{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# US Presidential Debate Sep 26 2016\n",
    "\n",
    "This is a Jupyter notebook detailing an analysis of the language used in the first debate.\n",
    "\n",
    "Prepared by Doug Blank, Bryn Mawr College  \n",
    "For full discussion, see: http://blankversusblank.blogspot.com/2016/09/post-debate-analysis.html\n",
    "\n",
    "Data from http://www.nytimes.com/2016/09/27/us/politics/transcript-debate.html\n",
    "\n",
    "There were some errors that I corrected, so you can use the version here [first_debate.txt](first_debate.txt)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we read the text into an array of lines:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "text = [line.strip().replace(\"\\n\", \" \").replace(\".\", \" \").replace(\"?\", \" \")\n",
    "        .replace(\"“\", \" \").replace(\"”\", \" \").replace(\":\", \" \")\n",
    "        .replace(\",\", \" \").replace(\"—\", \" \").replace(\"-\", \" \")\n",
    "        for line in open(\"first_debate.txt\").readlines()]\n",
    "text_all = \" \".join(text)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A sample to see what it looks like:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'HOLT  Good evening from Hofstra University in Hempstead  New York  I’m Lester Holt  anchor of  NBC Nightly News   I want to welcome you to the first presidential debate '"
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we break it down by speaker:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "holt = \"\"\n",
    "clinton = \"\"\n",
    "trump = \"\"\n",
    "\n",
    "current = None\n",
    "for line in text:\n",
    "    if not line:\n",
    "        continue\n",
    "    elif line in [\"(APPLAUSE)\", \"(CROSSTALK)\", \"(LAUGHTER)\"]:\n",
    "        continue\n",
    "    elif line.startswith(\"HOLT\"):\n",
    "        current = \"HOLT\"\n",
    "        holt += line[4:] + \" \"\n",
    "    elif line.startswith(\"TRUMP\"):\n",
    "        current = \"TRUMP\"\n",
    "        trump += line[5:] + \" \"\n",
    "    elif line.startswith(\"CLINTON\"):\n",
    "        current = \"CLINTON\"\n",
    "        clinton += line[7:] + \" \"\n",
    "    else:\n",
    "        if current == \"HOLT\":\n",
    "            holt += line + \" \"\n",
    "        elif current == \"TRUMP\":\n",
    "            trump += line + \" \"\n",
    "        elif current == \"CLINTON\":\n",
    "            clinton += line + \" \"\n",
    "        else:\n",
    "            raise Exception(\"No speaker?!\")\n",
    "            \n",
    "holt = holt.lower()\n",
    "clinton = clinton.lower()\n",
    "trump = trump.lower()\n",
    "\n",
    "clinton = clinton.strip()\n",
    "while \"  \" in clinton:\n",
    "    clinton = clinton.replace(\"  \", \" \")\n",
    "holt = holt.strip()\n",
    "while \"  \" in holt:\n",
    "    holt = holt.replace(\"  \", \" \").strip()\n",
    "trump = trump.strip()\n",
    "while \"  \" in trump:\n",
    "    trump = trump.replace(\"  \", \" \").strip()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Characters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10400, 42263, 33173)"
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(holt), len(trump), len(clinton)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And split the text into words:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "clinton_words = clinton.split(\" \")\n",
    "trump_words = trump.split(\" \")\n",
    "holt_words = holt.split(\" \")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of total \"words\" spoken"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6237, 8139, 1878)"
      ]
     },
     "execution_count": 170,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(clinton_words), len(trump_words), len(holt_words)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of unique words spoken"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "clinton_set = set(clinton_words)\n",
    "trump_set = set(trump_words)\n",
    "holt_set = set(holt_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1379, 1269)"
      ]
     },
     "execution_count": 172,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(clinton_set), len(trump_set)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of each word spoken, ranked from highest to lowest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def make_dict(words):\n",
    "    d = {}\n",
    "    for word in words:\n",
    "        count = d.get(word, 0)\n",
    "        d[word] = count + 1\n",
    "    return d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "clinton_dict = make_dict(clinton_words)\n",
    "trump_dict = make_dict(trump_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "common_words = [\"the\", \"to\", \"and\", \"or\", \"that\", \"of\", \"a\", \"in\", \"have\", \"it\", \"be\",\n",
    "                \"am\", \"are\", \"was\", \"were\", \"been\", \"be\", \"being\", \"is\", \"do\", \"would\",\n",
    "                \"but\", \"what\", \"so\", \"with\", \"about\", \"at\", \"on\", \"has\", \"can\", \"as\",\n",
    "                \"because\", \"when\", \"by\", \"an\", \"for\", \"this\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "i: 138\n",
      "we: 122\n",
      "you: 76\n",
      "he: 56\n",
      "our: 42\n",
      "not: 40\n",
      "think: 38\n",
      "well: 36\n",
      "people: 32\n",
      "they: 31\n",
      "know: 28\n",
      "going: 27\n",
      "donald: 26\n",
      "your: 25\n",
      "one: 24\n",
      "need: 23\n",
      "us: 22\n",
      "who: 21\n",
      "more: 21\n",
      "will: 21\n",
      "that’s: 21\n",
      "them: 21\n",
      "it’s: 21\n",
      "really: 20\n",
      "there: 20\n",
      "their: 19\n",
      "want: 19\n",
      "his: 18\n",
      "from: 18\n",
      "said: 17\n",
      "if: 17\n",
      "we’re: 17\n",
      "country: 16\n",
      "just: 16\n",
      "good: 16\n",
      "we’ve: 16\n",
      "jobs: 16\n",
      "lot: 16\n",
      "tax: 16\n",
      "make: 15\n",
      "get: 15\n",
      "new: 15\n",
      "out: 15\n",
      "up: 15\n",
      "business: 15\n",
      "got: 14\n",
      "how: 14\n",
      "work: 14\n",
      "some: 14\n",
      "should: 14\n",
      "go: 13\n",
      "all: 13\n",
      "economy: 13\n",
      "very: 13\n",
      "also: 12\n",
      "american: 12\n",
      "had: 12\n",
      "nuclear: 11\n",
      "see: 11\n",
      "he’s: 11\n",
      "no: 11\n",
      "down: 10\n",
      "debt: 10\n",
      "don’t: 10\n",
      "look: 10\n",
      "million: 10\n",
      "my: 10\n",
      "too: 10\n",
      "into: 10\n",
      "i’ve: 10\n",
      "actually: 10\n",
      "many: 10\n",
      "kind: 10\n",
      "fact: 10\n",
      "important: 10\n",
      "now: 9\n",
      "put: 9\n",
      "other: 9\n",
      "police: 9\n",
      "deal: 9\n",
      "did: 9\n",
      "again: 9\n",
      "information: 9\n",
      "middle: 9\n",
      "say: 9\n",
      "wealthy: 8\n",
      "those: 8\n",
      "first: 8\n",
      "isis: 8\n",
      "iran: 8\n",
      "let’s: 8\n",
      "plan: 8\n",
      "president: 8\n",
      "talk: 8\n",
      "back: 8\n",
      "over: 8\n",
      "then: 8\n",
      "support: 8\n",
      "years: 8\n",
      "much: 7\n",
      "i’m: 7\n",
      "lester: 7\n",
      "ever: 7\n",
      "things: 7\n",
      "why: 7\n",
      "me: 7\n",
      "class: 7\n",
      "something: 7\n",
      "even: 7\n",
      "paid: 7\n",
      "state: 7\n",
      "states: 7\n",
      "him: 7\n",
      "working: 7\n",
      "pay: 7\n",
      "trade: 7\n",
      "world: 7\n",
      "time: 7\n",
      "proposed: 7\n",
      "only: 7\n",
      "communities: 7\n",
      "home: 7\n",
      "where: 7\n",
      "taken: 6\n",
      "everyone: 6\n",
      "top: 6\n",
      "national: 6\n",
      "young: 6\n",
      "percent: 6\n",
      "everything: 6\n",
      "having: 6\n",
      "thing: 6\n",
      "man: 6\n",
      "money: 6\n",
      "called: 6\n",
      "which: 6\n",
      "they’ve: 6\n",
      "part: 6\n",
      "we’ll: 6\n",
      "made: 6\n",
      "different: 6\n",
      "nations: 6\n",
      "weapons: 6\n",
      "take: 6\n",
      "you’re: 6\n",
      "any: 6\n",
      "trying: 6\n",
      "government: 6\n",
      "after: 6\n",
      "worked: 6\n",
      "able: 6\n",
      "better: 6\n",
      "against: 6\n",
      "cyber: 5\n",
      "returns: 5\n",
      "obama: 5\n",
      "two: 5\n",
      "future: 5\n",
      "plans: 5\n",
      "come: 5\n",
      "iraq: 5\n",
      "there’s: 5\n",
      "right: 5\n",
      "heard: 5\n",
      "united: 5\n",
      "kinds: 5\n",
      "never: 5\n",
      "these: 5\n",
      "together: 5\n",
      "maybe: 5\n",
      "give: 5\n",
      "fair: 5\n",
      "system: 5\n",
      "believe: 5\n",
      "number: 5\n",
      "done: 5\n",
      "federal: 5\n",
      "sure: 5\n",
      "law: 5\n",
      "long: 5\n",
      "both: 5\n",
      "justice: 5\n",
      "use: 5\n",
      "incomes: 5\n",
      "looked: 5\n",
      "hope: 5\n",
      "let: 5\n",
      "doing: 4\n",
      "like: 4\n",
      "secretary: 4\n",
      "saying: 4\n",
      "benefit: 4\n",
      "facing: 4\n",
      "started: 4\n",
      "facts: 4\n",
      "troops: 4\n",
      "york: 4\n",
      "black: 4\n",
      "5: 4\n",
      "does: 4\n",
      "clean: 4\n",
      "says: 4\n",
      "families: 4\n",
      "hack: 4\n",
      "clear: 4\n",
      "provide: 4\n",
      "went: 4\n",
      "finally: 4\n",
      "worst: 4\n",
      "best: 4\n",
      "zero: 4\n",
      "america: 4\n",
      "determines: 4\n",
      "debate: 4\n",
      "deals: 4\n",
      "means: 4\n",
      "security: 4\n",
      "gun: 4\n",
      "try: 4\n",
      "understand: 4\n",
      "they’re: 4\n",
      "way: 4\n",
      "real: 4\n",
      "donald’s: 4\n",
      "family: 4\n",
      "taxes: 4\n",
      "nato: 4\n",
      "military: 4\n",
      "help: 4\n",
      "foreign: 4\n",
      "attacks: 4\n",
      "question: 4\n",
      "trillion: 4\n",
      "buy: 4\n",
      "sometimes: 4\n",
      "making: 4\n",
      "here: 4\n",
      "biggest: 4\n",
      "you’ve: 4\n",
      "job: 4\n",
      "reasons: 4\n",
      "around: 4\n",
      "russia: 4\n",
      "hard: 4\n",
      "year: 4\n",
      "invest: 4\n",
      "most: 4\n",
      "off: 4\n",
      "same: 4\n",
      "growth: 4\n",
      "energy: 4\n",
      "businesses: 4\n",
      "small: 4\n",
      "add: 4\n",
      "under: 4\n",
      "away: 4\n",
      "could: 4\n",
      "trickle: 4\n",
      "great: 4\n",
      "create: 4\n",
      "still: 4\n",
      "second: 4\n",
      "issues: 4\n",
      "problems: 4\n",
      "recession: 3\n",
      "street: 3\n",
      "father: 3\n",
      "grow: 3\n",
      "seen: 3\n",
      "intelligence: 3\n",
      "40: 3\n",
      "another: 3\n",
      "asked: 3\n",
      "half: 3\n",
      "americans: 3\n",
      "deserve: 3\n",
      "course: 3\n",
      "vote: 3\n",
      "opportunities: 3\n",
      "building: 3\n",
      "lose: 3\n",
      "installers: 3\n",
      "financial: 3\n",
      "white: 3\n",
      "start: 3\n",
      "private: 3\n",
      "live: 3\n",
      "muslim: 3\n",
      "near: 3\n",
      "may: 3\n",
      "she: 3\n",
      "before: 3\n",
      "keep: 3\n",
      "every: 3\n",
      "race: 3\n",
      "sailors: 3\n",
      "stand: 3\n",
      "life: 3\n",
      "prepared: 3\n",
      "responsibilities: 3\n",
      "college: 3\n",
      "matter: 3\n",
      "remember: 3\n",
      "took: 3\n",
      "her: 3\n",
      "release: 3\n",
      "budget: 3\n",
      "crime: 3\n",
      "face: 3\n",
      "criminal: 3\n",
      "investments: 3\n",
      "china: 3\n",
      "racist: 3\n",
      "though: 3\n",
      "war: 3\n",
      "build: 3\n",
      "met: 3\n",
      "policy: 3\n",
      "voted: 3\n",
      "tried: 3\n",
      "health: 3\n",
      "men: 3\n",
      "education: 3\n",
      "efforts: 3\n",
      "leadership: 3\n",
      "putin: 3\n",
      "lie: 3\n",
      "$5: 3\n",
      "abroad: 3\n",
      "absolutely: 3\n",
      "rising: 3\n",
      "enough: 3\n",
      "else: 3\n",
      "word: 3\n",
      "problem: 3\n",
      "barack: 3\n",
      "can’t: 3\n",
      "intend: 3\n",
      "african: 3\n",
      "happen: 3\n",
      "share: 3\n",
      "east: 3\n",
      "unfortunately: 3\n",
      "ways: 3\n",
      "defeat: 3\n"
     ]
    }
   ],
   "source": [
    "for pair in sorted([items for items in clinton_dict.items() if items[0] not in common_words and\n",
    "                   items[1] > 2],\n",
    "                   key=lambda pair: pair[1], reverse=True):\n",
    "    print(\"%s: %s\" % pair)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 8)"
      ]
     },
     "execution_count": 177,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clinton_dict[\"china\"], clinton_dict[\"plan\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9, 3)"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trump_dict[\"china\"], trump_dict[\"plan\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "i: 229\n",
      "you: 189\n",
      "we: 109\n",
      "it’s: 72\n",
      "they: 71\n",
      "very: 65\n",
      "our: 55\n",
      "not: 50\n",
      "country: 46\n",
      "going: 43\n",
      "all: 41\n",
      "they’re: 40\n",
      "look: 40\n",
      "me: 39\n",
      "said: 37\n",
      "think: 37\n",
      "them: 35\n",
      "just: 35\n",
      "she: 33\n",
      "don’t: 32\n",
      "will: 32\n",
      "say: 32\n",
      "i’m: 29\n",
      "that’s: 29\n",
      "people: 28\n",
      "no: 28\n",
      "doing: 27\n",
      "out: 27\n",
      "know: 27\n",
      "get: 26\n",
      "secretary: 25\n",
      "clinton: 25\n",
      "should: 25\n",
      "one: 25\n",
      "we’re: 24\n",
      "many: 24\n",
      "want: 23\n",
      "years: 22\n",
      "now: 21\n",
      "thing: 21\n",
      "things: 21\n",
      "did: 21\n",
      "their: 20\n",
      "like: 20\n",
      "your: 20\n",
      "can’t: 20\n",
      "if: 20\n",
      "much: 20\n",
      "other: 20\n",
      "companies: 20\n",
      "her: 20\n",
      "jobs: 19\n",
      "good: 19\n",
      "really: 19\n",
      "well: 19\n",
      "way: 19\n",
      "my: 19\n",
      "some: 19\n",
      "go: 18\n",
      "these: 18\n",
      "from: 17\n",
      "money: 17\n",
      "tell: 17\n",
      "new: 17\n",
      "over: 16\n",
      "great: 16\n",
      "you’re: 16\n",
      "into: 15\n",
      "believe: 15\n",
      "time: 15\n",
      "lot: 15\n",
      "tax: 15\n",
      "us: 15\n",
      "up: 15\n",
      "could: 15\n",
      "leaving: 15\n",
      "bad: 14\n",
      "ever: 14\n",
      "world: 14\n",
      "deal: 14\n",
      "isis: 14\n",
      "he: 14\n",
      "against: 14\n",
      "agree: 13\n",
      "got: 13\n",
      "more: 13\n",
      "first: 13\n",
      "bring: 13\n",
      "which: 13\n",
      "back: 13\n",
      "war: 12\n",
      "trillion: 12\n",
      "she’s: 12\n",
      "i’ll: 12\n",
      "right: 12\n",
      "lester: 12\n",
      "even: 12\n",
      "than: 12\n",
      "wrong: 12\n",
      "countries: 12\n",
      "where: 12\n",
      "see: 11\n",
      "done: 11\n",
      "had: 11\n",
      "how: 11\n",
      "tremendous: 11\n",
      "take: 11\n",
      "better: 10\n",
      "nato: 10\n",
      "i’ve: 10\n",
      "also: 10\n",
      "hillary: 10\n",
      "politicians: 10\n",
      "why: 10\n",
      "there: 10\n",
      "let: 10\n",
      "stop: 10\n",
      "president: 10\n",
      "doesn’t: 10\n",
      "taken: 9\n",
      "there’s: 9\n",
      "trade: 9\n",
      "job: 9\n",
      "didn’t: 9\n",
      "you’ve: 9\n",
      "never: 9\n",
      "him: 9\n",
      "china: 9\n",
      "times: 9\n",
      "maybe: 9\n",
      "whether: 8\n",
      "big: 8\n",
      "trump: 8\n",
      "come: 8\n",
      "taking: 8\n",
      "far: 8\n",
      "then: 8\n",
      "almost: 8\n",
      "campaign: 8\n",
      "russia: 8\n",
      "company: 8\n",
      "community: 8\n",
      "talking: 8\n",
      "deals: 8\n",
      "iran: 8\n",
      "need: 8\n",
      "sean: 8\n",
      "nuclear: 8\n",
      "happened: 8\n",
      "regulations: 7\n",
      "help: 7\n",
      "obama: 7\n",
      "greatest: 7\n",
      "who: 7\n",
      "saying: 7\n",
      "nafta: 7\n",
      "started: 7\n",
      "down: 7\n",
      "work: 7\n",
      "nobody: 7\n",
      "haven’t: 7\n",
      "defend: 7\n",
      "give: 7\n",
      "experience: 7\n",
      "nothing: 7\n",
      "before: 7\n",
      "30: 7\n",
      "middle: 7\n",
      "able: 7\n",
      "worst: 7\n",
      "under: 7\n",
      "paying: 7\n",
      "make: 7\n",
      "fact: 7\n",
      "hannity: 7\n",
      "day: 7\n",
      "business: 7\n",
      "losing: 7\n",
      "wait: 7\n",
      "taxes: 7\n",
      "last: 6\n",
      "seen: 6\n",
      "korea: 6\n",
      "mean: 6\n",
      "everybody: 6\n",
      "endorsed: 6\n",
      "ok: 6\n",
      "he’s: 6\n",
      "website: 6\n",
      "long: 6\n",
      "another: 6\n",
      "$20: 6\n",
      "oil: 6\n",
      "respond: 6\n",
      "spent: 6\n",
      "something: 6\n",
      "problem: 6\n",
      "little: 6\n",
      "debate: 6\n",
      "political: 6\n",
      "mess: 6\n",
      "happen: 6\n",
      "watch: 6\n",
      "mexico: 6\n",
      "release: 6\n",
      "east: 6\n",
      "important: 6\n",
      "laws: 5\n",
      "cyber: 5\n",
      "somebody: 5\n",
      "hundreds: 5\n",
      "terms: 5\n",
      "strongly: 5\n",
      "put: 5\n",
      "probably: 5\n",
      "billions: 5\n",
      "nation: 5\n",
      "york: 5\n",
      "million: 5\n",
      "fed: 5\n",
      "biggest: 5\n",
      "north: 5\n",
      "after: 5\n",
      "debt: 5\n",
      "everything: 5\n",
      "anywhere: 5\n",
      "used: 5\n",
      "called: 5\n",
      "stamina: 5\n",
      "land: 5\n",
      "dollars: 5\n",
      "income: 5\n",
      "question: 5\n",
      "major: 5\n",
      "formed: 5\n",
      "single: 5\n",
      "percent: 5\n",
      "off: 5\n",
      "true: 5\n",
      "disaster: 5\n",
      "only: 5\n",
      "cannot: 5\n",
      "fight: 5\n",
      "i’d: 5\n",
      "10: 5\n",
      "old: 5\n",
      "talks: 5\n",
      "we’ve: 5\n",
      "approve: 5\n",
      "getting: 5\n",
      "name: 5\n",
      "made: 5\n",
      "list: 5\n",
      "african: 5\n",
      "murders: 5\n",
      "totally: 5\n",
      "four: 5\n",
      "lots: 5\n",
      "returns: 5\n",
      "temperament: 5\n",
      "care: 5\n",
      "certainly: 5\n",
      "states: 4\n",
      "wealthy: 4\n",
      "places: 4\n",
      "certificate: 4\n",
      "different: 4\n",
      "interest: 4\n",
      "supposed: 4\n",
      "500: 4\n",
      "number: 4\n",
      "read: 4\n",
      "asked: 4\n",
      "american: 4\n",
      "thinking: 4\n",
      "ask: 4\n",
      "soon: 4\n",
      "frisk: 4\n",
      "around: 4\n",
      "raise: 4\n",
      "produce: 4\n",
      "place: 4\n",
      "sent: 4\n",
      "advantage: 4\n",
      "perhaps: 4\n",
      "year: 4\n",
      "during: 4\n",
      "lists: 4\n",
      "lawsuit: 4\n",
      "best: 4\n",
      "any: 4\n",
      "billion: 4\n",
      "proud: 4\n",
      "expand: 4\n",
      "anybody: 4\n",
      "gave: 4\n",
      "actually: 4\n",
      "iraq: 4\n",
      "cut: 4\n",
      "we’ll: 4\n",
      "find: 4\n",
      "minute: 4\n",
      "airports: 4\n",
      "relationships: 4\n",
      "does: 4\n",
      "once: 4\n",
      "two: 4\n",
      "audited: 4\n",
      "brought: 4\n",
      "japan: 4\n",
      "kind: 4\n",
      "real: 4\n",
      "budget: 4\n",
      "birth: 4\n",
      "excuse: 4\n",
      "what’s: 4\n",
      "week: 4\n",
      "talk: 4\n",
      "they’ve: 4\n",
      "needs: 4\n",
      "inner: 4\n",
      "cities: 4\n",
      "left: 4\n",
      "learn: 4\n",
      "tough: 4\n",
      "too: 4\n",
      "build: 3\n",
      "beautiful: 3\n",
      "night: 3\n",
      "myself: 3\n",
      "his: 3\n",
      "article: 3\n",
      "audit: 3\n",
      "feel: 3\n",
      "15: 3\n",
      "coming: 3\n",
      "concerned: 3\n",
      "reason: 3\n",
      "energy: 3\n",
      "yes: 3\n",
      "wonder: 3\n",
      "avenue: 3\n",
      "met: 3\n",
      "releases: 3\n",
      "trying: 3\n",
      "created: 3\n",
      "unbelievable: 3\n",
      "happy: 3\n",
      "strong: 3\n",
      "defective: 3\n",
      "quickly: 3\n",
      "5: 3\n",
      "fine: 3\n",
      "banks: 3\n",
      "800: 3\n",
      "course: 3\n",
      "shows: 3\n",
      "small: 3\n",
      "call: 3\n",
      "interview: 3\n",
      "200: 3\n",
      "$650: 3\n",
      "badly: 3\n",
      "michael: 3\n",
      "control: 3\n",
      "blumenthal: 3\n",
      "mine: 3\n",
      "life: 3\n",
      "roads: 3\n",
      "terror: 3\n",
      "000: 3\n",
      "looks: 3\n",
      "e: 3\n",
      "bit: 3\n",
      "start: 3\n",
      "thousands: 3\n",
      "dnc: 3\n",
      "hard: 3\n",
      "telling: 3\n",
      "arabia: 3\n",
      "communities: 3\n",
      "schedule: 3\n",
      "pennsylvania: 3\n",
      "history: 3\n",
      "opening: 3\n",
      "united: 3\n",
      "plants: 3\n",
      "isn’t: 3\n",
      "economy: 3\n",
      "cost: 3\n",
      "millions: 3\n",
      "told: 3\n",
      "numbers: 3\n",
      "own: 3\n",
      "assets: 3\n",
      "thinks: 3\n",
      "same: 3\n",
      "impact: 3\n",
      "reporter: 3\n",
      "estate: 3\n",
      "every: 3\n",
      "credit: 3\n",
      "went: 3\n",
      "anything: 3\n",
      "saudi: 3\n",
      "within: 3\n",
      "approved: 3\n",
      "person: 3\n",
      "d: 3\n",
      "keeping: 3\n",
      "election: 3\n",
      "end: 3\n",
      "admirals: 3\n",
      "fifth: 3\n",
      "manager: 3\n",
      "donald: 3\n",
      "yeah: 3\n",
      "mails: 3\n",
      "since: 3\n",
      "internet: 3\n",
      "certain: 3\n",
      "mainstream: 3\n",
      "fault: 3\n",
      "plan: 3\n",
      "order: 3\n",
      "ohio: 3\n",
      "shear: 3\n",
      "ahead: 3\n",
      "ago: 3\n",
      "winning: 3\n",
      "win: 3\n",
      "signed: 3\n",
      "love: 3\n",
      "nice: 3\n",
      "report: 3\n",
      "truth: 3\n",
      "defending: 3\n",
      "terrible: 3\n",
      "fly: 3\n",
      "obama’s: 3\n",
      "family: 3\n",
      "sell: 3\n",
      "rates: 3\n",
      "pay: 3\n",
      "frankly: 3\n",
      "oh: 3\n",
      "most: 3\n",
      "story: 3\n",
      "power: 3\n",
      "audit’s: 3\n",
      "worth: 3\n"
     ]
    }
   ],
   "source": [
    "for pair in sorted([items for items in trump_dict.items() if items[0] not in common_words and\n",
    "                   items[1] > 2],\n",
    "                    key=lambda pair: pair[1], reverse=True):\n",
    "    print(\"%s: %s\" % pair)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## How often did they speak"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "126"
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_all.count(\"TRUMP\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "93"
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_all.count(\"CLINTON\")"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}