Topic_Modeling_Financial_Reports.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.70796203613\n",
      "9.46735286713\n",
      "0.222633838654\n",
      "0.285443067551\n",
      "0.420050144196\n",
      "9.36676192284\n",
      "0.267348051071\n",
      "0.36127614975\n",
      "1.7410569191\n",
      "9.26254010201\n",
      "0.239897012711\n",
      "0.31273317337\n",
      "Review 4 of 33\n",
      "\n",
      "0.354735851288\n",
      "9.354129076\n",
      "0.265748023987\n",
      "0.31916308403\n",
      "1.42035603523\n",
      "10.8315441608\n",
      "0.735523939133\n",
      "0.578503131866\n",
      "0.693676948547\n",
      "7.50943803787\n",
      "0.352710962296\n",
      "0.342980861664\n",
      "1.46668791771\n",
      "5.90511989594\n",
      "0.319056987762\n",
      "0.313612222672\n",
      "Review 8 of 33\n",
      "\n",
      "0.6592669487\n",
      "5.63266682625\n",
      "0.276036977768\n",
      "0.256671905518\n",
      "0.528211116791\n",
      "4.584856987\n",
      "0.225298881531\n",
      "0.238942861557\n",
      "0.552925109863\n",
      "7.05968594551\n",
      "0.332903146744\n",
      "0.28605890274\n",
      "1.03310704231\n",
      "5.18860387802\n",
      "0.266998052597\n",
      "0.239639043808\n",
      "Review 12 of 33\n",
      "\n",
      "0.834041118622\n",
      "7.50169491768\n",
      "0.193987846375\n",
      "0.263561010361\n",
      "1.09483599663\n",
      "8.49080610275\n",
      "0.22155880928\n",
      "0.330150127411\n",
      "1.68242788315\n",
      "7.7941069603\n",
      "0.208788871765\n",
      "0.316426038742\n",
      "0.416291952133\n",
      "8.06660199165\n",
      "0.197158098221\n",
      "0.296669006348\n",
      "Review 16 of 33\n",
      "\n",
      "0.755106925964\n",
      "8.93793296814\n",
      "0.600372076035\n",
      "0.833534955978\n",
      "0.79233789444\n",
      "7.95226407051\n",
      "0.546429872513\n",
      "0.493972063065\n",
      "0.325896024704\n",
      "7.16856193542\n",
      "0.477825880051\n",
      "0.439838171005\n",
      "1.11599802971\n",
      "6.29571485519\n",
      "0.312236070633\n",
      "0.314493179321\n",
      "Review 20 of 33\n",
      "\n",
      "1.94952201843\n",
      "5.30968213081\n",
      "0.368964910507\n",
      "0.375640153885\n",
      "0.785202980042\n",
      "4.79327511787\n",
      "0.329118013382\n",
      "0.30588388443\n",
      "0.631713867188\n",
      "4.6356010437\n",
      "0.318578958511\n",
      "0.24152803421\n",
      "0.895048856735\n",
      "6.59443092346\n",
      "0.163467168808\n",
      "0.239860057831\n",
      "Review 24 of 33\n",
      "\n",
      "3.09714388847\n",
      "6.62759590149\n",
      "0.17006111145\n",
      "0.290781974792\n",
      "2.21812915802\n",
      "6.58287000656\n",
      "0.361886024475\n",
      "0.35689496994\n",
      "1.12074780464\n",
      "7.22348618507\n",
      "0.353237867355\n",
      "0.351586103439\n",
      "1.18457889557\n",
      "7.5023188591\n",
      "0.426394939423\n",
      "0.334240913391\n",
      "Review 28 of 33\n",
      "\n",
      "0.369579076767\n",
      "6.98576283455\n",
      "0.375415086746\n",
      "0.360445976257\n",
      "0.73940205574\n",
      "7.31580090523\n",
      "0.345856904984\n",
      "0.347598075867\n",
      "2.91692304611\n",
      "6.20147705078\n",
      "0.303822994232\n",
      "0.285761117935\n",
      "0.268045902252\n",
      "5.98839497566\n",
      "0.299963951111\n",
      "0.244870185852\n",
      "Review 32 of 33\n",
      "\n",
      "1.40888094902\n",
      "5.72494912148\n",
      "0.301513910294\n",
      "0.255808830261\n",
      "1.11089992523\n",
      "4.57677078247\n",
      "0.213597774506\n",
      "0.188228130341\n"
     ]
    }
   ],
   "source": [
    "## Cleaning of the data\n",
    "\n",
    "from bs4 import Comment, BeautifulSoup\n",
    "import re\n",
    "import time\n",
    "import requests\n",
    "import nltk\n",
    "from nltk.corpus import stopwords # Import thu\n",
    "\n",
    "import numpy as np  # a conventional alias\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "BOA= ['https://www.sec.gov/Archives/edgar/data/70858/000007085815000008/bac-12312014x10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000007085812000155/bac-12312011x10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000007085813000097/bac-12312012x10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000007085812000155/bac-12312011x10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000095012311018743/g25571e10vk.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312510041666/d10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312509041126/d10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312508041665/d10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312507042036/d10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312506056053/d10k.htm','https://www.sec.gov/Archives/edgar/data/70858/000119312505039878/d10k.htm']\n",
    "JPM=['https://www.sec.gov/Archives/edgar/data/19617/000001961715000272/corp10k2014.htm','https://www.sec.gov/Archives/edgar/data/19617/000001961714000289/corp10k2013.htm','https://www.sec.gov/Archives/edgar/data/19617/000001961713000221/corp10k2012.htm','https://www.sec.gov/Archives/edgar/data/19617/000001961712000163/corp10k2011.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012311019773/y86143e10vk.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012310016029/e82150e10vk.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012309003840/y74757e10vk.htm','https://www.sec.gov/Archives/edgar/data/19617/000119312508043536/d10k.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012307003015/y30834e10vk.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012306002875/y17599e10vk.htm','https://www.sec.gov/Archives/edgar/data/19617/000095012305002539/y05475e10vk.htm']\n",
    "CITI=['https://www.sec.gov/Archives/edgar/data/831001/000083100115000043/c-12312014x10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000110465914015152/a14-3681_610k.htm','https://www.sec.gov/Archives/edgar/data/831001/000120677413000852/citigroup_10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000120677412000799/citigroup_10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000120677411000316/citigroup_10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000120677410000406/citi_10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000119312509041237/d10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000119312508036445/d10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000119312507038505/d10k.htm','https://www.sec.gov/Archives/edgar/data/831001/000104746906002377/a2167745z10-k.htm','https://www.sec.gov/Archives/edgar/data/831001/000104746905004988/a2152655z10-k.htm']\n",
    "urls=BOA + JPM + CITI\n",
    "    \n",
    "def review_to_words(cleaned):\n",
    "\n",
    "    #\n",
    "    # 1. Remove non-letters        \n",
    "    letters_only = re.sub(\"[^a-zA-Z]\", \" \", cleaned) \n",
    "    \n",
    "    # 2. delete words of 1 or 2 characters\n",
    "    shortword = re.compile(r'\\W*\\b\\w{1,2}\\b')\n",
    "    deletesmallwords =shortword.sub('', letters_only)\n",
    "    \n",
    "    #\n",
    "    # 3. Convert to lower case, split into individual words\n",
    "    words = deletesmallwords.lower().split()                             \n",
    "    #\n",
    "    # 4. searching a set is much faster than searching\n",
    "    #   a list, so convert the stop words to a set\n",
    "    stops = set(stopwords.words(\"english\"))                  \n",
    "    # \n",
    "    # 5. Remove stop words\n",
    "    meaningful_words = [w for w in words if not w in stops]   \n",
    "    #\n",
    "    # 6. Join the words back into one string separated by space, \n",
    "    # and return the result.\n",
    "    return( \" \".join( meaningful_words ))   \n",
    "\n",
    "\n",
    "def visible(element):\n",
    "    if isinstance(element, Comment):\n",
    "        return False\n",
    "    if element.parent.name in ['style', 'script', '[document]', 'head', 'title', 'noscript']:\n",
    "        return False\n",
    "    elif re.match('<!--.*-->', element.string):\n",
    "        return False\n",
    "    return True\n",
    "\n",
    "\n",
    "\n",
    "def extract_visible_text(html):\n",
    "\n",
    "    try:\n",
    "        soup = BeautifulSoup(html, 'lxml')\n",
    "    except UnicodeEncodeError:\n",
    "        soup = BeautifulSoup(\"\")\n",
    "\n",
    "    # These two lines get rid of the comments in the html.\n",
    "    comments = soup.findAll(text=lambda text: isinstance(text, Comment))\n",
    "    [comment.extract() for comment in comments]\n",
    "\n",
    "    # as well as those spesky noscript tags:\n",
    "    noscripts = soup.findAll('noscript')\n",
    "    [noscript.extract() for noscript in noscripts]\n",
    "    # Next, the text is obtained, and filtered out for only the visible parts\n",
    "\n",
    "    texts = soup.findAll(text=True)\n",
    "    visible_texts = filter(visible, texts)\n",
    "\n",
    "    return visible_texts\n",
    "\n",
    "\n",
    "def cleanning(url):\n",
    "\n",
    "    t0 = time.time()\n",
    "    new_text = requests.get(url).text\n",
    "    print time.time()-t0\n",
    "    \n",
    "    t0 = time.time()\n",
    "    cleaned_text = extract_visible_text(new_text)\n",
    "    print time.time()-t0\n",
    "    t0 = time.time()\n",
    "    t = []\n",
    "    for line in cleaned_text:\n",
    "        t.append(' '.join(re.sub(r\"\\s\", ' ', line).split()))\n",
    "    cleaned_text = ' '.join(t)\n",
    "    print time.time()-t0\n",
    "\n",
    "    t0 = time.time()\n",
    "    cleaned_text = review_to_words(cleaned_text)\n",
    "    print time.time()-t0\n",
    "    \n",
    "    return cleaned_text\n",
    "\n",
    "\n",
    "\n",
    "num_reviews = 33\n",
    "clean_annuals = []\n",
    "for i in xrange( 0, num_reviews ):\n",
    "    # If the index is evenly divisible by 4, print a message\n",
    "    if( (i+1)%4 == 0 ):\n",
    "        print \"Review %d of %d\\n\" % ( i+1, num_reviews )                                                                    \n",
    "    clean_annuals.append(cleanning(urls[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vectorizer = CountVectorizer(input='content')\n",
    "dtm = vectorizer.fit_transform(clean_annuals)  # a sparse matrix\n",
    "vocab = vectorizer.get_feature_names()\n",
    "dtm = dtm.toarray()  # convert to a regular array\n",
    "vocab = np.array(vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([u'aaa', u'aadvantage', u'aal', ..., u'zubrow', u'zuni', u'zurich'], \n",
       "      dtype='<U22')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(33, 10845)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dtm.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.  ,  0.02,  0.01, ...,  0.21,  0.21,  0.2 ],\n",
       "       [ 0.02, -0.  ,  0.01, ...,  0.22,  0.22,  0.2 ],\n",
       "       [ 0.01,  0.01,  0.  , ...,  0.22,  0.22,  0.2 ],\n",
       "       ..., \n",
       "       [ 0.21,  0.22,  0.22, ...,  0.  ,  0.01,  0.04],\n",
       "       [ 0.21,  0.22,  0.22, ...,  0.01,  0.  ,  0.03],\n",
       "       [ 0.2 ,  0.2 ,  0.2 , ...,  0.04,  0.03, -0.  ]])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "#change range in dtm to select one particular company\n",
    "dist = 1 - cosine_similarity(dtm)\n",
    "np.round(dist, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import os  # for os.path.basename\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.manifold import MDS\n",
    "\n",
    "# two components as we're plotting points in a two-dimensional plane\n",
    "# \"precomputed\" because we provide a distance matrix\n",
    "# we will also specify `random_state` so the plot is reproducible.\n",
    "mds = MDS(n_components=2, dissimilarity=\"precomputed\", random_state=1)\n",
    "\n",
    "pos = mds.fit_transform(dist)  # shape (n_components, n_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEACAYAAACpoOGTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nliCMMcZ4ZAnCGGOMR5YgjDHGePT/qjrc4fiGsToAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117f727d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xs, ys = pos[:, 0], pos[:, 1]\n",
    "\n",
    "# short versions of filenames:\n",
    "#names=['BOA-2014','BOA-2013','BOA-2012','BOA-2011','BOA-2010','BOA-2009','BOA-2008','BOA-2007','BOA-2006','BOA-2005','BOA-2004', 'JPM-2014','JPM-2013','JPM-2012','JPM-2011','JPM-2010','JPM-2009','JPM-2008','JPM-2007','JPM-2006','JPM-2005','JPM-2004','CITI-2014','CITI-2013','CITI-2012','CITI-2011','CITI-2010','CITI-2009','CITI-2008','CITI-2007','CITI-2006','CITI-2005','CITI-2004']\n",
    "names=['BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA', 'JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI']\n",
    "#names=[ '2014','2013','2012','2011','2010','2009','2008','2007','2006','2005','2004']\n",
    "# color-blind-friendly palette\n",
    "for x, y, name in zip(xs, ys, names):\n",
    "     color = 'orange' if \"CITI\" in name  else 'skyblue'\n",
    "     plt.scatter(x, y, c=color)\n",
    "     plt.text(x, y, name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Topic Model preparation and calculation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3049"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def split_text(filename, n_words):\n",
    "     \"\"\"Split a text into chunks approximately `n_words` words in length.\"\"\"\n",
    "     # input = open(filename, 'r')\n",
    "     # words = input.read().split(' ')\n",
    "     words =filename.split(' ')   \n",
    "     # input.close()\n",
    "     chunks = []\n",
    "     current_chunk_words = []\n",
    "     current_chunk_word_count = 0\n",
    "     for word in words:\n",
    "         current_chunk_words.append(word)\n",
    "         current_chunk_word_count += 1\n",
    "         if current_chunk_word_count == n_words:\n",
    "            chunks.append(' '.join(current_chunk_words))\n",
    "            current_chunk_words = []\n",
    "            current_chunk_word_count = 0\n",
    "     chunks.append(' '.join(current_chunk_words) )\n",
    "     return chunks\n",
    "    \n",
    "chunk_length = 1000\n",
    "chunks = []\n",
    "num_reviews = 33\n",
    "chunks = []\n",
    "\n",
    "for i in xrange( 0, num_reviews ):\n",
    "     chunk_counter = 0\n",
    "     texts = split_text(clean_annuals[i], chunk_length)\n",
    "     for text in texts:\n",
    "         chunk = {'text': text, 'number': chunk_counter, 'filename': i}\n",
    "         chunks.append(chunk)\n",
    "         chunk_counter += 1\n",
    "\n",
    "\n",
    "# we started with 33 files\n",
    "# and now we have:\n",
    "len(chunks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "# Now we create the sparse matrix\n",
    "vectorizer44 = CountVectorizer(input='content',ngram_range=(1, 1),min_df=30, max_df=0.70)\n",
    "# max_df at 70% just eliminates 33 frequent words\n",
    "# i choose min_df=30 because it is just 1% of 3049 chunks, \n",
    "# deletes 65% of unfrequent words or 7000 words.\n",
    "dtm44 = vectorizer44.fit_transform([c['text'] for c in chunks])\n",
    "vocab44 = np.array(vectorizer44.get_feature_names())\n",
    "dtm44 = dtm44.toarray()  # convert to a regular array\n",
    "vocab44 = np.array(vocab44)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3848,)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab44.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3049, 3848)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dtm44.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting LDA models with tf \n",
      "done in 301.228s.\n",
      "\n",
      "Topics in LDA model:\n",
      "Topic #0:\n",
      "tax accounting sfas fair company goodwill earnings reporting impairment statements taxes deferred fasb changes january effective guidance liabilities recorded intangible statement recognized entity adoption impact acquisition operations method issued effect\n",
      "Topic #1:\n",
      "firm chase jpmorgan report annual note internal statements pages reporting control accounting see sheets ended discussion results used private page respectively includes provide upon activities periods repurchase material operations heritage\n",
      "Topic #2:\n",
      "mortgage repurchase servicing securitization claims representations warranties corporation securitizations loss settlement gses mortgages assumptions private foreclosure liability msrs label investors adverse residential estimated future possible home whole unit principal first\n",
      "Topic #3:\n",
      "court district claims actions filed action corporation defendants class plaintiffs bank new litigation york complaint citigroup state damages states pending motion federal proceedings matters inc united dismiss settlement southern alleged\n",
      "Topic #4:\n",
      "plans stock plan benefit pension million awards options expense rate compensation years year postretirement shares benefits employee expected employees cost defined corporation common average date asset granted return weighted tax\n",
      "Topic #5:\n",
      "bank subsidiaries federal act banking capital company holding companies fdic banks subject reserve activities services requirements regulatory institutions institution regulation exchange depository required subsidiary new state global operations rule linked\n",
      "Topic #6:\n",
      "capital corporation bank america stock modifications home mortgage new term fair regulatory liquidity exchange portfolio fsp consumer requirements subsidiaries tdrs could required funding debt federal tier company note programs agencies\n",
      "Topic #7:\n",
      "average rate trading var rates basis deposits liabilities portfolio changes common revenue change impact measures model used shareholders term bearing activities stress testing results expense positions year goodwill reported loss\n",
      "Topic #8:\n",
      "could risks businesses operational regulatory significant operations affect economic results result ability adversely services impact subject control policies factors markets compliance changes conditions addition board well future legal products potential\n",
      "Topic #9:\n",
      "commercial percent consumer portfolio million leases allowance real nonperforming estate lease charge offs home fair mortgage residential corporation card outstanding includes america dollars commitments accounted unfunded properties option countrywide millions\n",
      "Topic #10:\n",
      "commitments amount company guarantees agreements transactions collateral corporation obligations respectively lending million derivative party future contracts letters loss maximum amounts protection payments sold sheet investment third event contract contractual instruments\n",
      "Topic #11:\n",
      "exposure liquidity funding ratings derivative collateral derivatives exposures rating portfolio debt counterparty term country protection default commitments markets long cross border receivables industry additional purchased grade amount notional senior counterparties\n",
      "Topic #12:\n",
      "million increased banking revenue higher services expense investment offset primarily increase lower driven partially compared global growth impact noninterest results card average deposits tax provision products decreased consumer portfolio corporate\n",
      "Topic #13:\n",
      "year revenue prior expense ended annual charge report allowance million pages fees provision card respectively chase mutual end washington noninterest millions selected average results offs compared mortgage discussion jpmorgan period\n",
      "Topic #14:\n",
      "citigroup citi company operations year approximately loss dollars million consumer revenues corporate sale cards foreign change quarter holdings taxes north discontinued america offices continuing citibank reserve statements first citicorp banking\n",
      "Topic #15:\n",
      "incorporated reference exhibit form stock registrant dated company bank report file plan january filed corporation york executive new amendment officer february agreement thereto exchange effective pursuant indenture inc amended october\n",
      "Topic #16:\n",
      "allowance consumer portfolio mortgage card impaired past home days principal real estate rate modified current loss pci residential estimated payment borrower expected collateral modification tdrs recorded delinquency fair rates default\n",
      "Topic #17:\n",
      "capital tier basel common regulatory ratio bank iii requirements rules banking preferred weighted ratios stock federal leverage reserve minimum trust january qualifying final required well guidelines percent tax amount issued\n",
      "Topic #18:\n",
      "stock corporation preferred debt trust company term issued common notes shares rate interests vies trusts million commercial paper activities vie long share liabilities series conduits funds note held per entities\n",
      "Topic #19:\n",
      "fair debt trading liabilities derivative level instruments derivatives rate mortgage changes foreign gains valuation hedges investments contracts hedge account recorded exchange backed unrealized sale available corporation held agreements rates cost\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Find 20 topics using LDA\n",
    "\n",
    "from __future__ import print_function\n",
    "from time import time\n",
    "\n",
    "#In LDA, the time complexity is proportional to (n_samples * iterations).\n",
    "#import sklearn\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer\n",
    "from sklearn.decomposition import LatentDirichletAllocation\n",
    "\n",
    "def print_top_words(model, feature_names, n_top_words):\n",
    "    for topic_idx, topic in enumerate(model.components_):\n",
    "        print(\"Topic #%d:\" % topic_idx)\n",
    "        print(\" \".join([feature_names[i]\n",
    "                        for i in topic.argsort()[:-n_top_words - 1:-1]]))\n",
    "    print()\n",
    "\n",
    "n_topics=20\n",
    "max_iter=100\n",
    "n_top_words=30\n",
    "print(\"Fitting LDA models with tf \")\n",
    "lda = LatentDirichletAllocation(n_topics=n_topics, max_iter=max_iter,\n",
    "                                learning_method='online', learning_offset=50.,\n",
    "                                random_state=0)\n",
    "t0 = time()\n",
    "word_topic=lda.fit(dtm44)\n",
    "print(\"done in %0.3fs.\" % (time() - t0))\n",
    "\n",
    "print(\"\\nTopics in LDA model:\")\n",
    "#tf_feature_names = tf_vectorizer.get_feature_names()\n",
    "print_top_words(lda, vocab44, n_top_words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3049, 21)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Find Average topics per document (transform chunks to 33 docs)\n",
    "\n",
    "doc_topic= lda.transform(dtm44)\n",
    "doc_topic.shape\n",
    "\n",
    "lista=()\n",
    "num_chunks = 3049\n",
    "for j in xrange(0,33):\n",
    "    for i in xrange(0, num_chunks ):\n",
    "        if chunks[i]['filename']==j:\n",
    "            lista = lista + (j,)\n",
    "            \n",
    "lista=np.asarray(lista)[...,None]\n",
    "all_data = np.append(doc_topic, lista, 1)\n",
    "all_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Find average\n",
    "results=[]\n",
    "for x in sorted(np.unique(all_data[...,20])):\n",
    "    results.append([np.average(all_data[np.where(all_data[...,20]==x)][...,0]), \n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,1]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,2]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,3]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,4]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,5]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,6]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,7]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,8]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,9]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,10]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,11]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,12]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,13]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,14]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,15]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,16]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,17]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,18]),\n",
    "                    np.average(all_data[np.where(all_data[...,20]==x)][...,19]),x])\n",
    "\n",
    "results=np.asarray(results)\n",
    "results.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(33, 21)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Normalize results\n",
    "results2 = np.round(results[:,0:20] / np.sum(results[:,0:20], axis=1, keepdims=True), decimals=4)   \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Grap evolution of share of topics per year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZEAAAEZCAYAAABWwhjiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1198ee290>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# Change numbertop variable to do the graph for an specific topic\n",
    "numbertop=19\n",
    "doctopic = results2[(10,9,8,7,6,5,4,3,2,1,0),numbertop]\n",
    "doctopic2 = results2[(21,20,19,18,17,16,15,14,13,12,11),numbertop]\n",
    "doctopic3 = results2[(32,31,30,29,28,27,26,25,24,23,22),numbertop]\n",
    "docnames=[2004,2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014] \n",
    "plt.xlabel('Year')\n",
    "plt.ylabel('Share')\n",
    "plt.title('Share of Topic #%d' % numbertop)\n",
    "topic_labels=['Bank of America', 'J.P. Morgan', 'Citigroup']\n",
    "\n",
    "l2=plt.plot(docnames, doctopic, 'c-', docnames, doctopic2, 'm-', docnames, doctopic3, 'b-')\n",
    "plt.legend(l2, topic_labels, loc='upper left', shadow=False,  prop={'size':10},frameon=False)\n",
    "plt.ylim((0))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Graph distance between documents using topic share"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.  ,  0.03,  0.02, ...,  0.38,  0.39,  0.39],\n",
       "       [ 0.03,  0.  ,  0.01, ...,  0.39,  0.39,  0.39],\n",
       "       [ 0.02,  0.01, -0.  , ...,  0.4 ,  0.4 ,  0.4 ],\n",
       "       ..., \n",
       "       [ 0.38,  0.39,  0.4 , ...,  0.  ,  0.01,  0.02],\n",
       "       [ 0.39,  0.39,  0.4 , ...,  0.01,  0.  ,  0.02],\n",
       "       [ 0.39,  0.39,  0.4 , ...,  0.02,  0.02,  0.  ]])"
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "#change range in dtm to select one particular company\n",
    "\n",
    "#dist2 = 1 - cosine_similarity(results2[:,(0,2,3,4,5,7,8,10,11,12,15,16,17,18,19)])\n",
    "dist2 = 1 - cosine_similarity(results2)\n",
    "np.round(dist2, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import os  # for os.path.basename\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.manifold import MDS\n",
    "\n",
    "# two components as we're plotting points in a two-dimensional plane\n",
    "# \"precomputed\" because we provide a distance matrix\n",
    "# we will also specify `random_state` so the plot is reproducible.\n",
    "mds = MDS(n_components=2, dissimilarity=\"precomputed\", random_state=1)\n",
    "\n",
    "pos = mds.fit_transform(dist2)  # shape (n_components, n_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEACAYAAABcXmojAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114dfa410>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xs, ys = pos[:, 0], pos[:, 1]\n",
    "\n",
    "# short versions of filenames:\n",
    "#names=['BOA-2014','BOA-2013','BOA-2012','BOA-2011','BOA-2010','BOA-2009','BOA-2008','BOA-2007','BOA-2006','BOA-2005','BOA-2004', 'JPM-2014','JPM-2013','JPM-2012','JPM-2011','JPM-2010','JPM-2009','JPM-2008','JPM-2007','JPM-2006','JPM-2005','JPM-2004','CITI-2014','CITI-2013','CITI-2012','CITI-2011','CITI-2010','CITI-2009','CITI-2008','CITI-2007','CITI-2006','CITI-2005','CITI-2004']\n",
    "names=['BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA','BOA', 'JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','JPM','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI','CITI']\n",
    "#names=[ '2014','2013','2012','2011','2010','2009','2008','2007','2006','2005','2004','2014','2013','2012','2011','2010','2009','2008','2007','2006','2005','2004','2014','2013','2012','2011','2010','2009','2008','2007','2006','2005','2004']\n",
    "# color-blind-friendly palette\n",
    "for x, y, name in zip(xs, ys, names):\n",
    "     color = 'orange' if \"CITI\" in name  else 'skyblue'\n",
    "     plt.scatter(x, y, c=color)\n",
    "     plt.text(x, y, name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}