RTI_exercise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bhpDsY45ODak"
   },
   "source": [
    "# fetch data, install packages, load packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "TGvEOmwjNJ8r",
    "outputId": "2f5360d0-5dfe-4498-e6dc-397fef90d8df"
   },
   "outputs": [],
   "source": [
    "# !git clone https://github.com/Jbollenbacher/data-scientist-exercise02.git\n",
    "# !pip install pandas sklearn umap-learn sentence-transformers matplotlib hdbscan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "S28KJCWfNysf"
   },
   "outputs": [],
   "source": [
    "import json\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "from collections import Counter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "fU1JzIeYN6WF"
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer #common bag of words models\n",
    "from sentence_transformers import SentenceTransformer # a deep-learning-based model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rpRVcyz0OJqg"
   },
   "source": [
    "# Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "DMA8BsBeN9Kl"
   },
   "outputs": [],
   "source": [
    "#load aviation data table\n",
    "\n",
    "av = pd.read_csv('./data-scientist-exercise02/data/AviationData.csv')\n",
    "\n",
    "for col in av.columns:  #convert all date columns to datetimes\n",
    "    if 'Date' in col: av[col] = pd.to_datetime(av[col])  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Xp8N4qFBON1E",
    "outputId": "87b25de1-0f13-45f8-b28a-00a0d95d0a40"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 76133 entries, 0 to 76132\n",
      "Data columns (total 3 columns):\n",
      " #   Column          Non-Null Count  Dtype \n",
      "---  ------          --------------  ----- \n",
      " 0   EventId         76133 non-null  object\n",
      " 1   narrative       76133 non-null  object\n",
      " 2   probable_cause  76133 non-null  object\n",
      "dtypes: object(3)\n",
      "memory usage: 1.7+ MB\n"
     ]
    }
   ],
   "source": [
    "# load narratives data\n",
    "\n",
    "files = os.listdir('./data-scientist-exercise02/data/') #get all files in data directory\n",
    "files = [f for f in files if 'NarrativeData' in f] #keep only narrative jsons\n",
    "\n",
    "narratives = list()\n",
    "for f in files: \n",
    "    narratives += json.load(open('./data-scientist-exercise02/data/'+f, 'rb'))['data'] #load each file, add all events to narratives list\n",
    "    \n",
    "narratives = pd.DataFrame(narratives) #convert list of dicts to pandas df\n",
    "narratives.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "CtczgzCFOkys"
   },
   "outputs": [],
   "source": [
    "#merge/join tables\n",
    "df = av.merge(right = narratives, on = 'EventId', how='inner') #add narratives to av data\n",
    "df = df[df['narrative']!=''] #drop null entries "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "MYhyIjB2OxzP"
   },
   "source": [
    "# initial text cleaning and inspection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "rj4rnuDNO-mX",
    "outputId": "4d681b62-90b4-4ae6-cc80-73d8f9de771c"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "#seperate first sentence from narrative.  it concerns only evidence sources, not probable causes\n",
    "df['evidence_statement'] = df['narrative'].apply(lambda x: x.split('.', 1)[0])\n",
    "df['narrative'] = df['narrative'].apply(lambda x: '.'.join(x.split('.')[1:]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 140
    },
    "id": "n7lzp4z5OupH",
    "outputId": "482c5cc0-cb31-4b4d-d706-d5e400aafab7"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.google.colaboratory.intrinsic+json": {
       "type": "string"
      },
      "text/plain": [
       "\"TWO AIRPLANES WERE CONVERGING HEAD-ON WHILE TAXIING ON THE SAME EAST-WEST TAXIWAY. THE PILOT OF THE CESSNA 182 WAS TAXIING EAST AND SAW THE TAILWHEEL EQUIPPED GREAT LAKES 2T-1AK (BIPLANE) 'S-TURNING' ON THE TAXIWAY. SHE ASSUMED THE BIPLANE PILOT SAW HER AIRCRAFT. AS THE TWO AIRPLANES GOT CLOSER, THE CESSNA PILOT SLOWED AND MOVED TO THE RIGHT SIDE OF THE TAXIWAY, SO THE AIRPLANES COULD PASS WING TIP TO WING TIP. THE WESTBOUND BIPLANE THEN 'S-TURNED' LEFT TOWARDS THE CESSNA. THE CESSNA PILOT IMMEDIATELY VEERED RIGHT AND TAXIED OFF THE PAVEMENT INTO THE DIRT, BUT THE AIRPLANES COLLIDED. AFTER THE COLLISION, THE WESTBOUND BIPLANE PILOT STATED TO AIRPORT OFFICIALS THAT HE NEVER SAW THE CESSNA. THE SUN WAS SETTING AT THE TIME OF THE COLLISION. \""
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#manually inspect a sample text. repeat as needed\n",
    "df['narrative'].sample(1).reset_index(drop=True).iloc[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uN7Mpr-CPICl"
   },
   "outputs": [],
   "source": [
    "#standardize common abbreviations, acronyms, names\n",
    "\n",
    "aliases = { # manually created a list of common abbreviations by inspecting the data \n",
    "        'federal aviation administration': 'faa', \n",
    "        'flt': 'flight',\n",
    "        'hrs':'hours',\n",
    "        'plt':'pilot',\n",
    "        '&':'and',\n",
    "        'acft':'aircraft',\n",
    "        ' rnw':' runway',\n",
    "        'feet':'ft',\n",
    "        'ldg':'landing',\n",
    "        'lndg': 'landing',\n",
    "        'dept':'department',\n",
    "        ' agl': ' above ground level',\n",
    "        ' abt':' about',\n",
    "        'aprx': 'approximately',\n",
    "        'Approx': 'approximately',\n",
    "         }\n",
    "\n",
    "def replace_aliases_in_text(text, aliases):\n",
    "    for k in aliases.keys():\n",
    "        text = text.replace(k, aliases[k])\n",
    "    return text\n",
    "\n",
    "df['narrative'] = df['narrative'].apply(lambda x: replace_aliases_in_text(x.lower(), aliases))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "6PNhjE24PK8o",
    "outputId": "b42a1c21-d2e8-40aa-cc72-583bb8aa29b9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 77026 entries, 1 to 77256\n",
      "Data columns (total 34 columns):\n",
      " #   Column                Non-Null Count  Dtype         \n",
      "---  ------                --------------  -----         \n",
      " 0   EventId               77026 non-null  object        \n",
      " 1   InvestigationType     77026 non-null  object        \n",
      " 2   AccidentNumber        77026 non-null  object        \n",
      " 3   EventDate             77023 non-null  datetime64[ns]\n",
      " 4   Location              76945 non-null  object        \n",
      " 5   Country               76517 non-null  object        \n",
      " 6   Latitude              23722 non-null  float64       \n",
      " 7   Longitude             23713 non-null  float64       \n",
      " 8   AirportCode           43300 non-null  object        \n",
      " 9   AirportName           46027 non-null  object        \n",
      " 10  InjurySeverity        77021 non-null  object        \n",
      " 11  AircraftDamage        74747 non-null  object        \n",
      " 12  AircraftCategory      16496 non-null  object        \n",
      " 13  RegistrationNumber    74271 non-null  object        \n",
      " 14  Make                  76938 non-null  object        \n",
      " 15  Model                 76909 non-null  object        \n",
      " 16  AmateurBuilt          76473 non-null  object        \n",
      " 17  NumberOfEngines       73211 non-null  float64       \n",
      " 18  EngineType            73921 non-null  object        \n",
      " 19  FARDescription        16640 non-null  object        \n",
      " 20  Schedule              11366 non-null  object        \n",
      " 21  PurposeOfFlight       73533 non-null  object        \n",
      " 22  AirCarrier            3812 non-null   object        \n",
      " 23  TotalFatalInjuries    55697 non-null  float64       \n",
      " 24  TotalSeriousInjuries  53658 non-null  float64       \n",
      " 25  TotalMinorInjuries    54641 non-null  float64       \n",
      " 26  TotalUninjured        65752 non-null  float64       \n",
      " 27  WeatherCondition      75140 non-null  object        \n",
      " 28  BroadPhaseOfFlight    70917 non-null  object        \n",
      " 29  ReportStatus          77026 non-null  object        \n",
      " 30  PublicationDate       63965 non-null  datetime64[ns]\n",
      " 31  narrative             77026 non-null  object        \n",
      " 32  probable_cause        77026 non-null  object        \n",
      " 33  evidence_statement    77026 non-null  object        \n",
      "dtypes: datetime64[ns](2), float64(7), object(25)\n",
      "memory usage: 20.6+ MB\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-aef30f16-3bf6-44df-845f-2a67d96160db\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>EventId</th>\n",
       "      <th>InvestigationType</th>\n",
       "      <th>AccidentNumber</th>\n",
       "      <th>EventDate</th>\n",
       "      <th>Location</th>\n",
       "      <th>Country</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "      <th>AirportCode</th>\n",
       "      <th>AirportName</th>\n",
       "      <th>InjurySeverity</th>\n",
       "      <th>AircraftDamage</th>\n",
       "      <th>AircraftCategory</th>\n",
       "      <th>RegistrationNumber</th>\n",
       "      <th>Make</th>\n",
       "      <th>Model</th>\n",
       "      <th>AmateurBuilt</th>\n",
       "      <th>NumberOfEngines</th>\n",
       "      <th>EngineType</th>\n",
       "      <th>FARDescription</th>\n",
       "      <th>Schedule</th>\n",
       "      <th>PurposeOfFlight</th>\n",
       "      <th>AirCarrier</th>\n",
       "      <th>TotalFatalInjuries</th>\n",
       "      <th>TotalSeriousInjuries</th>\n",
       "      <th>TotalMinorInjuries</th>\n",
       "      <th>TotalUninjured</th>\n",
       "      <th>WeatherCondition</th>\n",
       "      <th>BroadPhaseOfFlight</th>\n",
       "      <th>ReportStatus</th>\n",
       "      <th>PublicationDate</th>\n",
       "      <th>narrative</th>\n",
       "      <th>probable_cause</th>\n",
       "      <th>evidence_statement</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20150901X92332</td>\n",
       "      <td>Accident</td>\n",
       "      <td>CEN15LA392</td>\n",
       "      <td>2015-08-31</td>\n",
       "      <td>Houston, TX</td>\n",
       "      <td>United States</td>\n",
       "      <td>29.809444</td>\n",
       "      <td>-95.668889</td>\n",
       "      <td>IWS</td>\n",
       "      <td>WEST HOUSTON</td>\n",
       "      <td>Non-Fatal</td>\n",
       "      <td>Substantial</td>\n",
       "      <td>Airplane</td>\n",
       "      <td>N452CS</td>\n",
       "      <td>CESSNA</td>\n",
       "      <td>T240</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Part 91: General Aviation</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Instructional</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>VMC</td>\n",
       "      <td>LANDING</td>\n",
       "      <td>Preliminary</td>\n",
       "      <td>2015-09-04</td>\n",
       "      <td>the flight instructor and private pilot under...</td>\n",
       "      <td></td>\n",
       "      <td>On August 31, 2015, about 1230 central dayligh...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>20150831X30510</td>\n",
       "      <td>Accident</td>\n",
       "      <td>WPR15LA253B</td>\n",
       "      <td>2015-08-28</td>\n",
       "      <td>Las Vegas, NV</td>\n",
       "      <td>United States</td>\n",
       "      <td>36.099444</td>\n",
       "      <td>-115.162500</td>\n",
       "      <td>LAS</td>\n",
       "      <td>McCarran International</td>\n",
       "      <td>Non-Fatal</td>\n",
       "      <td>Minor</td>\n",
       "      <td>Airplane</td>\n",
       "      <td>N187AA</td>\n",
       "      <td>DASSAULT</td>\n",
       "      <td>FALCON2000</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Turbo Jet</td>\n",
       "      <td>Part 91: General Aviation</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Business</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>VMC</td>\n",
       "      <td>STANDING</td>\n",
       "      <td>Preliminary</td>\n",
       "      <td>2015-09-03</td>\n",
       "      <td>the c90 sustained substantial damage, and the...</td>\n",
       "      <td></td>\n",
       "      <td>On August 28, 2015, about 1010 Pacific dayligh...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-aef30f16-3bf6-44df-845f-2a67d96160db')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-aef30f16-3bf6-44df-845f-2a67d96160db button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-aef30f16-3bf6-44df-845f-2a67d96160db');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "          EventId  ...                                 evidence_statement\n",
       "1  20150901X92332  ...  On August 31, 2015, about 1230 central dayligh...\n",
       "5  20150831X30510  ...  On August 28, 2015, about 1010 Pacific dayligh...\n",
       "\n",
       "[2 rows x 34 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.info()\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "wKqWOEnVPTns"
   },
   "source": [
    "# explore some summary stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "RWnYDG-rPR0B"
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import ENGLISH_STOP_WORDS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "lm0FANJNPiF3"
   },
   "outputs": [],
   "source": [
    "#define a helper function\n",
    "def get_word_freq_dict(text):\n",
    "    wordlist = text.split()\n",
    "    wordfreq = Counter(wordlist)\n",
    "    return wordfreq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "WUKo9ha8PXW7",
    "outputId": "a854f459-890a-4b89-d831-b9041617e60f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean 125.02924669574843\n",
      "median 104.0\n",
      "fraction <500 0.9923823165570675\n",
      "fraction <200 0.8704342730646526\n"
     ]
    }
   ],
   "source": [
    "# Check wordcounts of texts\n",
    "\n",
    "narrative_wordcounts = df['narrative'].apply(lambda x: len(x.split(' ')))\n",
    "print('mean',narrative_wordcounts.mean())\n",
    "print('median',narrative_wordcounts.median())\n",
    "print('fraction <500', (narrative_wordcounts < 500).sum()/len(narrative_wordcounts))\n",
    "print('fraction <200', (narrative_wordcounts < 200).sum()/len(narrative_wordcounts))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oiNuBjI1PjuF"
   },
   "source": [
    "**Conclusion**: texts are almost all short \n",
    "\n",
    "This suggests transformer-based embeddings may be appropriate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 519
    },
    "id": "6TdrniZPPpNz",
    "outputId": "9788fe70-c533-4dd8-94d3-6fc6d3daec20"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 141970 entries, 0 to 141969\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count   Dtype \n",
      "---  ------  --------------   ----- \n",
      " 0   word    141970 non-null  object\n",
      " 1   count   141970 non-null  int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 2.2+ MB\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-b8b606ba-129e-4d81-97db-236adecbafe7\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>word</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>pilot</td>\n",
       "      <td>145235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>airplane</td>\n",
       "      <td>104788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>aircraft</td>\n",
       "      <td>61914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>engine</td>\n",
       "      <td>55174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>landing</td>\n",
       "      <td>52916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>flight</td>\n",
       "      <td>50984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>fuel</td>\n",
       "      <td>42399</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>left</td>\n",
       "      <td>41691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>reported</td>\n",
       "      <td>39592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>right</td>\n",
       "      <td>35235</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b8b606ba-129e-4d81-97db-236adecbafe7')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-b8b606ba-129e-4d81-97db-236adecbafe7 button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-b8b606ba-129e-4d81-97db-236adecbafe7');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "       word   count\n",
       "0     pilot  145235\n",
       "1  airplane  104788\n",
       "2  aircraft   61914\n",
       "3    engine   55174\n",
       "4   landing   52916\n",
       "5    flight   50984\n",
       "6      fuel   42399\n",
       "7      left   41691\n",
       "8  reported   39592\n",
       "9     right   35235"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#explore most common words\n",
    "\n",
    "all_texts = '\\n'.join(list(df['narrative'])) #combine all narratives to single string\n",
    "word_freqs = get_word_freq_dict(all_texts) # get word counts\n",
    "word_freqs = {k:word_freqs[k] for k in word_freqs.keys() if k not in ENGLISH_STOP_WORDS} #remove stopwords\n",
    "word_freqs = pd.DataFrame(list(zip(word_freqs.keys(),word_freqs.values())), columns = ['word','count']) #to dataframe\n",
    "word_freqs = word_freqs.sort_values(by='count', ascending=False).reset_index(drop=True) #show most common words\n",
    "word_freqs.info()\n",
    "word_freqs.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lqnl7RO9PyT_"
   },
   "source": [
    "**Conclusion**: vocab is very large\n",
    "\n",
    "This suggests many unique words, e.g. proper nouns, typos, nonstandard abreviations. Adds to noise in any analysis.\n",
    "\n",
    "**Conclusion**:  top ~100 words dont look super informative.\n",
    "\n",
    "This suggest low dimensional bag of words embeddings may fail"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pQuG5XnQP3KH"
   },
   "source": [
    "# produce text embeddings \n",
    "\n",
    "create BERT, TFIDF, and word-count embeddings for each narrative"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "1wDBKA2AQ9_x"
   },
   "outputs": [],
   "source": [
    "import torch #supports SentenceTransformer usage\n",
    "torch.cuda.is_available()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "tSZDBCcjP9On"
   },
   "outputs": [],
   "source": [
    "#make BERT embeddings. warning, very slow. 10min on GPU, much longer on CPU\n",
    "bert_model = SentenceTransformer('stsb-distilbert-base')\n",
    "bert_embs = bert_model.encode(list(df['narrative']), \n",
    "                                   show_progress_bar=True, \n",
    "                                   device = torch.device('cuda'))#delete this param if cuda unavailable\n",
    "df['bert_emb'] = list(bert_embs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "vudyFQpkcphN"
   },
   "outputs": [],
   "source": [
    "#make rawcount embeddings. \n",
    "rawcount_model = CountVectorizer(ngram_range=(1,1), stop_words = 'english', max_features=750) \n",
    "rawcount_embs = rawcount_model.fit_transform(list(df['narrative']))\n",
    "df['rawcount_emb'] = pd.Series(list(rawcount_embs)).apply(lambda x: x.toarray().flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "GL5-rk8PP7M5"
   },
   "outputs": [],
   "source": [
    "#make tfidf embeddings\n",
    "tfidf_model = TfidfVectorizer(ngram_range=(1,1), stop_words = 'english', max_features=750) \n",
    "tfidf_embs = tfidf_model.fit_transform(list(df['narrative']))\n",
    "df['tfidf_emb'] = pd.Series(list(tfidf_embs)).apply(lambda x: x.toarray().flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "u0UjRMFmQAHr"
   },
   "outputs": [],
   "source": [
    "df.to_parquet('./drive/MyDrive/df_with_embs.parquet')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TEJp0eFlcj5O"
   },
   "source": [
    "# project and cluster text embeddings"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "2p1l7WZI30w7"
   },
   "source": [
    "###initial projection and inspection of different embeddings\n",
    "\n",
    "project down to 2d using umap to visualize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Rowlfnewc68w"
   },
   "outputs": [],
   "source": [
    "import umap\n",
    "import matplotlib.pyplot as plt\n",
    "import hdbscan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Wy615n4Lf9jv"
   },
   "outputs": [],
   "source": [
    "df = pd.read_parquet('./drive/MyDrive/df_with_embs.parquet')\n",
    "df = df[~df['tfidf_emb'].isna()] #drops empty-string and whitespace narratives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "id": "qzU1-tK4P95B",
    "outputId": "32b2e5a4-f425-4587-95a4-806375c58470"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f8cc33dcc10>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# project and visualize bert embeddings\n",
    "umap_model = umap.UMAP(n_neighbors=7, n_components=2)\n",
    "bert_emb_umap_proj = umap_model.fit_transform(np.array(list(df['bert_emb'])))\n",
    "plt.scatter(bert_emb_umap_proj[:,0], bert_emb_umap_proj[:,1], alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "id": "Cxn0GgBU0vpJ",
    "outputId": "8be1afa8-18ef-4ca1-9273-c379a419d9eb"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb30d1ac150>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAD4CAYAAADxeG0DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXRc53nn+e9769atfUGhsBErSVCkuEmUKGqzLdmWZTm2bMtK247TsZO440zSTnoykzkdtzvdPZNOxuNJn5xMJ+2JE3tiJ3ESJY4sybEtybZkObZEiaIkbhIpECQIgNiBqkKtt+4yfxQJcQdAFFAo4PmcowPyonjvUyrgV2+9912U67oIIYSoX1qtCxBCCLE0EuRCCFHnJMiFEKLOSZALIUSdkyAXQog6p9fioslk0u3p6anFpYUQom69/PLLk67rNl16vCZB3tPTw4EDB2pxaSGEqFtKqYErHZeuFSGEqHMS5EIIUeckyIUQos5JkAshRJ2TIBdCiDpXk1ErQiyHyWyRgcks2ZJN2OehOxkmGfbXuiwhlp20yMWaMJktcngwhWVDY9iHZcPhwRST2WKtSxNi2UmQizVhYDJL0NAJ+SsfMkN+naChMzCZrXFlQiw/6VoRQP13S2RLNo1h30XHQn6dqWypRhUJsXKkRS7WRLdE2OchV7QuOpYrWoR9nhpVJMTKkRa5uGK3xPnj9dIq706GOTyYAir154oWedNiV2e8xpVdn3r/hCRWlrTIBdmSPRfe54X8OtmSXaOKFi8Z9rOrM47ugalsCd0DuzrjdRl+a+ETklhZ0iIXc90SF4Z5PXZLJMP+ugzuS62FT0hiZUmLXNCdDJM3rbk+5vPdEt3JcI0rW5/WwicksbIkyMWa6pZYC+TGrVgs6VoRwNrpllgL1tqNW7H8pEUuxCojn5DEYkmLXIhVSD4hicVYcItcKfVVpdS4UurIBcf+i1JqWCn16rn/fmZ5yhRCCHE1i+la+UvggSsc/yPXdW8+9993qlOWEGtX2XbIFEymcyaZgknZdmpdkqhzC+5acV33OaVUz/KVIsTalzct/vyHJ3jitdPMllwaAhoP3bqJX75nC16P3LIS16caPzmfVUodOtf10nC1BymlPqOUOqCUOjAxMVGFywpRX8q2w58/c4K/ev4UjuPSlYxQKjt85bk+vvbjvlqXJ+rYUoP8S8Bm4GZgBPhvV3ug67pfdl13r+u6e5uampZ4WSHqT8G0+Parp4n6IRmPopQiGY8QNuCbB/prXZ6oY0sKctd1x1zXtV3XdYA/B/ZVpywh1h7LgdmSS2MsMndMKUUiFiZdkFmb4votKciVUm0X/PUh4MjVHivEeqdrEPdrTKZmcV0XAMdxmExlSQRlJLC4fosZfvi3wPPAVqXUkFLq08AXlVKHlVKHgHcCv7VMdQpR9wKGzsO3bSJXhqnULLbtMJXKUrDhE3f21ro8UccWM2rl565w+CtVrEWINc3r0fjFd2xBU4p/eKmfszNZGoM6v3JnL5+4a3OtyxN1TD7PCbGCvB6NT997A5++94ZalyLWEBm4KoQQdU6CXAgh6pwEuRBC1DkJciGEqHMS5EIIUeckyIUQos5JkAshRJ2TIBdCiDonQS6EEHVOglwIIeqcBLkQQtQ5CXIhhKhzsmiWEOtM2Xb4r998jX86eJYcEAY+dkcnn//w7lqXJq6TtMiFWEfKtsPvffM1vnbwLBbQFtUpAV95YZDf/9ahWpcnrpMEuRDrSMG0ePTgWQJAMqajKUhGdQzgkRcGa12euE4S5EKsI5YDeSohfp6moDGqk61dWWKJJMiFWEd0DYLAZNqaO+a4MJWxCNeuLLFEEuRCrCMBQ+ehWzZQoBLmjguTGQsT+OgdnbUuT1wnGbUixDri9Wj87sM3oYB/OniWkXMt8U/KqJW6plzXXfGL7t271z1w4MCKX1cIIeqZUupl13X3XnpcWuRiVXv84ACPvDTAVK5MY8jLR2/r5oO3dNe6LCFWFQlysWo9fnCAP3jsCNkSeDxwZrzIiaEj2I7LQ3t7al3eqvLogdN8Y/8ppmaLhH0ePnTrJn7ujh6ChvyKrwdys1OsWv/j+0eYKYFXh0jIg88LqTL86fePUradWpe3ajx64DR//IPjTKTy2Hg4NV7mj75znC9+6xB505r/BKLuydu1WLX6pyGkQyjoQSlFKKhD3uJ0qjKxxRswlr2GN4an+Zc3p5jOmySCBm/b0si29sSyX3cx/v7F0yjLoojCwKU14Wdypsi3Do5w7+427t3WVusSxTKTIBerlhdAA6XU3DFXA4PKxJbl9sbwNP/vs2/SN5ZifNaiUID//j1ob/HSHAuzKR7ig3vauHlj8/IXcw1TuTKO5sVQLoan8iudbPAzOlNk/5sTEuTrgHStiFVra5vBjAnZQqV7IFu0SJVgc7OOvgI/uY+8MMiRs1OcTVmUijALpIFjY2XyRZPBdJ6v/rSfV0+NL38x19AY8jKdLqNrnrljmWyRkB+yJemCWg8kyMWq9Zvv2c6muEbJhJmshVmCrqjiV+/ZSmAFbuIdGp4iX3TJm5C+ZJRu31gOHRfHVfzg2MSy13ItH9vXg/LA5HQJx3XJZIsULehp87MhHqhpbWJlSNeKWLXu3d4OwGMHhxjJmjQGvLx/dxvv2d2J17P8bRBH85DKQfEK38uVYCpv09MUYDJvLnst1/LQ3h6Gpwv8f8/1Mz5dIuiHXRuDbEkmeOeNTTWtTawMCXKxqt27vX0u0Ffa9rYwB4euvJSUCRRKJfKmwcaG2q9S8tn7b+S+HS0X3Zjd2x3j1PgsX3mun1TJojPi52duamVvb2utyxVVJkEuxFV8/LZuHnlplCu1t13AdMpoyuXd21dHq3dbe2JuRE3etHjqtUG+/dpZ4mEfXY1BJtIlvvrT0wAS5muM9JELcRWW7XCtW4VdySifvL2n5qNWrmQmV+L5/iniYR+NET+6ptEU8+HXPTx1tLY3Z0X1SYtciCvoG03xjy8PkdCh6EDWqbTCPVSGP25sgn//vt2cHE/ze4+9Sipn0RL1c/+O5hUP9gN9ozx1dJyJnElTyOD+Hc0kYmHSBYvupre6fXRNIxbyMZGrbZ++qD5pkQtxBfv7pzG8Hjo3hIgHIKpBAAgBm5u93LihiUNnpviHlwdJF0zQXF4cGOc/fuswX/r+cYamV2abhgN9o3z9hTMUbYdNLWGKtsPXXzjDsbMzxAI6U7Nv3aq1HId0rkRTaPknUomVJS1yIa5gOm+SjPjoioXIFW38ARe/7iGdyxML+bi5M8H+/gl8GmiaxtBMiaDXQPOUeeH0BNGQwT1boSOxvDdCnzo6Tjxo0BTzA8x9PdA/yZ2bGvn2a2cBiIUMJtIlLMfm/h2rrytILI20yIW4gkTQQAFtDSFuaAoTDelk8nkMr8a7trdz88ZGprJlmqMhZmZN/B5FyK8TM3zkSjZRn5cjQ6llr3MiZ86F93lNMT/pgs39N3Xy8K0dKBfOTOVJBrz88l09cqNzDZIWuRBXcPumBE8cGiHi19jYEsXwemkOBbh/Rxu3bUoCLpGAl2zJpGi7hP0eHNclXy4T0jWiIS8jqfyy19kUMphIFy8K84l0kaaQQdDQed+ebt63R5b9XeskyEXdK9sOn/zyD3l+oDR37I5uH3/1mXdd98Sh3tY4D1LpK7fyJnu64ty+KUFva3zumm/fkuQ7R0Yo22UKJYVplymWXO6+IUkmVyYe8Fbj6V3T/Tua+foLZ4BKS3wiXSSVN/nkHV3Lfm2xeix4hyCl1FeBDwDjruvuPHcsAfw90AOcBj7quu7MfOeSHYJEtZRth1/48g94YeDykRh72jUe+fX3Ltss0HTB5HuHhnnmjTFOTeZIBDzc1NPElmSEom1zz9amZe8jhyuPWpHuk7XpajsELSbI3wFkga9fEORfBKZd1/2CUup3gAbXdf/9fOeSIBfVkimY7P7fn77q95/57XewMRlZtuvnTYuZXIkzU1n6x3PYrktzxMfOjviKhLhYX5a81Zvrus8ppXouOfwh4N5zf/4a8Cwwb5ALUS3zLWd7eiK7rEEeNHSChk57Q4g7e5ftMkJc01L7yFtc1x059+dRoOVqD1RKfQb4DEBXl/TfieqYbznbYnn1LeOaLpgcPTPF6+M5LNuiKxFhT3cDrbFgrUsTdapqNztd13WVUlftp3Fd98vAl6HStVKt615oaDrLkaEUqULlRpN8vF37AobO1gY4foU7My0aRAOey79RQ+mCyYt947w5kSMeMNB9XoamcmQKZe7Z1ixhLq7LUu8CjSml2gDOfa3ZIg5D01l+dHyCsgVt8SBlC350fGLFZtiJ2vB6NL7yK/fQeUn+NQKfencv0aBvVe3vOZLKM5IpEQ8YBP06hrcybb5oWpwYzdS6PFGnltoifxz4FPCFc18fW3JF1+nIUIqoz0s0VBnydf7rkaGUtMrXuI5EmL/4lbs5NJTm1MQspuXQ2xzmpu4EQcPLbLFMxO9dkTXM51MoO1i2TTT41jR5w6th2hqZomyULK7PgoNcKfW3VG5sJpVSQ8B/phLgjyilPg0MAB9djiIXIlUo0xa/uFm2UpMyRO21xYO0xkPkS2VcFB5NYTsuCvBo2opt1jyfgFdD93jIFy2C/sqvn1l2sC2HqF+mdYjrs5hRKz93lW+9u0q1LEk84CWTK8+1xIEVm5Qhai9g6MwWy5Qsl4ChYTsutuMS9ut4NIVpVe+2TNl2KJgWllO52Row9AW39tviQdqiWd6cyAGgezTSuRLRoMENrdGq1SjWlzXTBNjZEedHxyt7J0ZDlVDPlMrc07M6Fv1fy8q2Q6ZgksqbFCyHoK4RCxpEA8aKdWd4PRoRv5eCaVEs2xgejbC/ErC246JrlTqnskVOjKY5OpRhaGqWjGkRNLzE/V7iYT/JiEFbQ4BtrTGSYf9l1ynbDmcms5wczzCVLYHmoTPhY1tbA4ZHmzfcYwGDfb3NhIxzo1ZKJboaZdSKWJoFTwiqpuWaECSjVlZe2XaYzpX40fFR/vR7xxjIvfW9IPDLb+/mF+/pvWIoLlc9s8UyHk2b617JFssUzDKj6QJHBtP89OQ4hwcyzNgX/9sYcMcNDWxvT5AIe9nUHCYS8BH0KsJ+L2bZ4fmT4xw8M0U6UyZlmmTyJWZyFrGAh11dTdzVm2RnexzD6yHg1fF7tUW12IW4liXP7Kwmmdm5eOmCyUgqz2zRQlPQFPYTDxk1D4lMweTpo8N87h+PXXFLtPPed0OMz3/45mV9Yz3/yWAslWc0azKVLXDgxCQvD44zmQbTgYXcMdme1GiMhtiYjNKVDHJqbJaxVJHhbJ7+MZPS/Kdgcxw2tyXY1Bgm6vfSkwzS2RShKxEitgr66kV9WvLMTlE76YLJ3z7fz7cO9nNq0sUCYjrcdkOUPR2t3NgeYkM8TCxgLHsL8ML+Ydd1GJrO8aUfXDvEAb57Is30N17i//rYrfQ0VbcvOF0weWN4hjdGs0wXTSiD11C8dGqC1/oz5BwoL+J8xyYdmmZnOTk2SyQAqRykCiwowM87mYKTqWl8TLMhAptb49zZm2SmNcZNXQkJc1FVEuR14J9eGuCvfnKS4Qu6LaYtePJYhiePZegMQXNDgKDuwWfoRPwGnfEAWzridMT8NIR8NEcDBI2lvdwXdlso5TKWLvL6SIa+eZdJq9g/lOfvnx/gNx64cVG1XOnmIlQ+DQxNZ3ltMMNktoBP15lKF8gUSgxMFXhjpLCo8L3QRIlKcufme+S1lYBTs3BqNsWJsRR39bZQMi3u3b5BultE1UiQ14Gnjp5l6hqBMpiDwVzhsuMBBtjYonFLdxubGwM0hP1MzpaYzpUwdJ2mqJ/mqI+I3wuuw/nZ7AGvh4hfrwyTK5WZzpuULIdS2SZgeDDLNgOTOVKFMpPpy697Lc/0jXDHySS3bW5aUJhf+OZh6JU+7+lcCct2yZsWQzMliiUb3aMxmsozOJ0nmy/z+khh3k8JK+1MBs4cHOPA6XGOj8zywO62uWVxhVgKCfI6kCtaFOd/2GUKwLExh2Njw3PHGnSIBaFkg10GXYeQDyJBP5sawzRFgmSKZXQNTNvl7GwBLxqxoBeFg9/jxe/T0TTFWCbH8eH0omoaHi/zzZdPkS2VuXVjksZzN0HPj3rJlixcIOrTiZ2bNOOiKJUsbMfFdV1SxTK5Yhmf7sG0bTw6GJbGcKrARKbAmxPmorpSVlrftMs/vNRP/8Qsn3rbJrZ3JKR1LpZEgrwONDf4Yex6ovxyMxbMXDgT3KRyB3CmyCvDRQJAWwQcIG9DwANBv4/JWUXRscnOlinaYHgh4NNIZRY3/X0WePrIDK+dmaEtFqA5FiJs6DQEfbQ1BGkIGmg6nHWhXLbJlxwCfg9Rv5egz0O2aDGTK2PaFqAYTxU5m8oxnilw9Owsk9fbl7LCBmZh4PAEk7MlPvfBXfS2RCXMxXWTIK8DH9/bzfffWP79H10qmX5y9q1jIcBTKmGZlcyfm0RuArnrW8OkAAxkYCBTgMFK10yrD1ob/PS0RNkQCxH0KmZLFqPTeUzHwefTKZoWKMgVbTL5EtlikdQsTNnXvt5q9tzpDDxxmH/3nq3cukk2RRbXR4K8Dty3s4Ntidd4Y3rlr50DVqKzebQE46NFRlJFGoIeyrbNbAmyxcqIkwu7ShSVN5214rlTGdKPv8rvf+RWtrY3SMtcLJr8xNSJD912Q61LWHYOlR6k49M2J9MwXqx8Qri0v3sthfh5r42W+fIzJzg7k1tVqzWK+iBBXif2dcfYHKt1FStjLQb1Qjz7+jQ/OjbGWFoWehOLI0FeJ3qaozywS3ZWWssywI/eHOGl01OkC6tt8KRYzSTI60Q0YHD31maWb/dJsRrsfzPDP750mq/95CSHh2fIm7JGuZif3OysE16PxqbmKLu6g/x0QD56r1VZ4CensrxwKstjL/dz5+Y2Ht7byc7ORrkJKq5KfjLqSGPYxzu2baBBllhf82zg5Aw8c2yEv/7pKY4OTctNUHFVEuR1xOvRuLmzgV1dYVStixErYjgPzx2f4NGXhxhb5HIIYv2QIK8znY1h7t7aRmeo1pWIlTJRgq+9OMynv/IjHj04QKZgSutcXESCvM5E/Dp3bWqiu1m2BVtvjk+5/PGTR3ju+CizxbKEuZgjQV5nAoZOc9TPja0yfmU9OpOGRw6cxnGhICNaxDkS5HXG69FIhHz0NMkWduuRA4xM5ylbDpY0yMU5EuR1yOvR2FjlXXZE/fDpOqbtoMtvrzhHfhTqVCSgI/c716c93Y2AO7dTkhAS5HWqIeSnKykv33r0/t1tNEX8q3KC0PkNsKdzpoyuWUHyll6nIn6dD96yidef6qt1KWIF3ZDwcOum5lUX4pPZIi+dnOSR/ad5pv/iXaN2tsAffvROtrUnalTd2re6fhrEggUMnft3tNImm7GvGw0avGdHe1VD/EDfKH/w2CF+6xsH+IPHDnGgb3TR55jMFnnujTH+/oWTl4U4wJEx+PWvPc8bwzVYUH+dkCCvU16PRlcywi+8fWOtSxEroM2AfVsSPHRrZ9XOeaBvlK+/cIai7bCpJUzRdvj6C2cWHeYDk1kmc0VeH81e9TH9GXj6yOLfJMTCSNdKHfN6ND54azfHxjJ8+8hUrcsRy6AjAje0xulIBPnIze30tsardu6njo4TDxo0xSobYJ//+tTRcfb2ti74PNmSjWVDYZ79Us+kq7PvrLicBHmda4kF+MgtXYxM53j5rPyirCUJHW7f2MzP372JDfEgQcNT1fNP5Ew2tVw8H6Ep5qd/7Oot6ysJ+zzoHtB1rrktYEiXFYKWi3St1DmvR+PWjUnevbOj1qWIKtO0yn/TORNNU1WfANQUMpi4pJU8kS7SFFrcjZfuZJhkyE9vk/+qj4kDt21pup4yxQJIkK8BsYDBB29up9lX60pENU2a8GL/OP/44mmeOjxMJj9P38Ui3b+jmVTenAvziXSRVN7k/h3NizpPMuznHdta+MTdvdzcevmbQAPwm+/fxtt6F3desXDKdVd+h8S9e/e6Bw4cWPHrrnVfee5Nfu87J2pdhqiyZh90NofY1hblnt5mbtmUJBm+eut3MQ70jfLU0XEmciZNIYP7dzQvqn/8UumCyZHBKV49k2Y6Z9IQNLi5O8auzkZiARlitVRKqZdd19172XEJ8rUjXTD5n77+E54/JTsIrTU60BSEhpBOUzREb1OUvZsS7OqI0xILrrpx5WJ5SJCvE0PTWb74z4d4/OhMrUsRy6zJgLu3JvjZ23q4paeJoEzZX/MkyNeRdMHkp2+O8yffP8TR8ZV/fatNAfX/LFZGZxjes72Dh/Z1sq0tTt602P/mGM+8PsFgJkdY19m3qZH7d22gIyEraNabqwW5vIWvQbGAwX07NhDxe/jqv5zkxyfSlGtd1BJIiC/cYBa+e2SItFniI7d0Mpoq8kzfCJmsS8SvUy67/PTEOOlimX91W7eE+RohQb5GeT0at/Q04bjQGRvlmdfPcmZxw4NFnUrnYWgqxwt9U+RNh1IR4iEvAUOnbDjkzTLDM3mODKUkyNeIqgS5Uuo0MEtl82/rSk1/sfKChs7ejU00RwLs3Zzklf4pHjswzJQsSLem2UDZ1ZgpWNiui0fT5pa89Xo0fLqOiyJVqOfPaeJC1WyRv9N13ckqnk9UQdDQ2dwSJR4yiAe8RIM6jx4cYCBT68rEcjEBXdk0BHTypsOk41AwrUqL3HYoWRZeQyce8Na6VFEl0rWyDng9Gq2xINGAQTzoIx7y8dirp3ll+BrzqUXdCgGt0RB7NyWYzJQYy2dJZcuUbZty2QXHob0lys6O6q3bImqrWkHuAk8ppVzgz1zX/fKlD1BKfQb4DEBXV9eiLzCZLTIwmSVbsgn7PJVpwVWaFLFeBA2dbRtiNIQMGsMGT7x2hhePZ0jJ3cQ15V07m/j527vpbYuhezQifv3iUSu9zTJqZY2pyvBDpVS767rDSqlm4GngN1zXfe5qj1/s8MPJbJHDgymChk7Ir5MrWuRNi12dcQnzJZjMFvmLZ/r4zqEBzszWuhqxVPduifBv79lG74Y4Xo+GaTkkFrluilg+faMp9vdPM503SQQNbt+UWPRqllcbfliV6WCu6w6f+zoOPArsq8Z5zxuYzM6FOEDIrxM0dAYmZRjGUiTDfj5+Rzfv2dHOxkhlvLaoP41e+NTeDfy7d944F+K248rmzKtI32iKJw6NULYcOhtDlC2HJw6N0Deaqsr5l9y1opQKAZrrurPn/nw/8H8subILZEs2jeGLV4QK+XWmstVdRGg9ak+EuOfGFoplB/fkKJm8S8oEGdiyuikgqsHGVoOP39HLnZuSxEN+PJrCdlxsxyHil5uZq8X+/mka/F7ikUqOnf+6v3+6KmvMV6OPvAV4VCl1/nzfcF33e1U475ywz0OuaM21yAFyRYuwr7rrM69HXo/G9vYGBqcKTOVKDE3P4smUmZClzVetVj+8a2crt3Q1ckdvko5EmLJdGZliWpWWeMTvlfVXVpHpvElnY+iiY/GIj8GpXFXOv+Qgd123H7ipCrVcVXcyzOHBykeQS/vIxcVOT2Q4ODDDTL6y8twt3Q30NEWv+W+SYT/v3tGKaTm4CmZyS9tbMaIg5K/sGOM6lQkGcj916Qzg7t4Iv3nfdnZ2Ji4Kaq9HwyurC65aiaBBarY01xIHSM2WSASr85rVxVt2MuxnV2cc3QNT2RK6B7nReQWnJzI8dXSMsuXSnghRtlyeOjrG6Yn5B423xoJ85LYuPnnnRm7f3MzWxut/jw8YgAtFBzJIiF+vCOABgsCuVp1furuLz39gN3t6ktLarjO3b0owUyyTmq10B6dmS8wUy9y+KVGV89fNOPJk2C/BPY+DAzPE/F5i4cq7/PmvBwdm5m2VQ2WNln2bm/FoGj/tC3Hm2VMUFlmDRmVmoWmC3MFYvLgGGxo0upoaaAh72ZgM05OMcP/ODbUuTSxBb2ucB6n0iQ9O5UgEDR7c3Va1PVjrJsjFlY2m85wYzZApWrx0aoKbuy/eTisWNhieXng/nNejsbsrQcDQGZyY5btHJ1lod7kCEl6IBqBfUvyqAoBPA68OJbOy12VTxEPA8HBjWwN3bWsh4PVSdhzCXg9Bv9wLWgt6W+NV3Tz7QhLkdWw0nefF/inChpfmqJ+AT+eVU5Ps2ZgkFqyMWEhnK33lFzp/Y8xyQNcgYOgXfVQPGjpbWqN88p5e/D6dV0+PMzztcLUOmhDQHIOuZASUxsh0lkq7XJynASEF0SA0xwzaGkI0BHw0hnzs7IjT0RgmVzQ5MZbFsUHzKcIeD5br0p28vok7edNiPFMgU7TwaIrGkEFj2C/dMmuQBHkdOzGaIWx4CQcrL+PNnQmeOzHOG8Mpbt/SRDprki6WuX9Hy9y/KdsOs8UyHk3D0CtD1WaL5ctGOQQNna2tMd5/8waaY0GGp/N4PC5h3WDWLJIpQ7FUImR4Cfi8RAyNTMnCsW32T0uIXyhE5dNKYxTifgOldMySw/ZNMd57U8dFXYYbm6NzM5i9Xo2IgmeOnuXEWAGXMhtiYTa3BAn4DHRNkQgZNEcDl20qkS6YHB2a5vRknnShjGm5RAM6uzribG+PyyYUa4y8mnUsU7Rojr4VAq3xIO+4oZmXTk0zPJ2jIWhw/46Wuf7xdMHkQP8Ep6by4EJHIsCODXGiQR8F07po1EPZduifmGUsXcanayQjBhqKgF9nSzRKa9SPx6Momg6mbXN4MIXfVnj93gV3xawX5zu2smkgbQImGnBsJE3RdvnZ27rn9rMMGjqtsQC5ks14Os+LI2mGpouE/TrFsuLQ8AxvjqV4+9Y2mmN+ZnJlimWHrsbQXDiXbYc3xzL0j+fpn8xwcizDmclZSg40hr189LbNPHxbt4T5GiKvZB2L+nWyeWuuRQ4QNgzes6OVd2y9eAPdF06M8H8/fYwjg5WYjfhge2eUfRub2dvdQMivsyEeJGDoFEyL46Npnj02znAqh6MUPk2jZNuE/V78Xo3OeBBXQSKkM5Vz0ZWHeEgjlZelURfCAc7m4Pe+c4JX+if41XdvoysZZnK2hKF7cFyXvq0WJHgAAB0USURBVIkchwfTRAM6Dh50j4ahazi4nJnK054I4rgurgszudJcME9lixwfyXBwcJJjg9MMTNlzN63PZsscfeINTo7P8lsPbJcNkdcICfI6dkNrlBf7pwAIByuhnjXL7OtovOhxR85M8h//6SB9F8wGLpXgub4MJ0cyjM+0cffWZqbyJh7lUjJdDpye5sXTY5xNldBdMAxFMurHtH2MTntpayjTGDQo2zCayjNTKGE5DuOZIgaVpVTFwnz7jRlOjD7PXdvauakzwcZkkIlsmclsEct1QCnyJQcXUCi8uodcuYxHU7i2i0cpSpZL2XYYyxR4ZWCK/vEsJ8ezF4X4eQ7wl/uHGc8U+bV33cDmluiKtM7P35splh3Kto1luxTKFhPpAmemC9iOQ3PUX7lnIAt6LYoE+SrxwokRPv9PBzl5Qdju6dD5n9+9m60b4jSGfZfdpGqNBdm3qdJXPp4pEvXr7OtopDUWvOhxX36276IQv9BwDn7SP0JHUxgPGqcnUhwdyTAyXWCiUBkDHtGgZLuglXAdGPV48NBIY9jHTL5E3rQply3OpgpM5YrEAzC+2HGL69yJFJx4YZiGF4fZ2O7nbZvbCOg6hqaRKpRojgYpmTalsoWNwqP8pHImSilc18VrK06MWvSNpsmZDtmSyWiqcM3ho995fYq+kef5zx+5hT3LvHlzumDy6sAkf/OTUzzzZuqaWw++d2uc3/3QTRLmiyBBvgq8cGKE3/qrg4xc8tP9ypDFp752kDt7gnz67hvY2dVw2aiD1ljwsuC+ULpg8kLf1DWv3z8N//DCSaJBHyPTBUYKF0/iSTkQNsF1HSI+G9OysV2XM5NZ0DTKZZupvMl0vkAqX2ZWQvy6zTgwM1jk9cFTvHdXE15dkc1ZFM0ytu1SsGyCmkZb3MdsqUyxbFMybVrjfizbZSJXYjJrcvD0BJP5+a93IgVf+O6r/MFH9rGrs3H+f3Ad8qbFy6cn+et/OcUPT86/SNSTx1MEv3uEP/r5O5alnrVIgnwV+PZrI5eF+IWeP51nOv0qn33PLnqawgQMD5rmwacrGkK+a7ak+idmmZ2nn8MF3pxxYObqCZwF/GWYyZpE/X6ODM6g6RptMR9npnNMpHKMp8tkS8jNziooAD86OsHbbmxkYzLCZNbE51VsaYqyZUO0MvHKdXAdF4+mODyUAhSjM3leOTPF+LSz4AGgh0ccXhlILVuQz+RKDEwVee3Mwlf6e+bItRsf4mIS5KvA+Oz8s2eOz8D3jgzzvl0dBHxeWuN+imVFtmjRecGIhQuVbYfTk1liQSgsoHU2Hw9gO2A5FofPTqPQ+EmxRCpbYiZb+Z6EePXMOHA2naejIcjD+7q4oSWKQmE7Di6KqWyR189mKuPF00VM2+bI0AxDE+aiZ+QeG0kvy3MAKFkujmNTtirj6ReysqYp6zosiswMWAUaFrh34g+PTZPKV25y2bZb2TzAdpnIXPnXNlMwKZQstnfEqlJnDrBsyBctsnmLMxMZjp8tMZqBnPPWMLv1ykf1f6FG0wVOjuc4NTaLR1P4vRolq9IKH00XmMwWGcsUUUpRKDvkS2UKcM0+6CsxTavKlb/FpytCPi+6Z+HLIzdK9/iiSJCvAvfvap3/QVQ+bo+lCgQMD865j9R+r4d08cofolN5k5lCmWyhevPlJ0pweNzktdEiZ2YrLfASMo9zQxgSQYgZlUWuqrVJh+tASzSIaTsMTedxgWypzNlUnuGpPMVSZcXKQtliLFPCcd3r+qW2cZnMLs/nqYaQj42NAW7duPCum19825ZlqWWtkiBfBd5x4wa6Frge2FAqj64pNK0SFY7jomuXfw4t2w4v9U3w9z89yYuD0uGxnOIaeDWwXSibkKd6Kz4aHoXfp+EzdDLFMqmcSTLso2w7eLwaWdOkXHbImTaWbVGyrm+xst7WKIcHU8sS5kFDZ1dXI59992Y+cGOSwDyP/90HtvDpe2+oeh1rmfSRrwJej8b3/sN7ecd/epLJazxOAzQFluMS8XsoWw6FskVr7PJfjdOTs/zt/j7OyG54y27WATNTCfBqi4R8FMsOhlKk8yazIR/RoEEy7MdpqrTSp3MFymULv8dDQ8igZJrMLKKnZG9XkD1dCbyah4HJ7LKsMho0dG7qaeJPeiqLus233o9YHPk/t0oEDZ3nf/99/OqdnVd9zJ72AJuaogQNDdN2cFyHpoj/sm3woDK2/NT48vV7irfYLE+IAzhlh7Jt4SoXn+5B4ZItltE1aGsIsbUtiqY0bKUI+Dy0RAK0JHw0LHCXt+4YfPZdNxIP+gj5dbKllekk83o0ogGDRMggGjAkxJdIWuSriNej8bkP7eahvR184Z+P8FL/LA4QN2DThiB3bm7l/Te3kQj7523J5ExH7vyvAV6vh7aon86GMJpW+TTmuJU+eE1BWyxAW9SPcjXQXLIFk1BJx6eXFnTH8/MP3kwyUmmBy/aJ9UuCfBXa1p7gDz+xjxdPjnNsOIuFw+bGEHduaVrwbLdk2MCnoChhviwWOoxuKRTQ2RimpyVGMmLg8+rkTZuiaRH0eQkYHuIBL63xEDgK03V5YyjFqdECCxlM2OyHkK+y1opsn1jfJMhXqWTYz8/c1MXPXOduqNs3xNja5ePFgZXb4SEAixq/7KF+R7sEWP7hls0+2JgM050IEA/5sWwHTdlM5kzirovXY9AQMrihOUzf6Cw/eXOU4SlzwTdad3RFKdk2Y2mLeNAr2yfWMQnyNao1FuQTt2/h1OARJpa76Qg0KNjRHeHg6dkF9Re3BSGXr8wYXYHyqqpBZ1E3E69Xc4NB0KsxkzeJ+A0KZZtMvsR4psizr6f5wdFBzqaufxJWPmexu6MBw6OIyiqIdU2CfA1riviJRmBi+Sbt4Qe2tfnpbYlx79YWUvmjHBm/dju7KwzdTWEGprLk5t8XelXxAuUVCPGkH3qbokQCBuOzJfJFC8txmc6bHByY5sfHJkgt8ePM/uE8X/+Xk/z8XZuIzjcmUKxqEuRr2GimyJllDPEGDd65q4lbuhtpjvjoaYrw9hs6mMwMMHqVZuKv3dVFWyLI2UwRn+FFMcN4ZnFdMrVUZvGzJhcjBPgNaAp78eo6w9N5LNsFTeH1ahwfmeEnx1NV+//1/zx7irjfwy/fu7VKZxS1IEG+ivWNptjfP8103iQRNLh9U2JRm7dOz5aWNXTefmOSe7a3sSHip7MxRGPYz8N7O7BdeOboAG+eWyMpBPzruzr57ffvZLZo8fLAFJlCmYhfpyURBpVlKl3pZlnP4lplT0/lgXjETzpb4my6QMDQMC2HmUyOoRmr6m96j74yIEFe5yTIV6m+0RRPHBqhwe+lszFEarbEE4dGeBAWHOZu1eYXXi7pAb/PQ9m0aYr455bS7W6K8vCt7ezpipMv2wS9HjY1hdjUEsPr0dA1aI740TZQGUbnKnBcQkYR03YZmnbW3cJbfio3fsMB0DSwLZjJFEh5FK7j4tN1SqbJeNZhehnemWdkV6e6J0G+Su3vn6bB7yUeqUz2Of91f//0goNcU9Va8eNiPqC9OYDpOORLNtnyW521Xo/GppYYbQ2hK451Dxg6YZ8HN+xjd3ec7mQIy3HQPYpsweLESIYnDw9ftWtmLfECiQAEDIXCJah7SJdsvF5wcHDLitmSjWlbWGXILNNd4Yjs3Vn35BVcpabzJp2NoYuOxSM+BqcWMehtmYK8p9FLS9SPWXYZnMmxx2m46Ptej3bRRs6Xfi8Z8ZMtzlI0bZSC5Lk3qVPjGd6cShPya0SKDrPLUn1t+aiMD/cCHg00B6bTLiZQwsYLtETAchwsG1L55V8a+P27rz6bWNQHCfJVKhE0SM2W5lriAKnZEongwoeJGVr1pz23BaAxGsTv81I2LfJmmYh/gfPBqayxYTsu7YkwPq9O2XYYzRQ4OpTie0dGMJSiKR4mGigxNFXCsSrrctfbEMWrOT+qvwjgQOqSYf4mMDoLYU9lyeDlDvF/c9sGfuEdvct8FbHcJMhXqds3JXji0AhQaYmnZkvMFMs8uLttwedojfvZntQ4NlmdGNSAaMCD36NhlcuYDnTEQzQvYuxawbTwaBqGpogFDGayRYan8nzn4BkKZRvL8KCXvTRGgugeHQ9w7/Y2Dg9Ps//U9IK2L6t3ZaBoL+9InqQX/uvHbubOzU3EZAx53ZMgX6V6W+M8SKVPfHAqRyJo8ODutkWNWtnd2cA927tIv3ia4So07TbFYWNTDI8GNoobWsO8c1vLojbttRxQyiVfsBgYT/PPr56lfyrLYMom5AfTtLEpgusQC3qZLdnc0dvEu7a3MjyT4ycnJnnswCATa/z+3HIPx/zCx27ivp3ty3wVsVIkyFex3tb4ooL7Uq2xIJ+6eyMNQS9Pv36Wkckclg2GF3wKYkE/ubLNG1Pla45vCQEP3dpGSyyA5VbWyO5KBNje3kBXcnFbubiuQ6Zgk8kV+esXTvPDY1NzM0ELxUofsgeLgq5BQZHw6/h0jcawj8awj+5khHfe2MRXf9zHD96ss9lEq8Qd3X7u29lR6zJEFUmQr3GtsSCfufcGfuntvXPrP7uuU5nuXTAplV36JzI88lI/+wcubwfe1R3kF9/Wy46ORhpCBpbtLHkN6aJZ5kvPvMF337h8M94SMJUDXBPTKnPPlk6iAe/cUqfRgEEy7ON/ud9Ha/g0j74ysmxLyK5VN3W01LoEUWUS5OvEFUeSNFRGxezpTvDw3h7SBZMz0zkyeROUIhE0aAz78Xu1qi38r5TGj14fvWKIn1eiskTr3s4kt21uoSHkm7v2+eexozPBb70/yO7uBE8eOcOZySxl1yWVgYYwNMZD2LbFbK7Eyeuc3doWgERYZzRjMVsCCzCAiAHJmAe/7uG1EbPubsQmQvJrv9bIKyrmxAIGu9qvfOOrbDtkCiaWA5lCibF0gaLlEvZ56E6GF7xqXt4s8ydPn5z3cXdvaebdO9vpagxh2VeOymTYzwM3baCnKcRIKofHo/PU4WF8hofWaBClKaZni4QGxzk0tvAFUjZGoac5SlMogOaB9zdU3jB6khGCPi+mZTEwmeOHR0c5PXFmRRbQqpY7O4Nsa6/OZtxi9ZAgF/Mq2w7TuRJly2F8tsDTR0cZncmjeTSiPp1NrTHet6ttQWF+aGBqQePDi7aDplxCfi/WNZq8sYDBLT2NZAoRciWLk2NZMoUiugLDq9EU83F3qI19mzVG01mePDp1zWULPnf/Fj6wp53maOAan0AMgoaObdnsHxhnphp3klfA7Z1+PnrHRna0N8z/YFFXJMjFvDIFk4JpUyrb/M3z/fzkzQm8GjRG/XREAswUykT8Gh/e0z3vuR4/OLiga2aLNn5DR1MKpa691IDXo9EY9tMYhrdtSXD4bIZC0cZxXQzlIRYx2LUhyvaObfzrgQk+/ZWDV1xLvCcEezc2zBPiFQFDp7khyIO7u3h1+MSCnlOt7G41+OCeLm7cEGdrW0zWHF+DJMjFFU1miwxMZsmWbLJFk5aon2eOnuVbr05gUxldMj1bJBMtsqMjwSv9UwsK8idPXL1v/EKzpRKFooXtOIuacLSlrbLGS7bg4CgXzVWEAxpb2uIEDZ07trTxzd+4k89+/Xn6Lug77w3Df3j4Zm5sb1jQvQCvR2NDPMhdW5p594kRfnBydc5DjSj44sf2sa1NulPWMglycZnJbJHDgymChk5j2Ec6b/L04REef20Im8r0cqjskDOWgcBkiqB//pEQZdtZ8DJeRdPGdFwifu+ibrLGAgY3dTUykspTKDsEvBpt8eBFk162tSf4619/J4cHZxjJFDB0na3NIbZuaFjUmPigodPbEuE3H9jJ6D8+z9GxBf/Ta/JR+cWsxg5E7Umdpsjlm3OLt1zYaFnsPZ/VoipBrpR6APhjKou4/YXrul+oxnlFbQxMZgkaOiF/5cejMeKjbyJLOvfWD4yuQLmV3eOHJhwevDl01fOdlymYC64hFgrQGL6+3dVjAWPe2YqtseDcio1502ImV2I0XcSnKxpCvgUHutejsTEZ5n973x6++dIZXuybYmwJu+s1GfDh2zrIFWyOj6UYnCpQLFZGzBRg0etZfmBnp+z+cw2XNlpyRYvDg6m62/ZuyUGulPIAfwq8BxgCXlJKPe667rGlnlvURrZk0xh+qxUXMnS8WmV/zZAXSmWw3ErLvAi4Gty3ff4Wea60sOEdGtAcrowXX2550+LURJYnXzvLD48PMVsoEw8a/MJdvfzsvo0LOkfA0NnSEuMDt3Txtq2tHBpK8TcvDi+qjiBw59YEn7yzm96WGLrHQ8CrYdoOk7NFZnImk9kib47O8uKpSd4czDI1z7jHd24O8fG7N1Vl2OhadWmj5fzXgcns+gpyYB/Q57puP4BS6u+ADwES5HUq7POQK1pzP9S6R6Mh5Cfug7JbWTfbKkPBqoyr3rcptqAZqKa9sPbkhjDcc2MrGxrmb+Uv1fBMnm8fHOSJw0PEDI32xhiTMxn+5IfHMXSND94yf7+/16PRHA2ww3U5M5Pn1ERli4zz1V+ri6Q9BPfe2MatPY3c0t1AT1P0ssdcGiiT2SLPvj7GU4eHOTE6QzpbWeL2wp3fgoDf5yWVLdZVIK20SxstUAnzqezKbVpeDdUI8nbgwqEIQ8Dtlz5IKfUZ4DMAXV1dVbisWC7dyTCHBys3JUN+nVzRYndnA6PpAicnU1gWeAzw6nBDW4iP3baw19O7gIZhkw6/9LZe7tvetqj+6us1mS3yL30jxAyNRKyy3ECyIcrkTJZHXhpYUJBDJcw7EmFaYkHOTBXwU5nYFNQhZEPugvewmAZbN/i5+4ZW3ruznc0t0UW1mpNhP3f2JsmXLBrDAX7cdxYnD35fZT9Rj4KuRh+Fsruo9evXo0sbLQC5okXY56lhVYu3Yjc7Xdf9MvBlgL179y7f1jViyZJhP7s64wxMZpnKlgj7PLxnVxs3tob4q+cHGJjKonTFxkSEu7Yk2beldUHnDRjXHn3SHYL/9PDN3Nm7uIW4lkYjU7ToSl48qqMhFmIqt/A+/fO8Ho1dbWH2dIV5/kwW0wJDp9LJDdza6eehPT3Egwbb2qJ0JyPX1fXREPKxuTnE2ZkCYY+HjGNTMitLJ2xI+GiKRbDtyrr24uqu1GjJmxa7Ouvrza8avy3DwIUr03ecOybqWDLsv+wjeTLsZ2Nr/JojQq7F79XYGINTV5ky/7kP717hEIfmsJeQoZiamaWxIYLtuNiOSyFXoDFyfV0SG1ti/Mo7e3GeOc7RMwUKFsSA992U5LcfvKkqXR1BQ2dPTxJNwZuTaaYLKUI+jfbGMCGfgeVYhL2eRa1fvx5dqdFSbzc6oTpB/hKwRSm1kUqAfxz4RBXOK1ahhYwIuZqAofOb9+/k//z2EcYvWJ8rBPzOB7bywK6V36mmrSHE+2/axN+8eBJ3epZ4NMxsNo+Jxr+5bWHdKpcKGjp3bG6hKRLg1GQOy3ZojvnZ1lrdyThBQ+fOLa20xYN87cd9HD2bwlAaSnNQliIeDHD7pkTVrrdWXanRUm+WHOSu61pKqc8CT1IZfvhV13WPLrkyUXf6RlPs759mOm+SCBrcvilxUf+s16PxgZs78Wjw2CtDZAoWybDBw7d21mxZ1aCh80vv2EwooPPYwdNMzuZIRv383L6eBfePX+28uzoS7OpY/iDtaYry6Xtv4KnDI7w6lMJxbXa0N/C+3a3SP75OKNdd+e7qvXv3ugcOHFjx64rl0zea4olDI3MbRl+4o5GEiRDVoZR62XXdvZcelwGmoir290/PhThUtqdr8HvZ3z9d48qEWPskyEVVTOfNizaKhkqYy6gJIZafBLmoikTQIDV78SSK1GxJRk0IsQIkyEVV3L4pwUyxPBfm5/vIZdSEEMtPglxURW9rnAd3t+HVNQancnh1TW50CrFCZBlbUTW9rXEJbiFqQFrkQghR5yTIhRCizkmQCyFEnZMgF0KIOidBLoQQdU6CXAgh6pwEuRBC1DkZRy549tgwj78yzGSuTDLk5YN72rl3e3utyxJCLJAE+Tr3v/71fr55ZBKo/DD0xGBstrLQlYS5EPVBulbWsS99/425EI/7Kruw96VhbCbN46/Ibn1C1AsJ8nXsGy+cBCohDhDzgQJOp2EyV65dYUKIRZEgX8cyBQgCpQtWn435Khu+J0PX3vFeCLF6SJCvY9FApQVe5q0wT537+sE90j8uRL2QIF/HPnHHZjwa+KgE+vkQf3hnUm50ClFHZNTKOvZr920DKn3lmQJ0Birhfv64EKI+KNd1V/yie/fudQ8cOLDi1xVCiHqmlHrZdd29lx6XrhUhhKhzEuRCCFHnJMiFEKLOSZALIUSdkyAXQog6J0EuhBB1ToJcCCHqnAS5EELUOQlyIYSocxLkQghR5yTIhRCizkmQCyFEnZMgF0KIOidBLoQQdU7WIxdCVNX3jwzxrYNDTObKJENePnxLB/ft7Kh1WWuatMiFEFXz/SND/Nmz/ZiWTW9LBNOy+bNn+/n+kaFal7amLSnIlVL/RSk1rJR69dx/P1OtwoQQ9edbB4doCHpojocAaI6HaAh6+NZBCfLlVI2ulT9yXfcPq3AeIUSdm8yV6W2JXHSsOR6ib2y2RhWtD9K1IoSommTIy3gqd9Gx8VSOZMhbo4rWh2oE+WeVUoeUUl9VSjVc7UFKqc8opQ4opQ5MTExU4bJCiNXmw7d0MJO358J8PJVjJm/z4VvkZudymnfzZaXU94HWK3zr88ALwCTgAr8HtLmu+8vzXVQ2XxZi7ZJRK8vnapsvz9tH7rrufQu8wJ8D376O2oQQa8h9OyW4V9pSR620XfDXh4AjSytHCCHEYi111MoXlVI3U+laOQ386pIrEkIIsShLCnLXdX+hWoUIIYS4PjL8UAgh6pwEuRBC1Ll5hx8uy0WVmgAGVvzC80tSGU65HslzX3/W6/OG+n3u3a7rNl16sCZBvloppQ5caYzmeiDPff099/X6vGHtPXfpWhFCiDonQS6EEHVOgvxiX651ATUkz339Wa/PG9bYc5c+ciGEqHPSIhdCiDonQS6EEHVOghxQSv0rpdRRpZSjlNp7yfc+p5TqU0odV0q9t1Y1roT1tnWfUuqBc69rn1Lqd2pdz0pSSp1WSh0+9zqv6TWlz+2VMK6UOnLBsYRS6mml1Jvnvl51L4V6IEFecQT4CPDchQeVUtuBjwM7gAeA/6GU8qx8eSvqj1zXvfncf9+pdTHL5dzr+KfA+4DtwM+de73Xk3eee53XzHjqq/hLKr+/F/od4Aeu624BfnDu73VLghxwXfd113WPX+FbHwL+znXdkuu6p4A+YN/KVieWyT6gz3Xdftd1TeDvqLzeYo1xXfc5YPqSwx8Cvnbuz18DPryiRVWZBPm1tQODF/x96NyxtWxBW/etAevxtb2QCzyllHpZKfWZWhdTAy2u646c+/Mo0FLLYpZqqeuR141rbVnnuu5jK11Prcyzdd+XqGzZd37rvv8GzLt1n6hLb3Ndd1gp1Qw8rZR641zLdd1xXddVStX1OOx1E+QL3bLuEsNA5wV/7zh3rG7J1n1z1txruxiu6w6f+zqulHqUSlfTegryMaVUm+u6I+d2OhuvdUFLIV0r1/Y48HGllE8ptRHYArxY45qWzTrbuu8lYItSaqNSyqByU/vxGte0IpRSIaVU5PyfgftZ26/1lTwOfOrcnz8F1PWn8nXTIr8WpdRDwH8HmoB/Vkq96rrue13XPaqUegQ4BljAv3Vd165lrcts3Wzd57qupZT6LPAk4AG+6rru0RqXtVJagEeVUlDJgG+4rvu92pa0fJRSfwvcCySVUkPAfwa+ADyilPo0lSW1P1q7CpdOpugLIUSdk64VIYSocxLkQghR5yTIhRCizkmQCyFEnZMgF0KIOidBLoQQdU6CXAgh6tz/D0kPXJ8Ut4GqAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# project and visualize wordcount embeddings\n",
    "umap_model = umap.UMAP(n_neighbors=7, n_components=2)\n",
    "rawcount_emb_umap_proj = umap_model.fit_transform(np.array(list(df['rawcount_emb'].dropna())))\n",
    "plt.scatter(rawcount_emb_umap_proj[:,0], rawcount_emb_umap_proj[:,1], alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 337
    },
    "id": "eHTkaW9u0vyR",
    "outputId": "c62123fa-81d0-47bf-8695-c42cc8fbd6d0"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.7/dist-packages/numba/np/ufunc/parallel.py:363: NumbaWarning: The TBB threading layer requires TBB version 2019.5 or later i.e., TBB_INTERFACE_VERSION >= 11005. Found TBB_INTERFACE_VERSION = 9107. The TBB threading layer is disabled.\n",
      "  warnings.warn(problem)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f8cc1807890>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# project and visualize tfidf embeddings\n",
    "umap_model = umap.UMAP(n_neighbors=9, n_components=2)\n",
    "tfidf_emb_umap_proj = umap_model.fit_transform(np.array(list(df['tfidf_emb'].dropna())))\n",
    "plt.scatter(tfidf_emb_umap_proj[:,0], tfidf_emb_umap_proj[:,1], alpha=0.05)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "2cVq3sIVs0Ak"
   },
   "source": [
    "**Conclusion**: tfidf embeddings look most seperated in 2d.\n",
    "\n",
    "Use tfidf for clustering first. also try BERT\n",
    "\n",
    "**Conclusion**: Cluster seperation is not superb in 2dembeddings.\n",
    "\n",
    "Perhaps try higher dimensional embedding for clustering if 2d is unsuccessful"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z6dYnZoWwG4U"
   },
   "source": [
    "### cluster attempt 1\n",
    "\n",
    "Use HDBSCAN to cluster the UMAP-projected embedddings.  \n",
    "Start with 2d, try higher dimensional projections if cluster seperation is poor. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 421
    },
    "id": "JlNW5eQPQC-i",
    "outputId": "2509be77-e3a3-4bfa-ab60-1393486c08fd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 5    38118\n",
      " 4    12819\n",
      "-1     8284\n",
      " 0     5964\n",
      " 2     4362\n",
      " 1     4051\n",
      " 3     3197\n",
      "Name: label, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f8e06044310>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#cluster from initial 2d tfidf embeddings\n",
    "clusterer = hdbscan.HDBSCAN(min_cluster_size=1000)\n",
    "clusterer.fit(tfidf_emb_umap_proj)\n",
    "df['label'] = clusterer.labels_\n",
    "\n",
    "#inspect cluster distribution\n",
    "print(df['label'].value_counts())\n",
    "\n",
    "#visualize clusters in 2d\n",
    "plt.scatter(tfidf_emb_umap_proj[:,0], tfidf_emb_umap_proj[:,1], c = list(df['label']), alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 615
    },
    "id": "E8CWUbvATFFS",
    "outputId": "ba67a1c2-7655-4059-b891-6aa49552fa49"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-6687cb99-cb79-4f5b-8001-77e8ab17c47a\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <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>narrative</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14439</th>\n",
       "      <td>while taxing for departure, the airplane's no. 12 tire on the right body landing gear contacted foreign object debris (fod), which resulted in a cut tire and loss of pressure.  an electronic message reflecting the loss of pressure in the no. 12 tire was displayed on the engine indicating and crew alerting system (eicas) in the cockpit; however, no further action was taken by the flight crew.  during the takeoff roll, the two aft tires, no. 11 (inboard) and 12 (outboard), on the right body landing gear (rblg) burst and their associated wheels fractured, which caused minor damage to the airplane.  the airplane returned back to the airport and landed uneventfully.  examination of the airplane's systems and wheel assemblies revealed no additional anomalies.</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18194</th>\n",
       "      <td>when the airplane turned downwind, the engine lost power. the pilot attempted to restart the engine, and it surged suddenly. the torque caused the airplane to roll inverted. the pilot managed to roll the airplane upright, but was unable to restart the engine and he made a forced landing on airport property, striking the ground hard and substantially damaging the airplane. examination and tesing of the carburetor revealed the enrichment cover was leaking, and there  was \"a substantial leak\" at one of the vapor vent ports. one of the vapor vent floats was not functioning. when the vapor vent port was plugged, the carburetor tested between 7 and 22 pounds per hour above the rich limit.</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28510</th>\n",
       "      <td>the pilot reported that she leaned the mixture prior to takeoff.  shortly after departure, during the climb, the pilot readjusted the mixture (aft) due to poor aircraft performance.  shortly after readjusting the mixture, the engine began to run rough and eventually quit.  the pilot initiated a forced landing to an open field near the airport.  during the landing roll, the aircraft's nose gear collapsed.  substantial damage was reported to the aircraft's left wing, left aileron and firewall.  the density altitude at nampa airport, during the same timeframe as the accident, was approximately 3,890 ft.</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9979</th>\n",
       "      <td>the pilot reported that at 70 miles from his destination he determined that he would be landing closer to minimum fuel than he had planned. he continued to the destination and passed over one airport because it did not have any services. at 35 miles from the destination, the low fuel caution-warning for fuel tank 2 illuminated, followed by a low fuel caution-warning for fuel tank 1. the pilot thought that he would still land with some reserve and continued to fly to his destination airport. approximately 1 mile from the destination airport the helicopter experienced a dual engine flame-out.  the pilot autorotated to a sports field where the helicopter landed hard, sustaining damage to the main rotor blades, tail rotor, tail boom, and horizontal stabilizer. the pilot stated that the helicopter and engines had no mechanical failures or malfunctions during the flight.</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72096</th>\n",
       "      <td>according to the surviving passenger, a boy, the pilot was flying the aircraft from the right seat. the aircraft was flown to the vicinity of the pilot's property where he made a pass over a farm belonging to a relative. the low pass was made near the operator of a combine who was harvesting grain in a field. during the pass, the aircraft collided with a power line, crashed and burned. the boy was the only occupant to escape from the plane.</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6687cb99-cb79-4f5b-8001-77e8ab17c47a')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-6687cb99-cb79-4f5b-8001-77e8ab17c47a button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-6687cb99-cb79-4f5b-8001-77e8ab17c47a');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           narrative  label\n",
       "14439                                                                                                                while taxing for departure, the airplane's no. 12 tire on the right body landing gear contacted foreign object debris (fod), which resulted in a cut tire and loss of pressure.  an electronic message reflecting the loss of pressure in the no. 12 tire was displayed on the engine indicating and crew alerting system (eicas) in the cockpit; however, no further action was taken by the flight crew.  during the takeoff roll, the two aft tires, no. 11 (inboard) and 12 (outboard), on the right body landing gear (rblg) burst and their associated wheels fractured, which caused minor damage to the airplane.  the airplane returned back to the airport and landed uneventfully.  examination of the airplane's systems and wheel assemblies revealed no additional anomalies.          5\n",
       "18194                                                                                                                                                                                            when the airplane turned downwind, the engine lost power. the pilot attempted to restart the engine, and it surged suddenly. the torque caused the airplane to roll inverted. the pilot managed to roll the airplane upright, but was unable to restart the engine and he made a forced landing on airport property, striking the ground hard and substantially damaging the airplane. examination and tesing of the carburetor revealed the enrichment cover was leaking, and there  was \"a substantial leak\" at one of the vapor vent ports. one of the vapor vent floats was not functioning. when the vapor vent port was plugged, the carburetor tested between 7 and 22 pounds per hour above the rich limit.      5\n",
       "28510                                                                                                                                                                                                                                                                               the pilot reported that she leaned the mixture prior to takeoff.  shortly after departure, during the climb, the pilot readjusted the mixture (aft) due to poor aircraft performance.  shortly after readjusting the mixture, the engine began to run rough and eventually quit.  the pilot initiated a forced landing to an open field near the airport.  during the landing roll, the aircraft's nose gear collapsed.  substantial damage was reported to the aircraft's left wing, left aileron and firewall.  the density altitude at nampa airport, during the same timeframe as the accident, was approximately 3,890 ft.       5\n",
       "9979   the pilot reported that at 70 miles from his destination he determined that he would be landing closer to minimum fuel than he had planned. he continued to the destination and passed over one airport because it did not have any services. at 35 miles from the destination, the low fuel caution-warning for fuel tank 2 illuminated, followed by a low fuel caution-warning for fuel tank 1. the pilot thought that he would still land with some reserve and continued to fly to his destination airport. approximately 1 mile from the destination airport the helicopter experienced a dual engine flame-out.  the pilot autorotated to a sports field where the helicopter landed hard, sustaining damage to the main rotor blades, tail rotor, tail boom, and horizontal stabilizer. the pilot stated that the helicopter and engines had no mechanical failures or malfunctions during the flight.      5\n",
       "72096                                                                                                                                                                                                                                                                                                                                                                                                                                                  according to the surviving passenger, a boy, the pilot was flying the aircraft from the right seat. the aircraft was flown to the vicinity of the pilot's property where he made a pass over a farm belonging to a relative. the low pass was made near the operator of a combine who was harvesting grain in a field. during the pass, the aircraft collided with a power line, crashed and burned. the boy was the only occupant to escape from the plane.       5"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#inspect categories\n",
    "cluster_to_inspect = 0\n",
    "pd.options.display.max_colwidth = 1000\n",
    "df[['narrative','label']][df['label']==cluster_to_inspect].sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ng416JrpuaNO"
   },
   "source": [
    "**Conlusion**: tfidf yeilds clear geometric seperation, but clusters dont seem useful in terms of text content. \n",
    "\n",
    "Try BERT, may yeild better results than tfidf. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "M1zzBgKmwLlR"
   },
   "source": [
    "### Cluster attempt 2\n",
    "\n",
    "Try BERT embeddings. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 473
    },
    "id": "0UEEgx8sWhoJ",
    "outputId": "9a9d9fe5-540d-489c-8290-b6165b5f896c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 7    57574\n",
      " 8     5394\n",
      " 5     5284\n",
      " 3     2862\n",
      "-1     2095\n",
      " 4     1177\n",
      " 2     1089\n",
      " 0      595\n",
      " 1      394\n",
      " 6      331\n",
      "Name: label, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f8cbfaaded0>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#cluster from initial 2d bert embeddings\n",
    "clusterer = hdbscan.HDBSCAN(min_cluster_size=250)\n",
    "clusterer.fit(bert_emb_umap_proj)\n",
    "df['label'] = clusterer.labels_\n",
    "\n",
    "#inspect cluster distribution\n",
    "print(df['label'].value_counts())\n",
    "\n",
    "#visualize clusters in 2d\n",
    "plt.scatter(bert_emb_umap_proj[:,0], bert_emb_umap_proj[:,1], c = list(df['label']), alpha=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 563
    },
    "id": "DoBLc9m7suTw",
    "outputId": "2de55185-b98d-4c26-f7de-26fe1c83fba3"
   },
   "outputs": [],
   "source": [
    "#inspect categories\n",
    "cluster_to_inspect = 8\n",
    "pd.options.display.max_colwidth = 1000\n",
    "df[['narrative','label']][df['label']==cluster_to_inspect].sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7orvuLLjzKDx"
   },
   "source": [
    "cluster descriptions from clustering attempt two:\n",
    "\n",
    "0. hot air balloon accidents \n",
    "1. empty string narratives (should have filtered out. oops.)\n",
    "2. reports featuring descriptions of injuries and aircraft ownership. probably insurance reports.\n",
    "3. reports with many proper nouns, especially location names\n",
    "4. glider accidents\n",
    "5. helicopter accidents\n",
    "6. extreme weather incidents\n",
    "7. pilot error\n",
    "8. student pilot accidents"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cqhXHUsVdya2"
   },
   "source": [
    "Conclusion: (BERT + UMAP + HDBSCAN) has clustered the text into meaninful categories. However, the categories do not deal primarily with probable causes of the accidents, except for categories 6,7,8."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sSqyVWJQeLiK"
   },
   "source": [
    "# Final visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "id": "QixSvSq7ePWK"
   },
   "outputs": [],
   "source": [
    "from matplotlib.lines import Line2D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {
    "id": "EBq2cHscsNx0"
   },
   "outputs": [],
   "source": [
    "cluster_names = {\n",
    "                 0:'hot air balloon accidents',\n",
    "                 1: 'no narrative given', \n",
    "                 2: 'insurance reports',\n",
    "                 3: 'accidents in foreign airspace',\n",
    "                 4: 'glider accidents',\n",
    "                 5: 'helicopter accidents',\n",
    "                 6: 'extreme weather incidents',\n",
    "                 7: 'pilot error',\n",
    "                 8: 'student pilot accidents',\n",
    "                 -1:'unclustered'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {
    "id": "rSEpLZsTgauG"
   },
   "outputs": [],
   "source": [
    "# reformat plotting data for conveneience\n",
    "x = bert_emb_umap_proj[:,0]\n",
    "y = bert_emb_umap_proj[:,1]\n",
    "labels = list(df['label'])\n",
    "names = [cluster_names[c] for c in labels]\n",
    "\n",
    "plot_data = pd.DataFrame(zip(x, y, labels, names), columns = ['x','y','label', 'name'])\n",
    "# plot_data = plot_data[plot_data['label']!=1] #drop empty string cluster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "FJumCl6tj-Fn",
    "outputId": "aa6f2082-c3b6-43d5-9970-af2f8b426756"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    57574\n",
       "1     5394\n",
       "2     5284\n",
       "3     2862\n",
       "9     2095\n",
       "4     1177\n",
       "5     1089\n",
       "6      595\n",
       "7      394\n",
       "8      331\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#reorder cluster labels by cluster size, and put unclustered data last\n",
    "cluster_rename_map = dict(zip(list(plot_data[plot_data['label']!=-1]['label'].value_counts().index), range(0,100)))\n",
    "cluster_rename_map[-1] = len(cluster_rename_map)\n",
    "plot_data['label'] = plot_data['label'].apply(lambda x: cluster_rename_map[x] if x in cluster_rename_map else x)\n",
    "plot_data['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "id": "dtKCyE_qmaBF"
   },
   "outputs": [],
   "source": [
    "plot_data = plot_data.sort_values('label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "id": "JA_Y8GWjjWG0",
    "outputId": "666e5c1e-f686-4645-aa5a-fa4d7bb524b8"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjwAAABECAYAAACF4e8fAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAACLklEQVR4nO3bMWpUURiG4TtBogRMMYUIA0ISsJtCEKa20SUIsXENNuIKUiVLCa5AsLztFEIatRnQSkGLQMIcN5BMIfxzwsfztH/zdfeFw5201toAABBsp/cAAIBqggcAiCd4AIB4ggcAiCd4AIB49zYdn86fDW1vuq0tWzff+dZ7Qqkvu7u9J5Q5/JH9c+Gfh096Tyi1vv7Ze0Kp6f3HvSeUutz/3ntCma/DUe8JpR79/d17Qqn1ej2M43jjbWPwtL3pcPXiXcmou+D8wXHvCaXmB7kfzdOT694TSn16/qH3hFKXv856Tyj1+uB97wmlLl6+7T2hzJvJSe8JpV59/th7QqnlcnnrzZMWABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBP8AAA8QQPABBv0lprtx0Xi8Uwm822uQcA4L+sVqthHMcbbxuDBwAggSctACCe4AEA4gkeACCe4AEA4gkeACDeP9k4ODJ1+Ux+AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x72 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#set font size of labels on matplotlib plots\n",
    "plt.rc('font', size=16)\n",
    "\n",
    "#set style of plots\n",
    "sns.set_style('white')\n",
    "\n",
    "#define a color palette. a reodered tab10 palette. put grey last for null cluster\n",
    "customPalette = ['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#bcbd22','#17becf','#7f7f7f']\n",
    "sns.set_palette(customPalette)\n",
    "sns.palplot(customPalette)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 346
    },
    "id": "B27tyFjVh3kk",
    "outputId": "89e7933c-3969-4db9-bed0-1f3cf521bd91"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#create a new figure\n",
    "plt.figure(figsize=(5,5))\n",
    "\n",
    "#loop through labels and plot each cluster\n",
    "legend_elements = list()\n",
    "for i, label in enumerate(plot_data['label'].unique()):\n",
    "\n",
    "    #add data points \n",
    "    plt.scatter(x=plot_data.loc[plot_data['label']==label,'x'], \n",
    "                y=plot_data.loc[plot_data['label']==label,'y'], \n",
    "                color=customPalette[i], \n",
    "                alpha=0.20)\n",
    "    \n",
    "    #add label\n",
    "    if label != len(plot_data['label'].unique())-1: #for all clusters, plot\n",
    "      plt.annotate(label, \n",
    "                  plot_data.loc[plot_data['label']==label,['x','y']].mean(),\n",
    "                  horizontalalignment='center',\n",
    "                  verticalalignment='center',\n",
    "                  size=20, weight='bold',\n",
    "                  color='k') \n",
    "      \n",
    "    #make legend element\n",
    "    cluster_name = list(plot_data.loc[plot_data['label']==label,'name'])[0] #get name\n",
    "    cluster_name = cluster_name.title() #make name title case\n",
    "    legend_elements.append(Line2D([0], [0], marker='o', \n",
    "                                  color='w', label=cluster_name,\n",
    "                                  markerfacecolor=customPalette[i], \n",
    "                                  markersize=10))\n",
    "      \n",
    "#hide tick labels\n",
    "plt.xticks(color='w')\n",
    "plt.yticks(color='w')\n",
    "\n",
    "#create legend\n",
    "plt.legend(handles = legend_elements, bbox_to_anchor=(1.05, 1.0))\n",
    "plt.title('Aviation Accident Narrative Clusters')\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "RTI_exercise.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}