\n",
    "We start by obtaining the correlation matrix of the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "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>3</th>\n",
       "      <th>4</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>16</th>\n",
       "      <th>...</th>\n",
       "      <th>44824</th>\n",
       "      <th>44825</th>\n",
       "      <th>44827</th>\n",
       "      <th>44829</th>\n",
       "      <th>44830</th>\n",
       "      <th>44831</th>\n",
       "      <th>44832</th>\n",
       "      <th>44833</th>\n",
       "      <th>44834</th>\n",
       "      <th>44835</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.109281</td>\n",
       "      <td>-0.178056</td>\n",
       "      <td>-0.184429</td>\n",
       "      <td>-0.113124</td>\n",
       "      <td>-0.381767</td>\n",
       "      <td>-0.123714</td>\n",
       "      <td>0.453836</td>\n",
       "      <td>0.095157</td>\n",
       "      <td>-0.022428</td>\n",
       "      <td>...</td>\n",
       "      <td>0.109598</td>\n",
       "      <td>0.006696</td>\n",
       "      <td>-0.009274</td>\n",
       "      <td>0.087997</td>\n",
       "      <td>0.414920</td>\n",
       "      <td>0.166689</td>\n",
       "      <td>0.193668</td>\n",
       "      <td>0.062510</td>\n",
       "      <td>0.063497</td>\n",
       "      <td>0.035290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.109281</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.045433</td>\n",
       "      <td>-0.072321</td>\n",
       "      <td>-0.001930</td>\n",
       "      <td>-0.124536</td>\n",
       "      <td>-0.092990</td>\n",
       "      <td>0.145530</td>\n",
       "      <td>0.414635</td>\n",
       "      <td>0.190070</td>\n",
       "      <td>...</td>\n",
       "      <td>0.056922</td>\n",
       "      <td>0.437089</td>\n",
       "      <td>0.076056</td>\n",
       "      <td>0.395492</td>\n",
       "      <td>0.221354</td>\n",
       "      <td>0.071303</td>\n",
       "      <td>0.067063</td>\n",
       "      <td>-0.001001</td>\n",
       "      <td>-0.022691</td>\n",
       "      <td>0.463866</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>-0.178056</td>\n",
       "      <td>-0.045433</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.598218</td>\n",
       "      <td>0.269298</td>\n",
       "      <td>0.405113</td>\n",
       "      <td>0.095899</td>\n",
       "      <td>-0.411802</td>\n",
       "      <td>0.077747</td>\n",
       "      <td>0.085698</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.131957</td>\n",
       "      <td>0.028953</td>\n",
       "      <td>0.002611</td>\n",
       "      <td>0.021390</td>\n",
       "      <td>-0.216956</td>\n",
       "      <td>-0.202139</td>\n",
       "      <td>-0.211420</td>\n",
       "      <td>-0.077814</td>\n",
       "      <td>-0.063961</td>\n",
       "      <td>0.044311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>-0.184429</td>\n",
       "      <td>-0.072321</td>\n",
       "      <td>0.598218</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.376614</td>\n",
       "      <td>0.393442</td>\n",
       "      <td>-0.003811</td>\n",
       "      <td>-0.424493</td>\n",
       "      <td>0.000048</td>\n",
       "      <td>0.036664</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.096665</td>\n",
       "      <td>-0.004609</td>\n",
       "      <td>0.029484</td>\n",
       "      <td>-0.024239</td>\n",
       "      <td>-0.241673</td>\n",
       "      <td>-0.130833</td>\n",
       "      <td>-0.139113</td>\n",
       "      <td>-0.019606</td>\n",
       "      <td>-0.019099</td>\n",
       "      <td>-0.004183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>-0.113124</td>\n",
       "      <td>-0.001930</td>\n",
       "      <td>0.269298</td>\n",
       "      <td>0.376614</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.268072</td>\n",
       "      <td>-0.001185</td>\n",
       "      <td>-0.275045</td>\n",
       "      <td>0.035769</td>\n",
       "      <td>0.029817</td>\n",
       "      <td>...</td>\n",
       "      <td>0.062795</td>\n",
       "      <td>0.044746</td>\n",
       "      <td>0.128153</td>\n",
       "      <td>0.036428</td>\n",
       "      <td>-0.169372</td>\n",
       "      <td>0.127369</td>\n",
       "      <td>0.171970</td>\n",
       "      <td>0.049209</td>\n",
       "      <td>0.078293</td>\n",
       "      <td>0.001451</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30387 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          3         4         7         8         9        10        11  \\\n",
       "3  1.000000  0.109281 -0.178056 -0.184429 -0.113124 -0.381767 -0.123714   \n",
       "4  0.109281  1.000000 -0.045433 -0.072321 -0.001930 -0.124536 -0.092990   \n",
       "7 -0.178056 -0.045433  1.000000  0.598218  0.269298  0.405113  0.095899   \n",
       "8 -0.184429 -0.072321  0.598218  1.000000  0.376614  0.393442 -0.003811   \n",
       "9 -0.113124 -0.001930  0.269298  0.376614  1.000000  0.268072 -0.001185   \n",
       "\n",
       "         12        13        16    ...        44824     44825     44827  \\\n",
       "3  0.453836  0.095157 -0.022428    ...     0.109598  0.006696 -0.009274   \n",
       "4  0.145530  0.414635  0.190070    ...     0.056922  0.437089  0.076056   \n",
       "7 -0.411802  0.077747  0.085698    ...    -0.131957  0.028953  0.002611   \n",
       "8 -0.424493  0.000048  0.036664    ...    -0.096665 -0.004609  0.029484   \n",
       "9 -0.275045  0.035769  0.029817    ...     0.062795  0.044746  0.128153   \n",
       "\n",
       "      44829     44830     44831     44832     44833     44834     44835  \n",
       "3  0.087997  0.414920  0.166689  0.193668  0.062510  0.063497  0.035290  \n",
       "4  0.395492  0.221354  0.071303  0.067063 -0.001001 -0.022691  0.463866  \n",
       "7  0.021390 -0.216956 -0.202139 -0.211420 -0.077814 -0.063961  0.044311  \n",
       "8 -0.024239 -0.241673 -0.130833 -0.139113 -0.019606 -0.019099 -0.004183  \n",
       "9  0.036428 -0.169372  0.127369  0.171970  0.049209  0.078293  0.001451  \n",
       "\n",
       "[5 rows x 30387 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corrmatrix = data_int_NZ.corr()\n",
    "corrmatrix.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We find which columns are members of a cluster, *any* cluster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_cl099 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.99)>1]\n",
    "in_cl095 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.95)>1]\n",
    "in_cl090 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.9)>1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We take a reduced correlation matrix, considering only columns that are members of clusters. We save a list of pairwise links for future use in a network visualisation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr099 = data_int_NZ[in_cl099].corr()\n",
    "corr095 = data_int_NZ[in_cl095].corr()\n",
    "corr090 = data_int_NZ[in_cl090].corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12</td>\n",
       "      <td>12</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12</td>\n",
       "      <td>29</td>\n",
       "      <td>-0.406414</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>47</td>\n",
       "      <td>-0.250751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>12</td>\n",
       "      <td>70</td>\n",
       "      <td>-0.267992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>84</td>\n",
       "      <td>-0.856772</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  col1 col2      corr\n",
       "0   12   12  1.000000\n",
       "1   12   29 -0.406414\n",
       "2   12   47 -0.250751\n",
       "3   12   70 -0.267992\n",
       "4   12   84 -0.856772"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links099 = corr099.stack().reset_index()\n",
    "links099.columns = ['col1', 'col2', 'corr']\n",
    "links095 = corr095.stack().reset_index()\n",
    "links095.columns = ['col1', 'col2', 'corr']\n",
    "links090 = corr090.stack().reset_index()\n",
    "links090.columns = ['col1', 'col2', 'corr']\n",
    "\n",
    "links099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1403</th>\n",
       "      <td>12</td>\n",
       "      <td>28754</td>\n",
       "      <td>0.996142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2734</th>\n",
       "      <td>29</td>\n",
       "      <td>1025</td>\n",
       "      <td>0.996293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7372</th>\n",
       "      <td>47</td>\n",
       "      <td>37806</td>\n",
       "      <td>0.990106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8540</th>\n",
       "      <td>70</td>\n",
       "      <td>12946</td>\n",
       "      <td>0.992204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12227</th>\n",
       "      <td>84</td>\n",
       "      <td>32450</td>\n",
       "      <td>0.992771</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      col1   col2      corr\n",
       "1403    12  28754  0.996142\n",
       "2734    29   1025  0.996293\n",
       "7372    47  37806  0.990106\n",
       "8540    70  12946  0.992204\n",
       "12227   84  32450  0.992771"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links_filtered099=links099.loc[ (np.abs(links099['corr']) > 0.99) & (links099['col1'] != links099['col2']) ]\n",
    "links_filtered095=links095.loc[ (np.abs(links095['corr']) > 0.95) & (links095['col1'] != links095['col2']) ]\n",
    "links_filtered090=links090.loc[ (np.abs(links090['corr']) > 0.9) & (links090['col1'] != links090['col2']) ]\n",
    "\n",
    "links_filtered099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "links_filtered099.to_csv('../data/links099_int.csv')\n",
    "links_filtered095.to_csv('../data/links095_int.csv')\n",
    "links_filtered090.to_csv('../data/links090_int.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we separate the actual clusters and their given members."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.99: 609\n"
     ]
    }
   ],
   "source": [
    "clusters099 = []\n",
    "cols = links_filtered099.col1.values\n",
    "cols2 = links_filtered099.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered099)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters099.append(cluster)\n",
    "    \n",
    "print('Number of clusters with threshold 0.99: %i' % len(clusters099))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.95: 1178\n"
     ]
    }
   ],
   "source": [
    "clusters095 = []\n",
    "cols = links_filtered095.col1.values\n",
    "cols2 = links_filtered095.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered095)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters095.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.95: %i' % len(clusters095))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.90: 1311\n"
     ]
    }
   ],
   "source": [
    "clusters090 = []\n",
    "cols = links_filtered090.col1.values\n",
    "cols2 = links_filtered090.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered090)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters090.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.90: %i' % len(clusters090))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We finally remove all but one member of every cluster, and save the resulting dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop099 = []\n",
    "for x in clusters099:\n",
    "    drop099.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_int_099 = data_int_NZ.drop(drop099, axis=1)\n",
    "\n",
    "data_int_099.to_csv('../data/model/data_int_099.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop095 = []\n",
    "for x in clusters095:\n",
    "    drop095.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_int_095 = data_int_NZ.drop(drop095, axis=1)\n",
    "\n",
    "data_int_095.to_csv('../data/model/data_int_095.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop090 = []\n",
    "for x in clusters090:\n",
    "    drop090.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_int_090 = data_int_NZ.drop(drop090, axis=1)\n",
    "\n",
    "data_int_090.to_csv('../data/model/data_int_090.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AUC data\n",
    "\n",
    "We start by obtaining the correlation matrix of the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "corrmatrix = data_auc_NZ.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "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>3</th>\n",
       "      <th>4</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>16</th>\n",
       "      <th>...</th>\n",
       "      <th>44824</th>\n",
       "      <th>44825</th>\n",
       "      <th>44827</th>\n",
       "      <th>44829</th>\n",
       "      <th>44830</th>\n",
       "      <th>44831</th>\n",
       "      <th>44832</th>\n",
       "      <th>44833</th>\n",
       "      <th>44834</th>\n",
       "      <th>44835</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.117007</td>\n",
       "      <td>-0.156911</td>\n",
       "      <td>-0.150367</td>\n",
       "      <td>-0.079332</td>\n",
       "      <td>-0.415777</td>\n",
       "      <td>-0.128657</td>\n",
       "      <td>0.443976</td>\n",
       "      <td>0.087506</td>\n",
       "      <td>0.031878</td>\n",
       "      <td>...</td>\n",
       "      <td>0.118129</td>\n",
       "      <td>-0.006491</td>\n",
       "      <td>-0.004430</td>\n",
       "      <td>0.067373</td>\n",
       "      <td>0.383341</td>\n",
       "      <td>0.160449</td>\n",
       "      <td>0.123000</td>\n",
       "      <td>0.102598</td>\n",
       "      <td>0.103587</td>\n",
       "      <td>0.011084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.117007</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.058498</td>\n",
       "      <td>-0.084942</td>\n",
       "      <td>-0.007788</td>\n",
       "      <td>-0.149104</td>\n",
       "      <td>-0.101978</td>\n",
       "      <td>0.180674</td>\n",
       "      <td>0.413428</td>\n",
       "      <td>0.213327</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064271</td>\n",
       "      <td>0.429839</td>\n",
       "      <td>0.081999</td>\n",
       "      <td>0.384449</td>\n",
       "      <td>0.240133</td>\n",
       "      <td>0.062137</td>\n",
       "      <td>0.062299</td>\n",
       "      <td>0.006499</td>\n",
       "      <td>-0.007263</td>\n",
       "      <td>0.456285</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>-0.156911</td>\n",
       "      <td>-0.058498</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.572860</td>\n",
       "      <td>0.279382</td>\n",
       "      <td>0.384045</td>\n",
       "      <td>0.106053</td>\n",
       "      <td>-0.376650</td>\n",
       "      <td>0.040023</td>\n",
       "      <td>0.006777</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.138010</td>\n",
       "      <td>0.026769</td>\n",
       "      <td>0.022621</td>\n",
       "      <td>0.008287</td>\n",
       "      <td>-0.198146</td>\n",
       "      <td>-0.213532</td>\n",
       "      <td>-0.188461</td>\n",
       "      <td>-0.050542</td>\n",
       "      <td>-0.057962</td>\n",
       "      <td>0.028088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>-0.150367</td>\n",
       "      <td>-0.084942</td>\n",
       "      <td>0.572860</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.397003</td>\n",
       "      <td>0.369870</td>\n",
       "      <td>0.010962</td>\n",
       "      <td>-0.393418</td>\n",
       "      <td>0.012009</td>\n",
       "      <td>-0.023976</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.096860</td>\n",
       "      <td>0.012073</td>\n",
       "      <td>0.057702</td>\n",
       "      <td>-0.035809</td>\n",
       "      <td>-0.214490</td>\n",
       "      <td>-0.150547</td>\n",
       "      <td>-0.126172</td>\n",
       "      <td>-0.034046</td>\n",
       "      <td>-0.017611</td>\n",
       "      <td>0.015154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>-0.079332</td>\n",
       "      <td>-0.007788</td>\n",
       "      <td>0.279382</td>\n",
       "      <td>0.397003</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.298946</td>\n",
       "      <td>-0.001096</td>\n",
       "      <td>-0.320142</td>\n",
       "      <td>-0.003282</td>\n",
       "      <td>-0.035315</td>\n",
       "      <td>...</td>\n",
       "      <td>0.005705</td>\n",
       "      <td>0.025212</td>\n",
       "      <td>0.117331</td>\n",
       "      <td>0.036779</td>\n",
       "      <td>-0.175996</td>\n",
       "      <td>0.140625</td>\n",
       "      <td>0.204140</td>\n",
       "      <td>0.030772</td>\n",
       "      <td>0.048990</td>\n",
       "      <td>-0.033766</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30673 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          3         4         7         8         9        10        11  \\\n",
       "3  1.000000  0.117007 -0.156911 -0.150367 -0.079332 -0.415777 -0.128657   \n",
       "4  0.117007  1.000000 -0.058498 -0.084942 -0.007788 -0.149104 -0.101978   \n",
       "7 -0.156911 -0.058498  1.000000  0.572860  0.279382  0.384045  0.106053   \n",
       "8 -0.150367 -0.084942  0.572860  1.000000  0.397003  0.369870  0.010962   \n",
       "9 -0.079332 -0.007788  0.279382  0.397003  1.000000  0.298946 -0.001096   \n",
       "\n",
       "         12        13        16    ...        44824     44825     44827  \\\n",
       "3  0.443976  0.087506  0.031878    ...     0.118129 -0.006491 -0.004430   \n",
       "4  0.180674  0.413428  0.213327    ...     0.064271  0.429839  0.081999   \n",
       "7 -0.376650  0.040023  0.006777    ...    -0.138010  0.026769  0.022621   \n",
       "8 -0.393418  0.012009 -0.023976    ...    -0.096860  0.012073  0.057702   \n",
       "9 -0.320142 -0.003282 -0.035315    ...     0.005705  0.025212  0.117331   \n",
       "\n",
       "      44829     44830     44831     44832     44833     44834     44835  \n",
       "3  0.067373  0.383341  0.160449  0.123000  0.102598  0.103587  0.011084  \n",
       "4  0.384449  0.240133  0.062137  0.062299  0.006499 -0.007263  0.456285  \n",
       "7  0.008287 -0.198146 -0.213532 -0.188461 -0.050542 -0.057962  0.028088  \n",
       "8 -0.035809 -0.214490 -0.150547 -0.126172 -0.034046 -0.017611  0.015154  \n",
       "9  0.036779 -0.175996  0.140625  0.204140  0.030772  0.048990 -0.033766  \n",
       "\n",
       "[5 rows x 30673 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corrmatrix.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We find which columns are members of a cluster, *any* cluster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_cl099 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.99)>1]\n",
    "in_cl095 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.95)>1]\n",
    "in_cl090 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.9)>1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We take a reduced correlation matrix, considering only columns that are members of clusters. We save a list of pairwise links for future use in a network visualisation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr099 = data_auc_NZ[in_cl099].corr()\n",
    "corr095 = data_auc_NZ[in_cl095].corr()\n",
    "corr090 = data_auc_NZ[in_cl090].corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12</td>\n",
       "      <td>12</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12</td>\n",
       "      <td>29</td>\n",
       "      <td>-0.380066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>190</td>\n",
       "      <td>-0.083151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>12</td>\n",
       "      <td>384</td>\n",
       "      <td>0.126252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>410</td>\n",
       "      <td>0.147638</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  col1 col2      corr\n",
       "0   12   12  1.000000\n",
       "1   12   29 -0.380066\n",
       "2   12  190 -0.083151\n",
       "3   12  384  0.126252\n",
       "4   12  410  0.147638"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links099 = corr099.stack().reset_index()\n",
    "links099.columns = ['col1', 'col2', 'corr']\n",
    "links095 = corr095.stack().reset_index()\n",
    "links095.columns = ['col1', 'col2', 'corr']\n",
    "links090 = corr090.stack().reset_index()\n",
    "links090.columns = ['col1', 'col2', 'corr']\n",
    "\n",
    "links099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>12</td>\n",
       "      <td>28754</td>\n",
       "      <td>0.993804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>492</th>\n",
       "      <td>29</td>\n",
       "      <td>1025</td>\n",
       "      <td>0.998821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>989</th>\n",
       "      <td>190</td>\n",
       "      <td>2115</td>\n",
       "      <td>0.991208</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1069</th>\n",
       "      <td>190</td>\n",
       "      <td>12754</td>\n",
       "      <td>0.990546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1336</th>\n",
       "      <td>190</td>\n",
       "      <td>35569</td>\n",
       "      <td>0.993667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     col1   col2      corr\n",
       "293    12  28754  0.993804\n",
       "492    29   1025  0.998821\n",
       "989   190   2115  0.991208\n",
       "1069  190  12754  0.990546\n",
       "1336  190  35569  0.993667"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links_filtered099=links099.loc[ (np.abs(links099['corr']) > 0.99) & (links099['col1'] != links099['col2']) ]\n",
    "links_filtered095=links095.loc[ (np.abs(links095['corr']) > 0.95) & (links095['col1'] != links095['col2']) ]\n",
    "links_filtered090=links090.loc[ (np.abs(links090['corr']) > 0.9) & (links090['col1'] != links090['col2']) ]\n",
    "\n",
    "links_filtered099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "links_filtered099.to_csv('../data/links099_auc.csv')\n",
    "links_filtered095.to_csv('../data/links095_auc.csv')\n",
    "links_filtered090.to_csv('../data/links090_auc.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we separate the actual clusters and their given members."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.99: 175\n"
     ]
    }
   ],
   "source": [
    "clusters099 = []\n",
    "cols = links_filtered099.col1.values\n",
    "cols2 = links_filtered099.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered099)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters099.append(cluster)\n",
    "    \n",
    "print('Number of clusters with threshold 0.99: %i' % len(clusters099))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.95: 1054\n"
     ]
    }
   ],
   "source": [
    "clusters095 = []\n",
    "cols = links_filtered095.col1.values\n",
    "cols2 = links_filtered095.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered095)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters095.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.95: %i' % len(clusters095))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.90: 1270\n"
     ]
    }
   ],
   "source": [
    "clusters090 = []\n",
    "cols = links_filtered090.col1.values\n",
    "cols2 = links_filtered090.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered090)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters090.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.90: %i' % len(clusters090))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We finally remove all but one member of every cluster, and save the resulting dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop099 = []\n",
    "for x in clusters099:\n",
    "    drop099.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_auc_099 = data_auc_NZ.drop(drop099, axis=1)\n",
    "\n",
    "data_auc_099.to_csv('../data/model/data_auc_099.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop095 = []\n",
    "for x in clusters095:\n",
    "    drop095.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_auc_095 = data_auc_NZ.drop(drop095, axis=1)\n",
    "\n",
    "data_auc_095.to_csv('../data/model/data_auc_095.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop090 = []\n",
    "for x in clusters090:\n",
    "    drop090.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_auc_090 = data_auc_NZ.drop(drop090, axis=1)\n",
    "\n",
    "data_auc_090.to_csv('../data/model/data_auc_090.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Zeroed AUC data\n",
    "\n",
    "We start by obtaining the correlation matrix of the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "corrmatrix = data_zauc_NZ.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "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>3</th>\n",
       "      <th>4</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>16</th>\n",
       "      <th>...</th>\n",
       "      <th>44824</th>\n",
       "      <th>44825</th>\n",
       "      <th>44827</th>\n",
       "      <th>44829</th>\n",
       "      <th>44830</th>\n",
       "      <th>44831</th>\n",
       "      <th>44832</th>\n",
       "      <th>44833</th>\n",
       "      <th>44834</th>\n",
       "      <th>44835</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.108550</td>\n",
       "      <td>-0.166967</td>\n",
       "      <td>-0.169697</td>\n",
       "      <td>-0.073474</td>\n",
       "      <td>-0.407935</td>\n",
       "      <td>-0.122661</td>\n",
       "      <td>0.442838</td>\n",
       "      <td>0.083154</td>\n",
       "      <td>0.023746</td>\n",
       "      <td>...</td>\n",
       "      <td>0.099348</td>\n",
       "      <td>-0.008651</td>\n",
       "      <td>0.000088</td>\n",
       "      <td>0.068016</td>\n",
       "      <td>0.388988</td>\n",
       "      <td>0.126047</td>\n",
       "      <td>0.111096</td>\n",
       "      <td>0.111046</td>\n",
       "      <td>0.109691</td>\n",
       "      <td>0.012503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.108550</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.051155</td>\n",
       "      <td>-0.090007</td>\n",
       "      <td>0.009373</td>\n",
       "      <td>-0.147544</td>\n",
       "      <td>-0.101121</td>\n",
       "      <td>0.176458</td>\n",
       "      <td>0.413428</td>\n",
       "      <td>0.213327</td>\n",
       "      <td>...</td>\n",
       "      <td>0.061921</td>\n",
       "      <td>0.429839</td>\n",
       "      <td>0.076448</td>\n",
       "      <td>0.387000</td>\n",
       "      <td>0.240133</td>\n",
       "      <td>0.059661</td>\n",
       "      <td>0.080297</td>\n",
       "      <td>0.005289</td>\n",
       "      <td>-0.009874</td>\n",
       "      <td>0.458139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>-0.166967</td>\n",
       "      <td>-0.051155</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.591054</td>\n",
       "      <td>0.287745</td>\n",
       "      <td>0.388247</td>\n",
       "      <td>0.101624</td>\n",
       "      <td>-0.387061</td>\n",
       "      <td>0.042510</td>\n",
       "      <td>0.013640</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.132656</td>\n",
       "      <td>0.034408</td>\n",
       "      <td>0.026978</td>\n",
       "      <td>0.018111</td>\n",
       "      <td>-0.196162</td>\n",
       "      <td>-0.202755</td>\n",
       "      <td>-0.176040</td>\n",
       "      <td>-0.060072</td>\n",
       "      <td>-0.064763</td>\n",
       "      <td>0.038204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>-0.169697</td>\n",
       "      <td>-0.090007</td>\n",
       "      <td>0.591054</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.402045</td>\n",
       "      <td>0.378447</td>\n",
       "      <td>0.005174</td>\n",
       "      <td>-0.414167</td>\n",
       "      <td>0.007318</td>\n",
       "      <td>-0.005253</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.085392</td>\n",
       "      <td>0.013499</td>\n",
       "      <td>0.054468</td>\n",
       "      <td>-0.022070</td>\n",
       "      <td>-0.227449</td>\n",
       "      <td>-0.132620</td>\n",
       "      <td>-0.101164</td>\n",
       "      <td>-0.036079</td>\n",
       "      <td>-0.036588</td>\n",
       "      <td>0.016330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>-0.073474</td>\n",
       "      <td>0.009373</td>\n",
       "      <td>0.287745</td>\n",
       "      <td>0.402045</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.287205</td>\n",
       "      <td>0.008489</td>\n",
       "      <td>-0.302262</td>\n",
       "      <td>0.011931</td>\n",
       "      <td>-0.019224</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.004209</td>\n",
       "      <td>0.035414</td>\n",
       "      <td>0.118670</td>\n",
       "      <td>0.044376</td>\n",
       "      <td>-0.156144</td>\n",
       "      <td>0.089251</td>\n",
       "      <td>0.181120</td>\n",
       "      <td>0.014313</td>\n",
       "      <td>0.042610</td>\n",
       "      <td>-0.019397</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30084 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          3         4         7         8         9        10        11  \\\n",
       "3  1.000000  0.108550 -0.166967 -0.169697 -0.073474 -0.407935 -0.122661   \n",
       "4  0.108550  1.000000 -0.051155 -0.090007  0.009373 -0.147544 -0.101121   \n",
       "7 -0.166967 -0.051155  1.000000  0.591054  0.287745  0.388247  0.101624   \n",
       "8 -0.169697 -0.090007  0.591054  1.000000  0.402045  0.378447  0.005174   \n",
       "9 -0.073474  0.009373  0.287745  0.402045  1.000000  0.287205  0.008489   \n",
       "\n",
       "         12        13        16    ...        44824     44825     44827  \\\n",
       "3  0.442838  0.083154  0.023746    ...     0.099348 -0.008651  0.000088   \n",
       "4  0.176458  0.413428  0.213327    ...     0.061921  0.429839  0.076448   \n",
       "7 -0.387061  0.042510  0.013640    ...    -0.132656  0.034408  0.026978   \n",
       "8 -0.414167  0.007318 -0.005253    ...    -0.085392  0.013499  0.054468   \n",
       "9 -0.302262  0.011931 -0.019224    ...    -0.004209  0.035414  0.118670   \n",
       "\n",
       "      44829     44830     44831     44832     44833     44834     44835  \n",
       "3  0.068016  0.388988  0.126047  0.111096  0.111046  0.109691  0.012503  \n",
       "4  0.387000  0.240133  0.059661  0.080297  0.005289 -0.009874  0.458139  \n",
       "7  0.018111 -0.196162 -0.202755 -0.176040 -0.060072 -0.064763  0.038204  \n",
       "8 -0.022070 -0.227449 -0.132620 -0.101164 -0.036079 -0.036588  0.016330  \n",
       "9  0.044376 -0.156144  0.089251  0.181120  0.014313  0.042610 -0.019397  \n",
       "\n",
       "[5 rows x 30084 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corrmatrix.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We find which columns are members of a cluster, *any* cluster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "in_cl099 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.99)>1]\n",
    "in_cl095 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.95)>1]\n",
    "in_cl090 = [column for column in corrmatrix.columns if sum(corrmatrix[column].abs() > 0.9)>1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We take a reduced correlation matrix, considering only columns that are members of clusters. We save a list of pairwise links for future use in a network visualisation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr099 = data_zauc_NZ[in_cl099].corr()\n",
    "corr095 = data_zauc_NZ[in_cl095].corr()\n",
    "corr090 = data_zauc_NZ[in_cl090].corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12</td>\n",
       "      <td>12</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12</td>\n",
       "      <td>29</td>\n",
       "      <td>-0.390829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>85</td>\n",
       "      <td>-0.256677</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>12</td>\n",
       "      <td>103</td>\n",
       "      <td>-0.324104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>106</td>\n",
       "      <td>-0.263684</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  col1 col2      corr\n",
       "0   12   12  1.000000\n",
       "1   12   29 -0.390829\n",
       "2   12   85 -0.256677\n",
       "3   12  103 -0.324104\n",
       "4   12  106 -0.263684"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links099 = corr099.stack().reset_index()\n",
    "links099.columns = ['col1', 'col2', 'corr']\n",
    "links095 = corr095.stack().reset_index()\n",
    "links095.columns = ['col1', 'col2', 'corr']\n",
    "links090 = corr090.stack().reset_index()\n",
    "links090.columns = ['col1', 'col2', 'corr']\n",
    "\n",
    "links099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "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>col1</th>\n",
       "      <th>col2</th>\n",
       "      <th>corr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>450</th>\n",
       "      <td>12</td>\n",
       "      <td>28754</td>\n",
       "      <td>0.993848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>712</th>\n",
       "      <td>29</td>\n",
       "      <td>1025</td>\n",
       "      <td>0.998800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1376</th>\n",
       "      <td>85</td>\n",
       "      <td>106</td>\n",
       "      <td>0.991310</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2177</th>\n",
       "      <td>103</td>\n",
       "      <td>8332</td>\n",
       "      <td>0.997594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2746</th>\n",
       "      <td>106</td>\n",
       "      <td>85</td>\n",
       "      <td>0.991310</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     col1   col2      corr\n",
       "450    12  28754  0.993848\n",
       "712    29   1025  0.998800\n",
       "1376   85    106  0.991310\n",
       "2177  103   8332  0.997594\n",
       "2746  106     85  0.991310"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "links_filtered099=links099.loc[ (np.abs(links099['corr']) > 0.99) & (links099['col1'] != links099['col2']) ]\n",
    "links_filtered095=links095.loc[ (np.abs(links095['corr']) > 0.95) & (links095['col1'] != links095['col2']) ]\n",
    "links_filtered090=links090.loc[ (np.abs(links090['corr']) > 0.9) & (links090['col1'] != links090['col2']) ]\n",
    "\n",
    "links_filtered099.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "links_filtered099.to_csv('../data/links099_zauc.csv')\n",
    "links_filtered095.to_csv('../data/links095_zauc.csv')\n",
    "links_filtered090.to_csv('../data/links090_zauc.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we separate the actual clusters and their given members."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.99: 218\n"
     ]
    }
   ],
   "source": [
    "clusters099 = []\n",
    "cols = links_filtered099.col1.values\n",
    "cols2 = links_filtered099.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered099)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters099.append(cluster)\n",
    "    \n",
    "print('Number of clusters with threshold 0.99: %i' % len(clusters099))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.95: 1038\n"
     ]
    }
   ],
   "source": [
    "clusters095 = []\n",
    "cols = links_filtered095.col1.values\n",
    "cols2 = links_filtered095.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered095)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters095.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.95: %i' % len(clusters095))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of clusters with threshold 0.90: 1290\n"
     ]
    }
   ],
   "source": [
    "clusters090 = []\n",
    "cols = links_filtered090.col1.values\n",
    "cols2 = links_filtered090.col1.values\n",
    "for column in cols:\n",
    "    if column not in cols2:\n",
    "        continue\n",
    "    cluster = searchcluster(column, links_filtered090)\n",
    "    cols2 = np.setdiff1d(cols2,cluster)\n",
    "    clusters090.append(cluster)\n",
    "\n",
    "print('Number of clusters with threshold 0.90: %i' % len(clusters090))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We finally remove all but one member of every cluster, and save the resulting dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop099 = []\n",
    "for x in clusters099:\n",
    "    drop099.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_zauc_099 = data_zauc_NZ.drop(drop099, axis=1)\n",
    "\n",
    "data_zauc_099.to_csv('../data/model/data_zauc_099.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop095 = []\n",
    "for x in clusters095:\n",
    "    drop095.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_zauc_095 = data_zauc_NZ.drop(drop095, axis=1)\n",
    "\n",
    "data_zauc_095.to_csv('../data/model/data_zauc_095.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "drop090 = []\n",
    "for x in clusters090:\n",
    "    drop090.extend(list(map(str, x[1:])))\n",
    "    \n",
    "data_zauc_090 = data_zauc_NZ.drop(drop090, axis=1)\n",
    "\n",
    "data_zauc_090.to_csv('../data/model/data_zauc_090.csv')"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}