Alveo NER.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sample Workflow: Alveo\n",
    "\n",
    "This worksheet pulls data from the Alveo API and performs some NER using Spacy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install -q -r requirements.txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import spacy\n",
    "import csv\n",
    "import geocoder\n",
    "import pandas as pd\n",
    "import re\n",
    "import utils\n",
    "import pyalveo"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Alveo requires a login and uses an API key to validate user requests. We read this from the file `secret.json`.  \n",
    "\n",
    "The data we will work with is represented by an [item list](http://alveo.edu.au/documentation/discovering-and-searching-the-collections/saving-your-search-results-to-an-item-list/) in Alveo - this is a list of items selected via a query as the starting point for a research project.   In this case I've selected three items from the [Braided Channels](https://app.alveo.edu.au/catalog/braidedchannels) collection that contains transcripts of oral history interviews.  Each item list has a URL and we refer to that here. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "API_KEY = utils.secret('alveo')\n",
    "API_URL = \"https://app.alveo.edu.au/\"\n",
    "item_list_url = \"https://app.alveo.edu.au/item_lists/1387\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We create an API client with the pyalveo module and use the client to get the item list details.  We then get the _primary text_ for each item.  We store these in a python list of texts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Item list name:  braided-sample\n",
      "Got 3 texts\n"
     ]
    }
   ],
   "source": [
    "client  =  pyalveo.Client(api_key=API_KEY, api_url=API_URL)\n",
    "itemlist = client.get_item_list(item_list_url)\n",
    "\n",
    "print(\"Item list name: \", itemlist.name())\n",
    "\n",
    "texts = []\n",
    "for itemurl in itemlist:\n",
    "    item = client.get_item(itemurl)\n",
    "    text = item.get_primary_text()\n",
    "    text = text.decode() # convert from bytes to a string\n",
    "    text = re.sub('\\W+', ' ', text)\n",
    "    texts.append(text) \n",
    "\n",
    "print(\"Got\", len(texts), \"texts\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## NER Using Spacy\n",
    "\n",
    "We will use Spacy to extract Named Entities from the text.   We download the appropriate models and initialise an NLP processor. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[93m    Linking successful\u001b[0m\n",
      "    /opt/conda/lib/python3.6/site-packages/en_core_web_md -->\n",
      "    /opt/conda/lib/python3.6/site-packages/spacy/data/en_core_web_md\n",
      "\n",
      "    You can now load the model via spacy.load('en_core_web_md')\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# download the spacy models we need\n",
    "model = 'en_core_web_md'\n",
    "spacy.cli.download(model)\n",
    "nlp = spacy.load(model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We then extract entities from the texts.  The results will be converted to a Pandas data frame. In this example we retain all of the entity types in the result and for each result include a _context_ string showing the words each side of the entity that was found. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found  578 entities in the texts\n"
     ]
    },
    {
     "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>context</th>\n",
       "      <th>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td></td>\n",
       "      <td>CARDINAL</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>A from 10_BC_DV This is DAT</td>\n",
       "      <td>CARDINAL</td>\n",
       "      <td>10_BC_DV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>It s 4 June 2000 Trish FitzSimons sound</td>\n",
       "      <td>DATE</td>\n",
       "      <td>4 June 2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>June 2000 Trish FitzSimons sound recording</td>\n",
       "      <td>NORP</td>\n",
       "      <td>Trish</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2000 Trish FitzSimons sound recording Erica</td>\n",
       "      <td>ORG</td>\n",
       "      <td>FitzSimons</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                       context     label         text\n",
       "0                                               CARDINAL            1\n",
       "1                  A from 10_BC_DV This is DAT  CARDINAL     10_BC_DV\n",
       "2      It s 4 June 2000 Trish FitzSimons sound      DATE  4 June 2000\n",
       "3   June 2000 Trish FitzSimons sound recording      NORP        Trish\n",
       "4  2000 Trish FitzSimons sound recording Erica       ORG   FitzSimons"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "places = []\n",
    "\n",
    "for text in texts:\n",
    "    doc = nlp(text)\n",
    "    for ent in doc.ents:\n",
    "        context = doc[ent.start-2:ent.end+3]\n",
    "        context = \" \".join([w.text for w in context])\n",
    "        d = {'label': ent.label_, 'text': ent.text, 'context': context}\n",
    "        places.append(d)\n",
    "\n",
    "entities = pd.DataFrame(places)\n",
    "print(\"Found \", entities.shape[0], \"entities in the texts\")\n",
    "entities.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We might be particularly interested in the GPE entities - locations.  We can select these as follows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>context</th>\n",
       "      <th>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>born in Brisbane in February 1954</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Brisbane</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>up in Brisbane and we had</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Brisbane</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>south of Boulia and that was</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Boulia</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>out to Alexandria in the Northern</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Alexandria</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>lot of Australia you ve Yes</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Australia</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "<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>context</th>\n",
       "      <th>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>born in Brisbane in February 1954</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Brisbane</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>up in Brisbane and we had</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Brisbane</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>south of Boulia and that was</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Boulia</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>out to Alexandria in the Northern</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Alexandria</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>lot of Australia you ve Yes</td>\n",
       "      <td>GPE</td>\n",
       "      <td>Australia</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "locations = entities[entities['label'] == 'GPE']\n",
    "locations.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can then plot the frequency of occurence of each place name in the texts. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fbf598019b0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAFnCAYAAAC2IbJmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3XeYpFWZ/vHvDUiQjIwJdhhFFkUUxUERMQBiFsMKggGMGFAQ02JaUfenqKw5siKgoogKuygrKAKCqAhDjsqyqChmRUTJ9++P8xZT01Pd02+o6Zp37s919dVd6anT3VVPnfe85zxHtomIiBXfKnPdgIiI6EYSekRETyShR0T0RBJ6RERPJKFHRPREEnpERE8koUdE9EQSekRETyShR0T0xGrL88k23nhjL1iwYHk+ZUTECm/RokV/sD1vWfdbrgl9wYIFnHvuucvzKSMiVniSfj6b+2XIJSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieWmdAlfV7S7yRdMnTdRpK+K+ln1fcNx9vMiIhYltn00I8EnjTluoOA79neAvhedTkiIubQMhO67TOAP025+hnAUdXPRwHP7LhdERFRU9OFRfewfR2A7esk3X26O0raF9gXYP78+Q2frh8uv/8DZnW/B1xx+ZhbEhF9NPaTorYPs73Q9sJ585a5cjUiIhpqmtB/K+leANX333XXpIiIaKJpQj8B2Kf6eR/gv7tpTkRENDWbaYtfAX4EbCnpWkkvBQ4BdpX0M2DX6nJERMyhZZ4Utb3XNDft0nFbIiKihawUjYjoiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieaJXQJR0o6VJJl0j6iqQ1u2pYRETU0zihS9oE2B9YaHtrYFVgz64aFhER9bQdclkNWEvSasBdgV+3b1JERDTROKHb/hVwKPAL4Drgetvf6aphERFRT5shlw2BZwD3Ae4NrC3pBSPut6+kcyWd+/vf/755SyMiYkZthlweD/yf7d/bvhU4Dthh6p1sH2Z7oe2F8+bNa/F0ERExkzYJ/RfA9pLuKknALsDl3TQrIiLqajOGfjbwdeA84OIq1mEdtSsiImparc2Dbb8TeGdHbYmIiBayUjQioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6otUGF333oKMeNKv7XbzPxWNuSUTEsqWHHhHRE0noERE9kYQeEdETSegRET2RhB4R0RNJ6BERPZGEHhHRE0noERE9kYQeEdETSegRET2RhB4R0RNJ6BERPZGEHhHRE0noERE9kYQeEdETSegRET3RKqFL2kDS1yVdIelySY/sqmEREVFP2x2LPgqcZPs5klYH7tpBmyIiooHGCV3SesBjgBcB2L4FuKWbZkVERF1teuj3BX4PHCFpG2ARcIDtG4fvJGlfYF+A+fPnt3i6WTh4/Vne7/rxtiMiYg60GUNfDdgW+LTthwI3AgdNvZPtw2wvtL1w3rx5LZ4uIiJm0iahXwtca/vs6vLXKQk+IiLmQOOEbvs3wC8lbVldtQtwWSetioiI2trOcnktcHQ1w+Vq4MXtmxQREU20Sui2LwAWdtSWiIhoIStFIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInqi7QYXrS046MRZ3e+aQ5465pZERKzY0kOPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInmid0CWtKul8Sd/qokEREdFMFz30A4DLO4gTEREttErokjYFngp8rpvmREREU2176B8B3gzc0UFbIiKihcabREt6GvA724skPW6G++0L7Aswf/78pk8XMbHuedoFs7rfb3Z6yJhbshwcvP4s73f9eNsRI7XpoT8K2E3SNcAxwM6SvjT1TrYPs73Q9sJ58+a1eLqIiJhJ44Ru+y22N7W9ANgTONX2CzprWURE1JJ56BERPdF4DH2Y7dOB07uIFRERzaSHHhHRE0noERE9kYQeEdETSegRET2RhB4R0RNJ6BERPZGEHhHRE0noERE9kYQeEdETSegRET2RhB4R0RNJ6BERPZGEHhHRE0noERE9kYQeEdETndRDj4juLDjoxFnd75pDnjrmlsSKJj30iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInmic0CX9k6TTJF0u6VJJB3TZsIiIqKfNFnS3AW+wfZ6kdYFFkr5r+7KO2hYRETU07qHbvs72edXPNwCXA5t01bCIiKink02iJS0AHgqcPeK2fYF9AebPn9/F00W08r1TN5/V/XbZ+X/H3JKIbrU+KSppHeAbwOts/3Xq7bYPs73Q9sJ58+a1fbqIiJhGq4Qu6S6UZH607eO6aVJERDTRZpaLgMOBy21/qLsmRUREE2166I8CXgjsLOmC6uspHbUrIiJqanxS1PYPAHXYloiIaCErRSMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6opNNomNufPKVpy7zPvt9ZudZxfqP5z5tVvd7w1e/Nav7XXvQmbO636aHPHpW9zv44IM7uc/KZsFBJ87qftcc8tQxt2S0Bx31oFnd7+J9Ll7mfS6//wNmFesBV1w+q/vN5v0Fc/ceGyU99IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJJPSIiJ5IQo+I6Ikk9IiInkhCj4joiST0iIieSEKPiOiJVgld0pMkXSnpKkkHddWoiIior3FCl7Qq8EngycBWwF6StuqqYRERUU+bHvrDgatsX237FuAY4BndNCsiIupqk9A3AX45dPna6rqIiJgDst3sgdLuwBNtv6y6/ELg4bZfO+V++wL7Vhe3BK6cRfiNgT80ath4Y016vElu26THm+S2dR1vktvWdbxJbludeJvZnresO63WoiHXAv80dHlT4NdT72T7MOCwOoElnWt7YYu2jSXWpMeb5LZNerxJblvX8Sa5bV3Hm+S2jSNemyGXc4AtJN1H0urAnsAJ3TQrIiLqatxDt32bpNcAJwOrAp+3fWlnLYuIiFraDLlg+3+A/+moLcNqDdEsx1iTHm+S2zbp8Sa5bV3Hm+S2dR1vktvWebzGJ0UjImKyZOl/RERPJKFHRPREEnqMnaQNJT14rtuxspK0u6R1q5/fLuk4SdvOdbuGSVpf0raSdhh8zXWbxkHSuZL2k7ThWOJnDH12qn/AFsCag+tsn9Ei3o7AFraPkDQPWMf2/zWII+D5wH1tv1vSfOCetn/StG1dkHQ6sBvlxPsFwO+B79t+fQexN6P87U6RtBawmu0bGsbaBnh0dfFM2xc2jLM98HHgAcDqlJlfN9per0m8KuZOwAMBA5fZPq1hnItsP7h6zb0POBR4q+1HNG1blyS9BHgDZaX5xcB2wI9tP65BrLUo74dBXalzga9X5Umatm8TYDOGJpE0fe9Luh/wYuC5VduOAL7jjhLxRCR0SVtQXmhbsWTCvG+DWGsCL6W8EYZjvaRF+14GHEBZPHUBsD3wI9s7N4z3TmAhsKXtf5Z0b+Brth/VINangTuAnW0/oPrg+Y7t7WrG+TglcYxke/+a8c63/dDqb/dPtt85SCx14oyI+3LKyuONbG9evXY+Y3uXBrEOAF4OHFdd9SzgMNsfbxDrXMpajK9R/rd7A/ez/bYGsTap2nQTsAgQsC2wFvAs27+qGW/wv3gfcLHtLw+ua9C2i5n5dVL7/1vFfDjlPfUQSQ8E3m57r5pxHgR8E/g+S/7dtgJ2Bd5o++01Y76fknwvA26vrrbt3erEGRF3FeBpwOD9+3ngo7b/1CZuq2mLHToCeCfwYWAnyieYGsb6InAF8ETg3ZRP68tbtu8AFvcadpJ0f+BdLeI9C3gocB6A7V8PDokbeITtbSWdX8X6c7XQq65zGz7/dFaTdC9gD6B2UpvBfpQ3/9kAtn8m6e4NY72U8ve7Ee588/6I0tOuzfZVkla1fTtwhKQfNmzXJ4BP2z5y+EpJewOfon4RvF9J+izweOD9ktag+XDr06rv+1Xfv1h9fz7w94Yxb7L9D0lIWt32pdV7rK6PAS+3/d3hKyU9HrgEaLJO5pmUjtfNDR47UjX8+GLgKcA3gKOBHYFTgYe0iT0pCX0t29+TJNs/Bw6WdCYlydd1P9u7S3qG7aMkfZmy+KmNm2zfVL3g1rB9haQtW8S7xbYlGUDS2i1i3VqVMh7Emkf5xK/F9lHDlyWtPUh0Db2b8nc/y/Y5ku4L/KxFvIGbbd9SRppA0mrM0GNcBrG410X1c9OOxN+rD9ILJH0AuA5o+n/dyvazpl5p+wuSmnw47gE8CTjU9l+qD9o3NWlY9f5E0qOmHFEeJOksyv+9ruskbUDpXZ8s6U/AbxvEudfUZF61+RRJt1I6UnVdDdwF6CShS1oE/AU4HDho6IPibEm1j9CnmpSEflN1CPKzavXpr4Cmva5bq+9/kbQ18BtgQcv2XVu94P4L+K6kPzOibk0Nx1Y9pg2qIYSXAP/ZMNbHgOOBu0v6f8BzgFqHlcMkPZLyYlsHmF+NMb/C9qvrxLH9Ncrww+Dy1cC/NG3XkO9LeiuwlqRdgVdTEkETR1DeSMdTEvkzKL97Ey+k9HpfAxxIqXPU9PddddSV1Xtk5G0zsf13Sf8N3KM6xwLlKLaNtSXtaPsHVdt2oOEH2NDwxTsk7QKsD5zYINQqVYdrieRbDcPeanvWRxBDQ5B/p3xIf4+hpF53CLKKuQrwDdvvHXW77WfXjbnUc0zIGPp2lGGRDYD3UP6hH7D94waxXkY5jHkw5Q27DvBvtj/TUVsfW7XvpJYnWnYFnkBJJCeP6lnUiHV/YJcq1vdsNx5iknQ25UPhhMEYq6RLbG9dM859gY9SzjeYMpTxuiYnfqfEXYUyVHLn3w74XNOTSiqzPXasLp5p+/w27atibkg5b3BRw8d/mPK6fd3QcNDalCHJmxqcz3gt5Wj3tyw+enOb8xmSHkYZ912/uuovwEtsn9cw3oMp/wdTjupq/+0kvZ3yenuN7Wuq6xZQOj3n2H5PjVj7zHT71CPaGnHPsP2YJo+dVfxJSOiTStJ6tv8qaaNRt7c9gdHGdG0aaNo2SWfbfsTwSTNJF9repmacH1N2tPpKddWewGvbzKyohpaOsv2CpjFGxNyWMsvlDkoiaZqQTqejWT2S7kKZJPAi4OeUJLcZcBRldkqtjoSkqyjnCv5Yty2ziL0eJY9c3yLG24DnUY6AoRwpHW37fQ1i7Qf8K3DX6qobKUNNjc6LdE3SO4B/AF+ltA3oLpfMaUKX9BHbr5P0TUaMg9Y5kyzpBba/JGnkG8j2hxq071u2nybp/6r2DY+vusksnCrus4H3U4aVVH3ZNaa4TdemoVhN2/Z14EOUE3PbA/sDC23vWTPO2VOTt6Qf296+SbuGYpwMPL3N0dFQrH8Ddqcc0YlyAuxrtv+9QazOZ/WoTMG7X9W2q+oMGUyJcxqwq+3bmrZlRMw1KENKC1hyOl/tMXRJlwMPG/x+ku4KLLL9gJpxBu99URK6gGsoR16NjgzV4Qy8Kt6odjR+v04112PogzPkh3YQazB+13S2yFJsP636fp+uYlY+QElKjYdGxtCmgVdShko2oZzLOJnFMxrqOE1l4/BjKB80zwVOHBxZtOiRXAOcJekEluzh1P7ABvYCHmr7JgBJh1BmHtVO6HQ4q0fSCyidrS9S5mUPrn85ZW77l2uGvBo4XdKJLDkO3ORvNvDfwPWU6YFtTxj+nCVz0WqUNtc16r3/MOCtkg62fUyDmF3OwAN4wOD1NlCN8XdizodcxnEY3RUtYzVdi8Pzs9xgzvmUGGNpW1em6YkMtDmCGDnzyXbtaaSSvg3sZfsv1eUNgC8NPshrxtodeAfwA9uvrs4hfNB27ROjKlNQH+Mpi6Wq4Y3TbD+sZrzO/mZDMWufVxkR48OUD/sFlGnBJ1eXn0D5Oz6/Tfyh59kIOMV27dWxkhbZfpiki20/qLruTNuPXtZjp4l33tR2jLquqbnuoWP7dknzVOaftjnJ+LFlPE/ts9LAf8wUEmi0sAg4V9JXKWOGwz2m46Z/yPJp2zQnMw+sZqnM2riOIAZJSGXevm3/rW6MoRkMNwOXSvpudXlX4AcN29XlrJ5VpybzKuZfq/H1um1rs2ZiOj+U9CDbFy/7rtO6pPp+KUvOaqk9GWImtv8kqWmvupMZeJLuSTnqXUvSQ1ncy1+PxeP9rc15Qq9cQ/vD6EVdN8r2Tl3HrKxHmQ71hOGnY/GKxWUaY9u+TDmZOZizuyflxGatk5lV4nkVMDijfzrwWdu3Tvug2cXdmjJUt1F1+Q/A3q63ucpgEdUiypTPgdMbtKfTFbaVu2jEOoDqQ2zWi8a6PEc1wo7Ai6ojsZtZfO5m1ucMbDedIlqLpJ2BPzd8+OsoCXd/ygy8nYAZZ8BM44mUk9ybUs5RDdwAvLVh25Yy50MuMJ5Dwi6prNBbiu0vLO+2DEja2fap1QnWpdTs7Q/H7eRkpqTPURZkDKZ3vRC43dWm4k2prL58m6u6JpIeB7zX9pwUcxrH9DZJb6RMQ33VlOl3nwROt/3BWcZ5mO1FKlNtR7Xt+3XbNhR7s2li/rxBrCdRkuWgXsrgw2HGmVwj4owqS7ARZc3I3rZrzb2vhoMPsd1oEdY0Mf/F9je6irdU/ElI6F1SWSn5ryx9Vrrp8MigFzawJuXNdp7t5zSM17rejKR3VTMpjhhxs+vEquIN3jxvpswpHj6ZuYZrzOGt4i011bHJ9Mdxxu16BkOXJL0SeAtlPropR66H2P70nDZsCpWyC8N/u180iHEV5WTyxQytcnYpoVAnztQPGQN/nHqkUzPmqcAu7ihRdjk7aJQ5HXKZ7lBwoOEh4dGUOZ5PpczY2IcyJ7gx268dvixpfRbP0Gmidb0Z24OjmpfVfeFPYxFLToN8xfDTUXpQddwuaXPb/wt3js130c6rq7m8g7//C4Cmi5W6nsHQGZeFcJ+RtA6l41W7muQ0Pdbh52gzpXI3ynmcewO/o/SuL6d0Uuq6FrjAdu2SFcOaHB3MwvnAf0v6GksOBzc6Aqbb2UFLmet56INDwWcD9wS+VF3eC7jGdu2xpaGz0nfOAZb0fdsjDzubqMaHL6o7T3bo8YM5y4OypnehrBatfRQh6RfASZQPsVO76km0pbKE+wjK9DNR3vAvdsMSsENxN6QURtuxinsGcLDt2mOkXc9g6IpKKd7PAptTeq0vcYMprtMNiwy0SYCSLqSceD+lei3vRJkxtG+DWA+nfLCezpKTBGac6LA8dHUEPBSv9eygmcxpD30whifpPV5yOew3JTWtNT446XadpKdSxs82bdHMqUcSq1JqXh/bImSX9Wa2BJ5OmSt+uKRvAce4qrHRRNWmqcMQtc4XuBRb26Jqn4Ar3EHFuipxNznROEpnNYRUilWdtazrZukTwBspH1a7AR+hHM3VMqYe68Cttv8oaRVJq9g+TaVaZRPvorwnNqBBYblxsv3ijkN2MTtoerbn/ItyqHbfocv3AS5vGOtplPoSWwOnUQ5tdmvZvscOfT0K2LRlvJcBG1bxrqYcsr6ig7/jhsAXKCcfm8Z4Z/V3+y2lh/0bygYBdeOsCbyeMnPnG5TZAmt28DsurGKeB1w0+GoYazvKGPWm1e96HLB9w1jnzea6JrFaxLkB+OvQ1w3D31v+H06p/nYfp8yC+ijww4axFrV9XYzrq3ptHF+9R39bvZYbv/8pddVvAa6sXrsXN339jvqaiJOi1Vnuw1i8OmwBJcHVKntbnZXe3/aHu20hSLoHJQEA/MT277p+jqaqoavnAk8GzgG+6oZn0qtx122A821vU/3en7P99JpxjqUkjuFhtA1t796kXUNxr6SUfp16Am2cvdGZ2vNIYAfKB9bw6249ymYUTU7WXk3poQ8cOnzZzcdvO6NSLOwmytHX8ymdqKPdoF6MSrnhk2yf2m0r26vWKHyZJc/ZPN/2rg3jdTY7aGT8SUjocOfZ30FR+8aH55JOc8dztCXtAXyQMsYnSjGnN9n+esN4dwMOpvT2DZwJvKfhm+H/KMWgjqVUSGxTwxxJP7H9cJW6zTtRkvIltmud7BrjLJcf2N5x2fecMcYJM93uejWEHgs8jnICfrii5w3AN23XrgE/zbjtUPPqj99qya32znDDSpDjoFKOen3K2ozBcKRdc9riOEi6wPZDlnVdjXjzR13vBrODRpmUhUVQai4soLRpG0m42TzvH0r6BEtXM2uzFP5twHaDXnk1NfIUoFFCp0wJPIPFKwmfT2nv4xvE2sb2Xxu2Y5RzVZbA/ydluOpvQJP9Sc+XtL2rEsiSHgE0GU+e6p3VHPep9anr9FofCfySMlRwNi1mtricB/q+pCO76mV5hnHb6oipFi291d7RkpputXcDQ0Xghm+iZoG5IRs3eMzy8geV2jqDqqF7AW2qVp7I4r/fmpTh5StpNjtoKRPRQ5f0RcoZ/QtYct++JkXkR82isNvNQ79zFkR1eRXgwuHrasZb5Cn1OCSda3thg1j/TNmX8B62t1apK72bG1QMHBF7AbBek96cSgW9LYFBz2M+5VzJHbSoxS3pS5QjuUtZsrZ3nTn8q1KW+e9FqZt/IvAV11ttOojVWW9/hudYn/Lh/zxKcadNaj7+IuCRXrK2+o+a/g/GQdJTGFpVbPukuWzPQNWj/gSlE2Dgh8ABnQ2RlJpMr7D9imXeeRYmpYe+kLLtVhefLi/1lLoj1RzoNk5SKds6+JR+LvA/LeKdJmlPFs+UeQ7NdmiB0pN+E2WaG7YvUtl2r1FCV9kG64Lqzb8jsK2kjzZ4AT+pyfPPwjZNP0gHXObtn0T5v65BSeynS3p3g15rZ739YSqlc3ejJPFtKZUEn0k5sqsdju622hu0b3PgWts3q6zWfTDwBVeFzmrG+n+U4cdBFck3q+yG1Hjnra5UQyGtP5RniH+eygY/nZiUHvrXKCczr+sg1qhqZkv1iBvEfTZDc59tH7+Mh4yKMXy4ujaL31irAH9rcrgq6Rzb22nJDSnajPFdRDkp+mDKiaDDgWe74Tx+dbCScEq8/wQ+bPuylnHWoCw+24sy1HcC8Hnbv6oZp7Pe/lDMoym91e9QhudOpdRDb1TwTKVO+D4srlvzTOBI2x9p0cYLKB2xBZQqiSdQNlN+SoNYF1HKGN9eXV6NMrNnzo4gNJ4aPYP/xcAqlA/ru9muPS11lEnpoW8MXCbpJyw5Llrn5NT9KeNQ62vJ+ibrMZRQmqrGaI+TtDENx9Bsd1arfcgfqt6SASQ9h7JBcVO32bakZwAftX24llGvZBR1u5Jw2I7APmpRFErSUZRprd8G3mX7kmU8ZFod9/YHtqYUk7qcMkHgdlUbijds44dUdlQadEhe7PZb7d1h+zZJzwI+YvvjKmV/m1qPxQW0xvE+qWtQwO1RlDUZX60u7067QoDDv9ttlA5AZ7VdJiWhH9xBjC0pc9A3oCy0GbiBckKoNpUVe4cAf6Isff8i5cNnFUl7Nx3nGx7WqE64bEt5UzTpve5HmfJ5f0m/oiyDb1Nb/gZJb6liPKbqgdYu2Ur5e23PlJWELdo10MVQzgspJ8z/GdhfiyurNjqxN6K3/zFqVM6cqpouen/KcMspkn4HrCvpnrZ/U6Nd2wEb2/52NSngvOr63VQWA7VJTLdK2ovS8x+835q8TqBs+HKeykbMoswa+rcWbWvNVVE1SS8CdnJVJVTSZyhHTk3jti7/vKwn6NUX5eRPV7HOpZS43Z3Se9i+uv7+lHnaTeNeRHnhblP9fABl/8k2bV0bWLeD3/melAVBj64uz6dUqqv9t6u+XwisUv38k47+L6tSev7zB19z+Ho7itJj+3dg6zE9x0JKydVfUGPxDmWa7YIR19+PUiaiTZu2onxw7VVdvg9wUIt4m1BKgDwb2GSu/p8j2nUlsNHQ5Q2BK1vE25pSH+bn1deiLl83kzKGvj1lxdkDKPWeV6VstdVkTPkDlDfXPyiHwttQdk//0owPHB3rzrFoSZd7qHbL8Jh1g7jn2d5WZU/LX7kMa9TatUTT7J064Hbbi7Um6RTKWO0hwN0owy7buWWZW41hB/uW7bmDxdNju5rGN91zibKT0azK3k6dnTXlttZrAro0ZZbL921/ey7bMyDpxZQRhMHsucdSagfVLotcxRtr+edJGXL5BGUjha9ReiN7A1s0jPUE22+uxvaupfSuT2PxisU6hutK/GPKbW0+CbsY1uh0nHHohO1SN9EsMe1GWUl4AOX3XI9Ss6OtAygn3zrfwb4J26t0HbP6oJ/JbOuYrzXDbWvPcNsyVcOGB7N0DfPaM8pGzHJ5k0odnEmY5XKEylaFgz0CDnKNYa8R1vZQgTrbp1fTSDsxKT30c20v1JIVEn/Y5FNL0qW2H1jNhviG7ZOa9kYk3U7pfYny5hjsui5KXZJGY4Yq21E9DzjH9pnVXNfHueZCKo2x1EFT03wwDAapbwL+l9JD+V7D+KfR8Q72k0bSG0ZcvTalhv7dbK8zyzifoZzAf7uH3uiS3gXtajFFAAAQU0lEQVTcyw0qIw7FuAI4kDJkcOeUyCYftJM4y2WYpE1Y/MEFgO1GxQMlHU85lzFcSmCh7We2bSdMTg/975JWBy6ohkyuo3kP4pvVi+0fwKtVVnXetIzHjGR71YZtWFbc3zC0DZXLydDaq2JdZj/sxpI1ROaUZ5jJU30AbU2pWd+0hOg4drCfKLbv3C+2Onl2AKVW+zHMvJfsVG8APgdcVU0zhDIEeS6lQFwb13c8LDJps1wAUKkg+VymLGSj2XoAgJdQjlQHJ83PoPxvOzEpPfTNKGOiq1M+9dcHPmX7qobxNqRUk7td0l0pqx3bHCZ1QlUdkhG92MbjrdXh6vp0W+pgrCS9wvZnGz72naOu94RsV9gVlR2kXk8pC3EUZQppo30xVRbWDaaLXuqaG35PE/MQyrmu41jyg7X2666a6fUeSjmHO2e52D66bTvbUikG92C3LP2sskvZurZ/P+X6e1A+HBt1Opd6nglJ6Hf3lOqFkra0fWWDWBO3/+c4aQylDlYEGrGJcl9I+iBltsdhwCfd9dS2DnT9uquGNR5BSeg/ds0FXuNSjZ/v3vZ/IOkwSkXJ46Zc/3xgR9uvahP/zngTktCvBN5h+9jq8hsoS/i3ahCr0/0/u6ZSB+Yij3HXkj5TKVd7OLCO7fkqVQRfYfvVc9y0zlQzZ26mLDwZ68yZSSDp85SKo2c2PSofF0nfoAxTTS0GV2ulqKTLpstng/N+rRpamZQx9McBh0naHbgHZYXcw5sEcvf7f3bK9h2SLpQ03x2VzFTZmWnqhtOdbDo7gQa795wAYPtCSY+Z+SErlnHMnOlaNVTwXuDetp8saSvKGpDDG4Q7hrKKdW9J/0Q50XqG7U921+LGTqi+2pqpdk5n/++JSOi2r5N0EmWX8zuAt3R4mPl3yorASXIv4FKVUgfD4961iwBVMxnuSqld/jlKoa8m5W5XGLZ/ObS6E7rZfLq3qlW6D6T09i9zy31dK0dSdnl6W3X5p5TzOLUTuu3vVOsWtqUcUe9HKac95wm96XzzEX4n6eG2l3hvVqt5W21iP2wiErrKriDXUWY+bAp8XtIZtt848yNHxup6/89x6PIE3g4uG01fZPtdkv6DFsvOVwC/lLQD4Gpm1P6UI7qYohqXPo4yy2sRpZe4RzVz41ktx6k3tn1stZ4Cl7oujT5YVSqZrk/ZbetMyorsX7doW2uSjrW9h8oOXkuNSzeYUvkm4FhJR7K4Fsxgzc2ebdo6bCISOuXEz39VP/+lesO+pWGsQ4d+vo3yIu6ihkhnhlf6qSr2NTxPuKbBgqe/S7o3Zd5xo6p8K4hXUvav3ISycOw7QG/Gzzv2CeDTto8cvrKaOPAp4BktYt+osvPWoCjc9sD1DWP9FHgoZTHhb4HfSPpj25klLZ1f9Z6fxeJdlBqz/RNJD6ccfbyouvpS4BFTJ4S0faKJ+KJM3H989fNatKhLAjyEUvDnGsoq0dfM9e9XtWt7Sn2N4ygv4EsomzD/DnhSw5jvoBQk+5cq1nWU7ezm/Pft+G837ca8wNPnun2T+MUMNUdmum2Wsbel7EB1ffX9p5TpfW1irg+8llLj5B9z/Lc7lLKZxZ+q9+x7KQXYNprLdi3ra1Jmubwc2Jfyx9pc0hbAZ2zvUiPGP1MOXQZbRH0VeKPtkZuyzgVJ5wJvpbxwDwOebPvHKpX1vuKGtWGG4q9BWcHatKc0saqZUE+0fc2U619MWQm5+Zw0bIJJusr2/UZcvwrw01G31Yy/GqXKqSgfEI16spJeSdnvdDtKh+QMyoyXxlUNu1IN6y2kbAT+yOrrL24wA295mJQhl/0os1rOBrD9M5WNEeq4gjL+9nRXU58kHdhpK9tbbfAiVamX/WMA21dMOclXSzVEtYDq/6nm+7FOsgOB70p6iquNl6vx2+dRCibF0r5ZlcB4nZfcfu7DtNtxi2rB3uuBzWy/XNIW1dqRbzUItyFlCOgc27e0adcYrEVZxbp+9fVr4OI5bdEMJiWh32z7lkFSqz756x46/Aulh35aNWPmGGaeKjQXOi/2pWn2Y6VBKYFJZvt/JN0MfFvSMylL17ejVB9stIJyJfBm4H3AzyX9nPK62Iyy8vStLWMfQTm598jq8rWU4nq1E7rt90naGnhJlQPOdIsdn7pQLQR6IGU/hbMpwy8fmvTX2qQMuXwA+AvljO9rKSe5LrP9thkfODrW2pSyrXsBO1NevMdPyOFb58W+VDZj7mo/1oknaUfgvyhvsD3c0ZLpPlPZn/R+lNfZVbb/voyHzCbmoKDe8NaHTYvg7Uc5Sh9MjHgGZaLEp9q2s6mqU7gx5TzXD4EfAZc0fZ9NmX23FHewmThMTkJfhVJJ7gmUF93JwOfaJqmqHsbuwHPd06Xw6nA/1kmmJfdjXYMy82CwJ6vds9WTXahqpMj2F6dc/3LKfgNfHv3IWcX+IWXO+Fkutf03p5wHqr0gUKXa4g6u1p5IWoeykcecVltUOVx4IGX8fAfKtOo/AT+yPbKm0AyxZhwW9Cxr3C/zeSYhoUd9Q5/461Jm9TTejzX6SWWPz8fYvmHK9esBp7nFxumSdgXeTtm56DuUeuYvsn16g1gXU0rI3lxdXoOy49XIzTmWN0mbUn6/HSjbXN7N9gYt4q1F2WWrdq2qZZnTMfTpJu1X3OTwbSVyAqVMwplTrn8sMBGFjWLOrTo1mQPY/qukpvt/DmJ8V9J5lKm4Ag6w/YeG4b4I/LiqmyLKkGlXKzQbkbQ/JYE/inI0eBZl2OXztDgpKunplCmRqwP3kfQQ4N1ddcDm+qTo00ZcJ8pq0bYnbfruGcBbbV80fKWkGylbtDWpqRH9cheNqEqpUmN99TaBh+rnDD4wtqpmV9WuE277AyrVGx9Nef+/0vY5bdrXgQXA14EDOx7OPJgyo+90ANsXSFrQVfA5Tei2fz74ufqkeh6wB2Xn+m/MVbtWEAumJnMA2+d2+QKJFdrhwNclvWowf796bXyS9h/4bxr6eU1KklpEmYjQxD+qL7N4ssCcsT3jnr0t3Gb7+jbTlGcy10MuoxYDyfZOc9muFcSaM9w2016SsZKwfaikvwHfr040mjLL6hDbn24Z++nDl1WqJH6gSSxJr6HMbDue0kM/VtKcznIZo0skPQ9YtVpAuT9lFk0n5vSkqErd5zMptc8Hi4GudoONZlc2kr4CnGr7P6dc/1LKRtnPnZuWxSSqErpGjal3FF+UOv+1T2RO6iyXcagWZL2NJWf0vaer6bdzPYa+IiwGmlSvA45X2fFkuHrb6pSCQrGSqwpmfZay8Oxiyn6WnVSmVNlIZtAbXIVSm+jCpuFYsgDWrfQ0D1RrAN7G4rLDnZqIaYuTvBho0qnUuh7sfnSp7VPnsj0xOaraQW+h1EbZDXiZ7Sd2FHs/SnlqKMOl19g+q2aM1VzK7r6Z8t4fnDd7FmVO+6HTP3rFVA0zv5GhUh0AXa2TmYiEPmxlWAwUsTxIOs/2ttNdbhjzLsAHKau6r6H0pO8OfNz2IZIeavv8uu1TKVU7mOVyxgTMchkLSRcCn6EcVd9ZP972omkfVCf+pCX0iOiGpKspvcGBQ4cve8qGxbOM+THKDlkHDsbjq4VKh1IS1JNsz6oe/3DZgJWFpEVtFnQtM34SekQ/STpihptt+yUNYl4FbDG1LIekVYE/UJWEnmWsa4EPzdDAaW9bUUk6mLL/wfEsubL7T13En+uTohExJrZfPN1tKps8N3HHqBpLtm+X9PvZJvPKqsA69PQE6DT2qb4Pz+M30MnMviT0iJWEpPUpM8ueR9lrd5MGYS6TtPfUevtVIbC6M2ius/3uBm1YYc12OKqpDLlE9FhVCGo3ShLfllLM7ZmUE493zPTYaeINNp7+B+XEnil16dei5sbTK+MYOiy9IQ3Q2YY0SegRPSXpaOAxlGqIxwCnUuqht+4lStqZUlpWlOmy32sQY6Ouxo5XFNNtSGN7/y7iZ8glor+2Bv5MGQq5ohrn7qQHV613aLXmYWVL5pWFjHFDmlXGETQi5l5VfnoPyp6Yp0g6E1hX0j3ntmUrtUuAsf39M+QSsZKQtJAylv4c4FrbO8xxk1Y6VZngsW1Ik4QesZKpCmk9pqttz2L2ptuKrqv/RcbQI3pK0r8t4y5J6MvZuD9EM4Ye0V83jviCsiH7m+eqUSszSdtLOkfS3yTdIul2SX/tLH6GXCL6r9p27gBKMj8W+A/bv5vbVq18qgqYewJfo8x42ZtSSqGTLTcz5BLRY1X10tcDz6eUpd7W9p/ntlUrN9tXSVrV9u3AEZI627EoCT2ipyR9EHg2cBjwoMGOQDGn/i5pdeACSR8ArgPW7ip4hlwieqra4vFm4DYW7y4EZXWnba83Jw1biUnaDPgtZWexA4H1gU8NtuBsHT8JPSKiHzLkEhGxnEh6FHAwsBlLFufqpHxueugREcuJpCsoQy1Tt6D7Yxfx00OPiFh+rrf97XEFTw89ImLMJA02596DslPTcSxZy+W8Tp4nCT0iYryqolzTse2dO3meJPSIiH5ILZeIiDGT9HpJLx1x/Wslva6z50kPPSJivCRdQim7cMuU69cAzrH94C6eJz30iIjx89RkXl15M2XlbieS0CMilgNJ95jNdW0koUdEjN8HgRMlPVbSutXX44BvAod29SQZQ4+IWA4kPRk4CNiaUiztUuCQLhcaJaFHRPREhlwiInoiCT0ioieS0CMieiIJPSJiOZF0D0mHS/p2dXmrUStIm0pCj4hYfo4ETgbuXV3+KdDZ0v8k9IiI5Wdj28cCdwDYvo2hjS7aSkKPiFh+bpR0N6pNuyVtD1zfVfDsWBQRsfy8HjgB2FzSWcA84DldBc/CooiI5UjSasCWlKJcV9q+tbPYSegREcuPpB2ABQyNkNj+QhexM+QSEbGcSPoisDlwAYtPhhroJKGnhx4RsZxIuhzYymNKvJnlEhGx/FwC3HNcwTPkEhExZpK+SRlaWRe4TNJPgJsHt9verYvnSUKPiBi/zjaxmEnG0CMilhNJ77f9r8u6rqmMoUdELD+7jrjuyV0Fz5BLRMSYSXoV8GrgvpIuGrppXeCszp4nQy4REeMlaX1gQ+B9lH1FB26w/afOnicJPSKiHzKGHhHRE0noERE9kYQevSdpA0mvbvjYBZKe13WbIsYhCT1WBhtQZhg0sQBIQo8VQhJ6rAwOoWwocIGkD0p6k6RzJF0k6V0AkrarLq8paW1Jl0raunrso6vHHjinv0XEMmSWS/SepAXAt2xvLekJlB1iXkHZYOAE4AO2z5D078CawFrAtbbfJ+lxwBttP21OGh9RQxYWxcrmCdXX+dXldYAtgDOAdwPnADcB+89J6yJaSEKPlY2A99n+7IjbNqIk+LtQeuo3Ls+GRbSVMfRYGdxAWWINcDLwEknrAEjaRNLdq9sOA94BHA28f8RjIyZaeujRe7b/KOksSZcA3wa+DPxIEsDfgBdIehJwm+0vS1oV+KGknYEzgdskXQgcafvDc/RrRCxTTopGRPREhlwiInoiCT0ioieS0CMieiIJPSKiJ5LQIyJ6Igk9IqInktAjInoiCT0ioif+P0A5FinmfT7WAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "% matplotlib inline\n",
    "grouped = locations.groupby('text')\n",
    "counts = grouped.size()\n",
    "counts.plot.bar()"
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}