02. Clustering.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](http://www.upm.es/sfs/Rectorado/Gabinete%20del%20Rector/Logos/UPM/EscPolitecnica/EscUpmPolit_p.gif \"UPM\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Trabajo final SITC\n",
    "## Análisis de sentimientos en Twitter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Departamento de Ingeniería de Sistemas Telemáticos. Universidad Politécnica de Madrid."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Realizado por:\n",
    "- Juan Bermudo Mera\n",
    "- Margarita Bolívar Jiménez\n",
    "- Lourdes Fernández Nieto\n",
    "- Ramón Pérez Hernández\n",
    "\n",
    "© 2017"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clustering aplicado sobre el fichero de tweets"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tabla de contenidos\n",
    "\n",
    "* [Preparación del dataset](#1.-Preparación-del-dataset)\n",
    "\t* [Apertura del fichero Excel y conversión a dataframe](#Apertura-del-fichero-Excel-y-conversión-a-dataframe)\n",
    "    * [Tokenización y stemming](#Tokenización-y-stemming)\n",
    "* [Clustering con K-means](#2.-Clustering-con-K-means)\n",
    "    * [K-means con unigramas](#K-means-con-unigramas)\n",
    "    * [K-means con unigramas y bigramas](#K-means-con-unigramas-y-bigramas)\n",
    "    * [K-means con unigramas, bigramas y trigramas](#K-means-con-unigramas,-bigramas-y-trigramas)\n",
    "    * [Asignación de temas/topics](#Asignación-de-temas/topics)\n",
    "* [Topic modeling con LDA](#3.-Topic-modeling-con-LDA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Se importan todas las librerías necesarias (Algunas necesitan instalación con: pip install <nombre_paquete>)\n",
    "\n",
    "# Procesado del fichero csv\n",
    "import pandas as pd\n",
    "\n",
    "# Tokenización y stemming\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "from string import punctuation\n",
    "from nltk.stem import SnowballStemmer\n",
    "from nltk.tokenize import TweetTokenizer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "# Clustering\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "# LDA\n",
    "from gensim import corpora, models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Preparación del dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### Apertura del fichero Excel y conversión a dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tweets = pd.read_excel('ficheros/Preprocesados/tweets_spainGeo.xlsx',header=0,encoding='iso8859_15')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Regalo para socios: 'Talento a la fuga', el li...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>La recuperación del ‘ladrillo’ dispara el 144,...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TELEVISIÓN SERIES - Cuenta atrás para conocer ...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Nuestra taza de hoy es un homenaje a uno de lo...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>@MCadepe nos vemos el 19 de mayooo!!!</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             content  Latitude  Longitude\n",
       "0  Regalo para socios: 'Talento a la fuga', el li...    40.416     -3.703\n",
       "1  La recuperación del ‘ladrillo’ dispara el 144,...    40.416     -3.703\n",
       "2  TELEVISIÓN SERIES - Cuenta atrás para conocer ...    40.416     -3.703\n",
       "3  Nuestra taza de hoy es un homenaje a uno de lo...    40.416     -3.703\n",
       "4              @MCadepe nos vemos el 19 de mayooo!!!    40.416     -3.703"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Se comprueba que se ha cargado correctamente\n",
    "tweets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26363, 3)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### Tokenización y stemming"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/monphdez/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "# Se descargan las palabras de parada (stopwords) en español \n",
    "nltk.download(\"stopwords\")\n",
    "spanish_stopwords = stopwords.words('spanish')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Se obtienen los signos de puntuación de la librería punctuation\n",
    "non_words = list(punctuation)\n",
    "\n",
    "# Añadimos signos de puntuación que se usan en español\n",
    "non_words.extend(['¿', '¡'])\n",
    "non_words.extend(map(str,range(10)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Se definen las funciones para realizar la tokenización y el stemming\n",
    "stemmer = SnowballStemmer('spanish')\n",
    "tknzr = TweetTokenizer(strip_handles=True, reduce_len=True)\n",
    "\n",
    "def stem_tokens(tokens, stemmer):\n",
    "    stemmed = []\n",
    "    for item in tokens:\n",
    "        stemmed.append(stemmer.stem(item))\n",
    "    return stemmed\n",
    "\n",
    "def tokenize(text):\n",
    "    # Eliminamos lo que no sean palabras\n",
    "    text = ''.join([c for c in text if c not in non_words])\n",
    "    # Tokenización\n",
    "    tokens = tknzr.tokenize(text)\n",
    "\n",
    "    # Stemming\n",
    "    try:\n",
    "        stems = stem_tokens(tokens, stemmer)\n",
    "    except Exception as e:\n",
    "        print(e)\n",
    "        print(text)\n",
    "        stems = ['']\n",
    "    return stems"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Clustering con K-means"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Definición de la función que realiza el clustering con K-means y muestra los resultados\n",
    "def kmeans_function(final_cluster, mx_tfidf, tweets, vectorizer, text, r_state):\n",
    "    for num_clusters in range(5,final_cluster):\n",
    "        km = KMeans(n_clusters=num_clusters, random_state=r_state)\n",
    "        km.fit(mx_tfidf)  \n",
    "        clusters = km.labels_.tolist()  \n",
    "        tweets[\"Cluster_\" + text + \"_\" + str(num_clusters)] = clusters\n",
    "        print('Número de clusters: %d' % (num_clusters))\n",
    "        print(tweets[\"Cluster_\" + text + \"_\" + str(num_clusters)].value_counts())\n",
    "\n",
    "        # Impresión de palabras más frecuentes por grupo.\n",
    "        order_centroids = km.cluster_centers_.argsort()[:, ::-1]\n",
    "        for i in range(num_clusters):\n",
    "            print(\"Cluster {} : Palabras más frecuentes :\".format(i))\n",
    "            for ind in order_centroids[i, :10]: \n",
    "                print(' %s' % vectorizer.get_feature_names()[ind])\n",
    "            print('')\n",
    "        print('')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Extraemos los tweets y los guardamos en una lista\n",
    "tweets_text = tweets[\"content\"].tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### K-means con unigramas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Convertimos la lista en la matriz Tf-idf\n",
    "vectorizer_unig = TfidfVectorizer(\n",
    "                max_df = 0.95,\n",
    "                max_features = 30000,\n",
    "                analyzer = 'word',\n",
    "                tokenizer = tokenize,\n",
    "                lowercase = True,\n",
    "                use_idf = True,\n",
    "                stop_words = spanish_stopwords,\n",
    "                ngram_range=(1,1))\n",
    "mx_tfidf_unig = vectorizer_unig.fit_transform(tweets_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['abrel', 'abren', 'abrepuert', 'abres', 'abri', 'abriend', 'abriendomadr', 'abriendopuert', 'abrig', 'abril']\n"
     ]
    }
   ],
   "source": [
    "# Palabras más frecuentes con las condiciones marcadas (pequeño subconjunto; sólo se verá \n",
    "# aquí para mostrar que se han extraido correctamente las features)\n",
    "print(vectorizer_unig.get_feature_names()[100:110])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26363, 30000)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mx_tfidf_unig.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de clusters: 5\n",
      "0    17365\n",
      "4     6714\n",
      "3      974\n",
      "2      711\n",
      "1      599\n",
      "Name: Cluster_unig_5, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " in\n",
      " mas\n",
      " spain\n",
      " at\n",
      " com\n",
      " tod\n",
      " si\n",
      " im\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " seguidor\n",
      " unfollowers\n",
      " tuitutil\n",
      " nuev\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " tod\n",
      " hoy\n",
      " madr\n",
      " com\n",
      " si\n",
      " dia\n",
      " hac\n",
      "\n",
      "\n",
      "Número de clusters: 6\n",
      "1    15172\n",
      "2     6753\n",
      "0     2512\n",
      "4      712\n",
      "5      615\n",
      "3      599\n",
      "Name: Cluster_unig_6, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " madr\n",
      " tod\n",
      " mejor\n",
      " mas\n",
      " dia\n",
      " com\n",
      " hac\n",
      " hoy\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " mas\n",
      " spain\n",
      " si\n",
      " com\n",
      " via\n",
      " tod\n",
      " nuev\n",
      " mejor\n",
      " graci\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " tod\n",
      " mas\n",
      " hoy\n",
      " madr\n",
      " com\n",
      " dia\n",
      " si\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "\n",
      "Número de clusters: 7\n",
      "0    16000\n",
      "3     6340\n",
      "1     1115\n",
      "6      837\n",
      "4      761\n",
      "5      711\n",
      "2      599\n",
      "Name: Cluster_unig_7, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " mas\n",
      " com\n",
      " si\n",
      " via\n",
      " spain\n",
      " tod\n",
      " mejor\n",
      " dia\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " …\n",
      " madr\n",
      " viern\n",
      " jardin\n",
      " dia\n",
      " hoy\n",
      " feliz\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " tod\n",
      " madr\n",
      " hoy\n",
      " com\n",
      " si\n",
      " dia\n",
      " hac\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " in\n",
      " madr\n",
      " at\n",
      " im\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " graci\n",
      " tuitutil\n",
      " unfollowers\n",
      " seguidor\n",
      " nuev\n",
      " …\n",
      " tod\n",
      " segu\n",
      " much\n",
      " devuelv\n",
      "\n",
      "\n",
      "Número de clusters: 8\n",
      "6    12307\n",
      "5     5441\n",
      "0     3012\n",
      "4     2558\n",
      "1     1142\n",
      "7      711\n",
      "3      599\n",
      "2      593\n",
      "Name: Cluster_unig_8, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " buen\n",
      " dias\n",
      " …\n",
      " tod\n",
      " viern\n",
      " hoy\n",
      " dia\n",
      " mejor\n",
      " com\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver\n",
      " pued\n",
      " mas\n",
      " hac\n",
      " tien\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " real\n",
      " mas\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " …\n",
      " tod\n",
      " mas\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " graci\n",
      " hac\n",
      " vid\n",
      " nuestr\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " tod\n",
      " nuev\n",
      " graci\n",
      " via\n",
      " per\n",
      " mejor\n",
      " dia\n",
      " hac\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "\n",
      "Número de clusters: 9\n",
      "1    12826\n",
      "6     6275\n",
      "5     2405\n",
      "3     1632\n",
      "2     1018\n",
      "8      711\n",
      "4      599\n",
      "0      593\n",
      "7      304\n",
      "Name: Cluster_unig_9, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " com\n",
      " si\n",
      " via\n",
      " tod\n",
      " graci\n",
      " per\n",
      " hac\n",
      " mejor\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " hoy\n",
      " …\n",
      " dia\n",
      " buen\n",
      " par\n",
      " toc\n",
      " dias\n",
      " madr\n",
      " com\n",
      " tod\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " sol\n",
      " ser\n",
      " tien\n",
      " …\n",
      " si\n",
      " par\n",
      " tod\n",
      " pued\n",
      " mas\n",
      " per\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " via\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " par\n",
      " tod\n",
      " mas\n",
      " com\n",
      " si\n",
      " dia\n",
      " mejor\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " whatsapp\n",
      " vip\n",
      " sab\n",
      " tartuf\n",
      " mejor\n",
      " show\n",
      " sal\n",
      " info\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "\n",
      "Número de clusters: 10\n",
      "5    14210\n",
      "6     5350\n",
      "4     1398\n",
      "0     1349\n",
      "7     1127\n",
      "8      759\n",
      "3      755\n",
      "9      711\n",
      "2      599\n",
      "1      105\n",
      "Name: Cluster_unig_10, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " tod\n",
      " list\n",
      " …\n",
      " reserv\n",
      " sab\n",
      " mejor\n",
      " vip\n",
      " whatsapp\n",
      " par\n",
      " madr\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much\n",
      " segu\n",
      " tod\n",
      " devuelv\n",
      " follow\n",
      " brev\n",
      " tuitutil\n",
      " buen\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " par\n",
      " much\n",
      " cad\n",
      " pued\n",
      " men\n",
      " person\n",
      " tod\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " buen\n",
      " spain\n",
      " via\n",
      " com\n",
      " dias\n",
      " dia\n",
      " hoy\n",
      " españ\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " madr\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " buen\n",
      " nuestr\n",
      " nuev\n",
      " hac\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " par\n",
      " ver\n",
      " pued\n",
      " hac\n",
      " tien\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " in\n",
      " madr\n",
      " at\n",
      " im\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "\n",
      "Número de clusters: 11\n",
      "9     12116\n",
      "6      5282\n",
      "7      2236\n",
      "3      2176\n",
      "10     1251\n",
      "8       710\n",
      "5       645\n",
      "1       599\n",
      "4       591\n",
      "2       481\n",
      "0       276\n",
      "Name: Cluster_unig_11, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " vip\n",
      " whatsapp\n",
      " sab\n",
      " mejor\n",
      " tartuf\n",
      " show\n",
      " sal\n",
      " new\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " bien\n",
      " segu\n",
      " graci\n",
      " …\n",
      " devuelv\n",
      " follow\n",
      " brev\n",
      " tuitutil\n",
      " si\n",
      " pas\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " mas\n",
      " dia\n",
      " hac\n",
      " com\n",
      " mejor\n",
      " hoy\n",
      " nuestr\n",
      " madr\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " viern\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      " madr\n",
      " felizviern\n",
      " fin\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " mas\n",
      " hoy\n",
      " com\n",
      " si\n",
      " madr\n",
      " dia\n",
      " mejor\n",
      " nuestr\n",
      " hac\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " sol\n",
      " plaz\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " mas\n",
      " si\n",
      " com\n",
      " nuev\n",
      " via\n",
      " per\n",
      " dia\n",
      " hac\n",
      " graci\n",
      " quier\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " graci\n",
      " par\n",
      " per\n",
      " madr\n",
      " si\n",
      " mas\n",
      " com\n",
      " mund\n",
      "\n",
      "\n",
      "Número de clusters: 12\n",
      "4     11075\n",
      "2      4548\n",
      "7      2068\n",
      "3      1977\n",
      "5      1366\n",
      "8      1334\n",
      "11     1101\n",
      "9       886\n",
      "10      711\n",
      "0       599\n",
      "1       593\n",
      "6       105\n",
      "Name: Cluster_unig_12, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " hoy\n",
      " com\n",
      " madr\n",
      " buen\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      " hac\n",
      " graci\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " hac\n",
      " com\n",
      " mejor\n",
      " hoy\n",
      " nuestr\n",
      " buen\n",
      " via\n",
      " madr\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " buen\n",
      " com\n",
      " dias\n",
      " via\n",
      " graci\n",
      " per\n",
      " hac\n",
      " mejor\n",
      " hoy\n",
      " quier\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much\n",
      " par\n",
      " pued\n",
      " person\n",
      " mejor\n",
      " tod\n",
      " men\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " via\n",
      " palaci\n",
      " gran\n",
      " buen\n",
      " sol\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " tod\n",
      " list\n",
      " …\n",
      " reserv\n",
      " sab\n",
      " vip\n",
      " mejor\n",
      " whatsapp\n",
      " tartuf\n",
      " par\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " buen\n",
      " tierr\n",
      " feliz\n",
      " madr\n",
      " par\n",
      " com\n",
      " libr\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver\n",
      " pued\n",
      " hac\n",
      " par\n",
      " tien\n",
      "\n",
      "\n",
      "Número de clusters: 13\n",
      "0     11127\n",
      "11     4514\n",
      "1      2278\n",
      "4      1991\n",
      "2      1202\n",
      "12     1100\n",
      "3       914\n",
      "9       710\n",
      "8       703\n",
      "5       599\n",
      "7       591\n",
      "10      529\n",
      "6       105\n",
      "Name: Cluster_unig_13, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " via\n",
      " graci\n",
      " per\n",
      " mejor\n",
      " dia\n",
      " hac\n",
      " quier\n",
      " nuev\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " real\n",
      " sol\n",
      " plaz\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " sab\n",
      " graci\n",
      " par\n",
      " madr\n",
      " mas\n",
      " mejor\n",
      " list\n",
      " per\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " hoy\n",
      " …\n",
      " dia\n",
      " par\n",
      " toc\n",
      " madr\n",
      " com\n",
      " h\n",
      " hac\n",
      " mas\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " mejor\n",
      " mas\n",
      " madr\n",
      " com\n",
      " nuestr\n",
      " dia\n",
      " hac\n",
      " list\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " lleg\n",
      " tiemp\n",
      " …\n",
      " par\n",
      " pas\n",
      " hac\n",
      " mas\n",
      " dia\n",
      " com\n",
      " madr\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " …\n",
      " mas\n",
      " com\n",
      " mejor\n",
      " madr\n",
      " dia\n",
      " nuestr\n",
      " nuev\n",
      " graci\n",
      " hac\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver\n",
      " pued\n",
      " mas\n",
      " hac\n",
      " tien\n",
      "\n",
      "\n",
      "Número de clusters: 14\n",
      "0     12870\n",
      "6      4808\n",
      "3      2055\n",
      "10     1381\n",
      "9      1364\n",
      "13      773\n",
      "12      710\n",
      "7       599\n",
      "11      535\n",
      "4       507\n",
      "8       348\n",
      "2       193\n",
      "1       115\n",
      "5       105\n",
      "Name: Cluster_unig_14, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " com\n",
      " via\n",
      " si\n",
      " españ\n",
      " dia\n",
      " per\n",
      " graci\n",
      " sol\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " adolf\n",
      " suarez\n",
      " madridbaraj\n",
      " mad\n",
      " aeropuert\n",
      " aenaaeropuert\n",
      " in\n",
      " madr\n",
      " im\n",
      " at\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " palaci\n",
      " madr\n",
      " cristal\n",
      " vistalegr\n",
      " real\n",
      " cibel\n",
      " florenceandthemachin\n",
      " deport\n",
      " florenc\n",
      " in\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " madr\n",
      " mejor\n",
      " com\n",
      " dia\n",
      " hac\n",
      " hoy\n",
      " nuestr\n",
      " via\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " cosl\n",
      " españ\n",
      " alarcon\n",
      " pozuel\n",
      " estacion\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " com\n",
      " hoy\n",
      " si\n",
      " dia\n",
      " hac\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " mañan\n",
      " …\n",
      " hoy\n",
      " par\n",
      " madr\n",
      " dia\n",
      " si\n",
      " com\n",
      " buen\n",
      " ser\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " list\n",
      " …\n",
      " reserv\n",
      " sab\n",
      " vip\n",
      " whatsapp\n",
      " mejor\n",
      " par\n",
      " madr\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " par\n",
      " cad\n",
      " si\n",
      " pued\n",
      " mejor\n",
      " person\n",
      " bonit\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " much\n",
      " …\n",
      " graci\n",
      " mas\n",
      " felic\n",
      " tod\n",
      " par\n",
      " mejor\n",
      " gan\n",
      " dia\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      "\n",
      "\n",
      "Número de clusters: 15\n",
      "9     10019\n",
      "8      5442\n",
      "12     2088\n",
      "6      1620\n",
      "7      1386\n",
      "11     1347\n",
      "10      711\n",
      "4       711\n",
      "14      599\n",
      "2       578\n",
      "13      553\n",
      "0       488\n",
      "3       399\n",
      "1       308\n",
      "5       114\n",
      "Name: Cluster_unig_15, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " cosl\n",
      " españ\n",
      " alarcon\n",
      " pozuel\n",
      " estacion\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " list\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " reserv\n",
      " vip\n",
      " whatsapp\n",
      " sab\n",
      " tartuf\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " nuestr\n",
      " …\n",
      " par\n",
      " tod\n",
      " app\n",
      " mejor\n",
      " seseñ\n",
      " hoy\n",
      " disfrut\n",
      " mas\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " ver\n",
      " …\n",
      " si\n",
      " par\n",
      " quier\n",
      " pued\n",
      " com\n",
      " madr\n",
      " tod\n",
      " hac\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " madridbaraj\n",
      " adolf\n",
      " suarez\n",
      " mad\n",
      " aeropuert\n",
      " aenaaeropuert\n",
      " in\n",
      " madr\n",
      " im\n",
      " at\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " com\n",
      " graci\n",
      " …\n",
      " tod\n",
      " par\n",
      " much\n",
      " si\n",
      " segu\n",
      " hac\n",
      " per\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much\n",
      " par\n",
      " pued\n",
      " si\n",
      " mejor\n",
      " hac\n",
      " cad\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " tod\n",
      " si\n",
      " mejor\n",
      " nuev\n",
      " hac\n",
      " per\n",
      " madr\n",
      " sol\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " si\n",
      " via\n",
      " tod\n",
      " per\n",
      " mejor\n",
      " quier\n",
      " hac\n",
      " tien\n",
      " nuev\n",
      " pued\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " felizviern\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " dia\n",
      " hoy\n",
      " …\n",
      " par\n",
      " madr\n",
      " tierr\n",
      " feliz\n",
      " tod\n",
      " libr\n",
      " celebr\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " sol\n",
      " plaz\n",
      " adecc\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " porqu\n",
      " aqu\n",
      " …\n",
      " tod\n",
      " si\n",
      " par\n",
      " ayudamosaqui\n",
      " som\n",
      " ser\n",
      " necesit\n",
      "\n",
      "Cluster 14 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Realización del clustering con k-means\n",
    "kmeans_function(16, mx_tfidf_unig, tweets, vectorizer_unig, \"unig\", 3333)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de clusters: 5\n",
      "0    22282\n",
      "3     2467\n",
      "4      712\n",
      "1      599\n",
      "2      303\n",
      "Name: Cluster_unig2_5, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " buen\n",
      " tod\n",
      " com\n",
      " in\n",
      " si\n",
      " hoy\n",
      " dia\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " whatsapp\n",
      " vip\n",
      " sab\n",
      " tartuf\n",
      " mejor\n",
      " show\n",
      " sal\n",
      " info\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " madr\n",
      " tod\n",
      " mas\n",
      " dia\n",
      " com\n",
      " hac\n",
      " buen\n",
      " hoy\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "\n",
      "Número de clusters: 6\n",
      "4    15143\n",
      "3     6245\n",
      "0     1837\n",
      "2     1551\n",
      "5      988\n",
      "1      599\n",
      "Name: Cluster_unig2_6, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " in\n",
      " at\n",
      " im\n",
      " comun\n",
      " españ\n",
      " aeropuert\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " tod\n",
      " list\n",
      " …\n",
      " reserv\n",
      " sab\n",
      " mejor\n",
      " vip\n",
      " whatsapp\n",
      " par\n",
      " madr\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " nuev\n",
      " madr\n",
      " hoy\n",
      " com\n",
      " si\n",
      " dia\n",
      " hac\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " mas\n",
      " com\n",
      " si\n",
      " via\n",
      " dia\n",
      " per\n",
      " graci\n",
      " hac\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " madr\n",
      " viern\n",
      " …\n",
      " jardin\n",
      " feliz\n",
      " dia\n",
      " hoy\n",
      "\n",
      "\n",
      "Número de clusters: 7\n",
      "0    16127\n",
      "3     6311\n",
      "1     1468\n",
      "5      711\n",
      "6      616\n",
      "4      599\n",
      "2      531\n",
      "Name: Cluster_unig2_7, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " buen\n",
      " spain\n",
      " tod\n",
      " com\n",
      " dias\n",
      " via\n",
      " si\n",
      " mejor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " par\n",
      " much\n",
      " cad\n",
      " pued\n",
      " mejor\n",
      " tod\n",
      " si\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " tien\n",
      " …\n",
      " si\n",
      " aqu\n",
      " tod\n",
      " par\n",
      " mas\n",
      " q\n",
      " porqu\n",
      " cuand\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " tod\n",
      " nuev\n",
      " hoy\n",
      " madr\n",
      " com\n",
      " dia\n",
      " si\n",
      " buen\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "\n",
      "Número de clusters: 8\n",
      "3    19831\n",
      "0     1457\n",
      "2     1110\n",
      "1     1026\n",
      "4      875\n",
      "6      755\n",
      "5      710\n",
      "7      599\n",
      "Name: Cluster_unig2_8, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " tod\n",
      " list\n",
      " …\n",
      " sab\n",
      " reserv\n",
      " mejor\n",
      " whatsapp\n",
      " vip\n",
      " par\n",
      " tartuf\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " …\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " hoy\n",
      " par\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver\n",
      " pued\n",
      " par\n",
      " mas\n",
      " hac\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " par\n",
      " mas\n",
      " com\n",
      " nuev\n",
      " hoy\n",
      " via\n",
      " spain\n",
      " mejor\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " madr\n",
      " tierr\n",
      " par\n",
      " feliz\n",
      " buen\n",
      " mas\n",
      " com\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " in\n",
      " madr\n",
      " at\n",
      " im\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "\n",
      "Número de clusters: 9\n",
      "8    14419\n",
      "0     5412\n",
      "6     1769\n",
      "2     1182\n",
      "4      783\n",
      "5      749\n",
      "7      740\n",
      "1      710\n",
      "3      599\n",
      "Name: Cluster_unig2_9, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " madr\n",
      " hoy\n",
      " dia\n",
      " si\n",
      " mejor\n",
      " nuestr\n",
      " nuev\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " si\n",
      " par\n",
      " madr\n",
      " per\n",
      " dia\n",
      " quier\n",
      " tod\n",
      " hoy\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " mas\n",
      " si\n",
      " falt\n",
      " cuand\n",
      " años\n",
      " per\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " tod\n",
      " graci\n",
      " …\n",
      " par\n",
      " much\n",
      " madr\n",
      " sab\n",
      " segu\n",
      " mas\n",
      " mejor\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " in\n",
      " at\n",
      " madr\n",
      " im\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " mas\n",
      " via\n",
      " spain\n",
      " si\n",
      " nuev\n",
      " mejor\n",
      " dia\n",
      " españ\n",
      "\n",
      "\n",
      "Número de clusters: 10\n",
      "0    12909\n",
      "2     5724\n",
      "5     2232\n",
      "9     1407\n",
      "6     1122\n",
      "8      778\n",
      "7      711\n",
      "1      599\n",
      "4      592\n",
      "3      289\n",
      "Name: Cluster_unig2_10, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " com\n",
      " tod\n",
      " nuev\n",
      " via\n",
      " graci\n",
      " per\n",
      " dia\n",
      " hac\n",
      " hoy\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " tod\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " madr\n",
      " nuestr\n",
      " hac\n",
      " graci\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " whatsapp\n",
      " vip\n",
      " sab\n",
      " tartuf\n",
      " mejor\n",
      " show\n",
      " sal\n",
      " info\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " sol\n",
      " plaz\n",
      " real\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver\n",
      " par\n",
      " pued\n",
      " hac\n",
      " tien\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much\n",
      " par\n",
      " cad\n",
      " pued\n",
      " men\n",
      " person\n",
      " tod\n",
      "\n",
      "\n",
      "Número de clusters: 11\n",
      "0     11410\n",
      "2      4910\n",
      "10     2733\n",
      "8      2323\n",
      "9      1263\n",
      "1      1219\n",
      "4       710\n",
      "7       599\n",
      "3       593\n",
      "5       498\n",
      "6       105\n",
      "Name: Cluster_unig2_11, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " mas\n",
      " si\n",
      " via\n",
      " graci\n",
      " per\n",
      " mejor\n",
      " quier\n",
      " hac\n",
      " dia\n",
      " hoy\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " si\n",
      " par\n",
      " hac\n",
      " per\n",
      " hoy\n",
      " dia\n",
      " quier\n",
      " tod\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " mas\n",
      " hoy\n",
      " si\n",
      " dia\n",
      " madr\n",
      " nuestr\n",
      " mejor\n",
      " hac\n",
      " nuev\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " siempr\n",
      " estadi\n",
      " vicent\n",
      " calderon\n",
      " …\n",
      " santiag\n",
      " bernabeu\n",
      " madr\n",
      " mejor\n",
      " atleti\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " buen\n",
      " sol\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " sab\n",
      " graci\n",
      " par\n",
      " madr\n",
      " mas\n",
      " list\n",
      " mejor\n",
      " dias\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " par\n",
      " buen\n",
      " dias\n",
      " …\n",
      " viern\n",
      " dia\n",
      " hoy\n",
      " mejor\n",
      " mas\n",
      " hac\n",
      "\n",
      "\n",
      "Número de clusters: 12\n",
      "0     10672\n",
      "2      4512\n",
      "10     2576\n",
      "8      2132\n",
      "11     1505\n",
      "9      1242\n",
      "1      1221\n",
      "4       710\n",
      "7       599\n",
      "3       593\n",
      "5       496\n",
      "6       105\n",
      "Name: Cluster_unig2_12, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " si\n",
      " via\n",
      " graci\n",
      " per\n",
      " quier\n",
      " hoy\n",
      " hac\n",
      " dia\n",
      " mejor\n",
      " ser\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " si\n",
      " par\n",
      " hac\n",
      " per\n",
      " hoy\n",
      " dia\n",
      " tod\n",
      " quier\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " hoy\n",
      " si\n",
      " dia\n",
      " madr\n",
      " nuev\n",
      " nuestr\n",
      " hac\n",
      " mejor\n",
      " noch\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " siempr\n",
      " estadi\n",
      " vicent\n",
      " calderon\n",
      " …\n",
      " santiag\n",
      " bernabeu\n",
      " madr\n",
      " mejor\n",
      " atleti\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " via\n",
      " palaci\n",
      " gran\n",
      " buen\n",
      " sol\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " graci\n",
      " sab\n",
      " par\n",
      " madr\n",
      " mejor\n",
      " dias\n",
      " per\n",
      " list\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " par\n",
      " buen\n",
      " dias\n",
      " …\n",
      " viern\n",
      " dia\n",
      " hoy\n",
      " mejor\n",
      " seman\n",
      " hac\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " list\n",
      " reserv\n",
      " mejor\n",
      " whatsapp\n",
      " madr\n",
      " vip\n",
      " sab\n",
      " par\n",
      "\n",
      "\n",
      "Número de clusters: 13\n",
      "0     11141\n",
      "11     5334\n",
      "5      2143\n",
      "1      2059\n",
      "3      1260\n",
      "10     1191\n",
      "7       710\n",
      "2       599\n",
      "6       591\n",
      "9       545\n",
      "12      419\n",
      "4       266\n",
      "8       105\n",
      "Name: Cluster_unig2_13, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " mas\n",
      " si\n",
      " graci\n",
      " via\n",
      " per\n",
      " quier\n",
      " dia\n",
      " hac\n",
      " hoy\n",
      " mejor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " mas\n",
      " dia\n",
      " hoy\n",
      " mejor\n",
      " nuestr\n",
      " buen\n",
      " hac\n",
      " madr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " graci\n",
      " dias\n",
      " buen\n",
      " madr\n",
      " si\n",
      " per\n",
      " mas\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " vip\n",
      " whatsapp\n",
      " sab\n",
      " tartuf\n",
      " mejor\n",
      " show\n",
      " sal\n",
      " info\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " palaci\n",
      " buen\n",
      " sol\n",
      " plaz\n",
      " parqu\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " stigmab\n",
      " años\n",
      " …\n",
      " pobrez\n",
      " mas\n",
      " vih\n",
      " person\n",
      " drogadiccion\n",
      " homosexual\n",
      " gay\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " si\n",
      " par\n",
      " hac\n",
      " per\n",
      " hoy\n",
      " dia\n",
      " quier\n",
      " madr\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " mas\n",
      " hoy\n",
      " dia\n",
      " si\n",
      " madr\n",
      " mejor\n",
      " hac\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " gran\n",
      " via\n",
      " madr\n",
      " españ\n",
      " …\n",
      " hoy\n",
      " graci\n",
      " dia\n",
      " par\n",
      " mas\n",
      "\n",
      "\n",
      "Número de clusters: 14\n",
      "0     10023\n",
      "13     5662\n",
      "8      2349\n",
      "11     1369\n",
      "9      1126\n",
      "1      1109\n",
      "2      1067\n",
      "4       710\n",
      "10      627\n",
      "7       599\n",
      "3       593\n",
      "12      542\n",
      "5       482\n",
      "6       105\n",
      "Name: Cluster_unig2_14, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " via\n",
      " graci\n",
      " per\n",
      " dia\n",
      " nuev\n",
      " hac\n",
      " mejor\n",
      " hoy\n",
      " quier\n",
      " ser\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " par\n",
      " hac\n",
      " hoy\n",
      " dia\n",
      " per\n",
      " quier\n",
      " madr\n",
      " tod\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " per\n",
      " com\n",
      " ver\n",
      " pued\n",
      " par\n",
      " hac\n",
      " ser\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " siempr\n",
      " estadi\n",
      " vicent\n",
      " calderon\n",
      " …\n",
      " santiag\n",
      " bernabeu\n",
      " madr\n",
      " atleti\n",
      " par\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " retir\n",
      " españ\n",
      " …\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " sol\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " graci\n",
      " madr\n",
      " list\n",
      " mejor\n",
      " esper\n",
      " vid\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " viern\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      " felizviern\n",
      " madr\n",
      " fin\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " mas\n",
      " list\n",
      " …\n",
      " reserv\n",
      " whatsapp\n",
      " mejor\n",
      " madr\n",
      " vip\n",
      " par\n",
      " sab\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " much\n",
      " …\n",
      " graci\n",
      " mas\n",
      " felic\n",
      " tod\n",
      " par\n",
      " mejor\n",
      " gan\n",
      " per\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " hoy\n",
      " dia\n",
      " nuestr\n",
      " nuev\n",
      " hac\n",
      " mejor\n",
      " sol\n",
      " pas\n",
      "\n",
      "\n",
      "Número de clusters: 15\n",
      "3     10443\n",
      "10     4378\n",
      "4      2077\n",
      "7      2073\n",
      "0      1944\n",
      "9      1176\n",
      "12      764\n",
      "6       710\n",
      "8       707\n",
      "1       599\n",
      "5       591\n",
      "2       375\n",
      "13      266\n",
      "14      155\n",
      "11      105\n",
      "Name: Cluster_unig2_15, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " mas\n",
      " si\n",
      " …\n",
      " quier\n",
      " pued\n",
      " much\n",
      " com\n",
      " hac\n",
      " madr\n",
      " ser\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dgtmadr\n",
      " dgt\n",
      " nivel\n",
      " autov\n",
      " pk\n",
      " autop\n",
      " retencion\n",
      " amarill\n",
      " decrecient\n",
      " crecient\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " siempr\n",
      " …\n",
      " com\n",
      " par\n",
      " mejor\n",
      " mas\n",
      " graci\n",
      " madr\n",
      " per\n",
      " nunc\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " com\n",
      " via\n",
      " graci\n",
      " dia\n",
      " mejor\n",
      " hoy\n",
      " hac\n",
      " nuev\n",
      " quier\n",
      " pas\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " in\n",
      " españ\n",
      " via\n",
      " gran\n",
      " palaci\n",
      " plaz\n",
      " sol\n",
      " adecc\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " at\n",
      " im\n",
      " in\n",
      " madr\n",
      " comun\n",
      " madridbaraj\n",
      " mad\n",
      " adolf\n",
      " aenaaeropuert\n",
      " suarez\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " madr\n",
      " tiempodejueg\n",
      " visavis\n",
      " nadal\n",
      " sv\n",
      " manoslimpi\n",
      " fatim\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " dia\n",
      " hac\n",
      " mas\n",
      " com\n",
      " mejor\n",
      " nuestr\n",
      " hoy\n",
      " madr\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " buen\n",
      " dias\n",
      " retir\n",
      " viern\n",
      " madr\n",
      " jardin\n",
      " feliz\n",
      " …\n",
      " hoy\n",
      " dia\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " graci\n",
      " par\n",
      " madr\n",
      " sab\n",
      " com\n",
      " mas\n",
      " mund\n",
      " hac\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " …\n",
      " hoy\n",
      " com\n",
      " madr\n",
      " dia\n",
      " nuev\n",
      " nuestr\n",
      " hac\n",
      " mejor\n",
      " graci\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " esadelawschool\n",
      " esbes\n",
      " esasti\n",
      " esanchezmolin\n",
      " esalud\n",
      " 󾭚\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " per\n",
      " …\n",
      " si\n",
      " com\n",
      " tod\n",
      " par\n",
      " ser\n",
      " quier\n",
      " hac\n",
      " pued\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " list\n",
      " reserv\n",
      " vip\n",
      " whatsapp\n",
      " sab\n",
      " tartuf\n",
      " mejor\n",
      " show\n",
      " sal\n",
      " info\n",
      "\n",
      "Cluster 14 : Palabras más frecuentes :\n",
      " h\n",
      " c\n",
      " temperatur\n",
      " kmh\n",
      " °\n",
      " vient\n",
      " hpa\n",
      " mm\n",
      " hr\n",
      " precipit\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Nueva prueba de clustering con k-means, al detectarse otros grupos con otra semilla\n",
    "kmeans_function(16, mx_tfidf_unig, tweets, vectorizer_unig, \"unig2\", 4444)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "      <th>Cluster_unig_5</th>\n",
       "      <th>Cluster_unig_6</th>\n",
       "      <th>Cluster_unig_7</th>\n",
       "      <th>Cluster_unig_8</th>\n",
       "      <th>Cluster_unig_9</th>\n",
       "      <th>Cluster_unig_10</th>\n",
       "      <th>Cluster_unig_11</th>\n",
       "      <th>...</th>\n",
       "      <th>Cluster_unig2_6</th>\n",
       "      <th>Cluster_unig2_7</th>\n",
       "      <th>Cluster_unig2_8</th>\n",
       "      <th>Cluster_unig2_9</th>\n",
       "      <th>Cluster_unig2_10</th>\n",
       "      <th>Cluster_unig2_11</th>\n",
       "      <th>Cluster_unig2_12</th>\n",
       "      <th>Cluster_unig2_13</th>\n",
       "      <th>Cluster_unig2_14</th>\n",
       "      <th>Cluster_unig2_15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Regalo para socios: 'Talento a la fuga', el li...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>La recuperación del ‘ladrillo’ dispara el 144,...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TELEVISIÓN SERIES - Cuenta atrás para conocer ...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Nuestra taza de hoy es un homenaje a uno de lo...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>@MCadepe nos vemos el 19 de mayooo!!!</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             content  Latitude  Longitude  \\\n",
       "0  Regalo para socios: 'Talento a la fuga', el li...    40.416     -3.703   \n",
       "1  La recuperación del ‘ladrillo’ dispara el 144,...    40.416     -3.703   \n",
       "2  TELEVISIÓN SERIES - Cuenta atrás para conocer ...    40.416     -3.703   \n",
       "3  Nuestra taza de hoy es un homenaje a uno de lo...    40.416     -3.703   \n",
       "4              @MCadepe nos vemos el 19 de mayooo!!!    40.416     -3.703   \n",
       "\n",
       "   Cluster_unig_5  Cluster_unig_6  Cluster_unig_7  Cluster_unig_8  \\\n",
       "0               0               0               0               0   \n",
       "1               0               1               0               6   \n",
       "2               0               0               0               0   \n",
       "3               0               1               0               6   \n",
       "4               0               1               0               6   \n",
       "\n",
       "   Cluster_unig_9  Cluster_unig_10  Cluster_unig_11        ...         \\\n",
       "0               1                5                3        ...          \n",
       "1               1                5                9        ...          \n",
       "2               1                5                3        ...          \n",
       "3               2                4                9        ...          \n",
       "4               1                5                9        ...          \n",
       "\n",
       "   Cluster_unig2_6  Cluster_unig2_7  Cluster_unig2_8  Cluster_unig2_9  \\\n",
       "0                4                0                3                8   \n",
       "1                4                0                3                8   \n",
       "2                4                0                3                8   \n",
       "3                4                1                3                4   \n",
       "4                4                0                3                8   \n",
       "\n",
       "   Cluster_unig2_10  Cluster_unig2_11  Cluster_unig2_12  Cluster_unig2_13  \\\n",
       "0                 0                10                10                 1   \n",
       "1                 0                 0                 0                 0   \n",
       "2                 0                10                10                 1   \n",
       "3                 9                 0                11                 0   \n",
       "4                 0                 0                 0                 0   \n",
       "\n",
       "   Cluster_unig2_14  Cluster_unig2_15  \n",
       "0                13                 7  \n",
       "1                 0                 3  \n",
       "2                13                 7  \n",
       "3                11                 0  \n",
       "4                 0                 3  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Se comprueba que las columnas se han añadido correctamente\n",
    "tweets.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### K-means con unigramas y bigramas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Convertimos la lista en la matriz Tf-idf\n",
    "vectorizer_unibi = TfidfVectorizer(\n",
    "                max_df = 0.95,\n",
    "                min_df = 0.00005,\n",
    "                max_features = 30000,\n",
    "                analyzer = 'word',\n",
    "                tokenizer = tokenize,\n",
    "                lowercase = True,\n",
    "                use_idf = True,\n",
    "                stop_words = spanish_stopwords,\n",
    "                ngram_range=(1,2))\n",
    "mx_tfidf_unibi = vectorizer_unibi.fit_transform(tweets_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26363, 30000)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mx_tfidf_unibi.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de clusters: 5\n",
      "1    18115\n",
      "0     6850\n",
      "4      703\n",
      "3      590\n",
      "2      105\n",
      "Name: Cluster_unibi_5, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " tod\n",
      " madr\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " si\n",
      " buen\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " buen\n",
      " in\n",
      " mas\n",
      " com\n",
      " spain\n",
      " tod\n",
      " si\n",
      " in madr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "\n",
      "Número de clusters: 6\n",
      "1    19776\n",
      "4     3563\n",
      "2      984\n",
      "5      747\n",
      "0      703\n",
      "3      590\n",
      "Name: Cluster_unibi_6, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " buen\n",
      " com\n",
      " in\n",
      " si\n",
      " tod\n",
      " hoy\n",
      " dias\n",
      " spain\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " mejor\n",
      " list\n",
      " reserv\n",
      " …\n",
      " vip\n",
      " sab\n",
      " whatsapp\n",
      " list vip\n",
      " madr\n",
      " sal\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " …\n",
      " madr\n",
      " tod\n",
      " dia\n",
      " hac\n",
      " com\n",
      " si\n",
      " hoy\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much graci\n",
      " much\n",
      " graci segu\n",
      " segu\n",
      " tod\n",
      " graci tod\n",
      " follow tuitutil\n",
      " devuelv follow\n",
      "\n",
      "\n",
      "Número de clusters: 7\n",
      "1    15942\n",
      "6     5981\n",
      "5     2458\n",
      "0      703\n",
      "4      590\n",
      "2      577\n",
      "3      112\n",
      "Name: Cluster_unibi_7, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " spain\n",
      " mas\n",
      " com\n",
      " si\n",
      " tod\n",
      " madr spain\n",
      " via\n",
      " in madr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " retir\n",
      " viern\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " jardin\n",
      " retir madr\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " aeropuert adolf\n",
      " suarez madridbaraj\n",
      " madridbaraj\n",
      " madridbaraj mad\n",
      " adolf suarez\n",
      " adolf\n",
      " suarez\n",
      " mad\n",
      " mad aenaaeropuert\n",
      " aenaaeropuert\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " madr\n",
      " mejor\n",
      " mas\n",
      " com\n",
      " hac\n",
      " dia\n",
      " hoy\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " mas\n",
      " madr\n",
      " tod\n",
      " hoy\n",
      " com\n",
      " si\n",
      " dia\n",
      " graci\n",
      " mejor\n",
      "\n",
      "\n",
      "Número de clusters: 8\n",
      "1    14339\n",
      "7     5360\n",
      "0     2403\n",
      "6     1475\n",
      "2      764\n",
      "4      729\n",
      "5      703\n",
      "3      590\n",
      "Name: Cluster_unibi_8, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " nuev\n",
      " seguidor\n",
      " seguidor unfollowers\n",
      " nuev seguidor\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " tuitutil\n",
      " tod\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " buen\n",
      " com\n",
      " spain\n",
      " si\n",
      " dias\n",
      " via\n",
      " tod\n",
      " madr spain\n",
      " buen dias\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " im at\n",
      " im\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madridbaraj mad\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much graci\n",
      " much\n",
      " graci segu\n",
      " segu\n",
      " tod\n",
      " graci tod\n",
      " follow tuitutil\n",
      " brev devuelv\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much mas\n",
      " much\n",
      " par\n",
      " cad\n",
      " si\n",
      " mas …\n",
      " mas bonit\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " hoy\n",
      " tod\n",
      " com\n",
      " si\n",
      " dia\n",
      " buen\n",
      " nuev\n",
      " hac\n",
      "\n",
      "\n",
      "Número de clusters: 9\n",
      "1    15688\n",
      "3     6141\n",
      "6     1142\n",
      "0      768\n",
      "4      717\n",
      "8      703\n",
      "2      590\n",
      "7      509\n",
      "5      105\n",
      "Name: Cluster_unibi_9, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " im at\n",
      " madr\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " aeropuert adolf\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " mas\n",
      " com\n",
      " spain\n",
      " via\n",
      " tod\n",
      " madr spain\n",
      " mejor\n",
      " dia\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " madr\n",
      " mas\n",
      " hoy\n",
      " tod\n",
      " com\n",
      " dia\n",
      " nuestr\n",
      " hac\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much graci\n",
      " much\n",
      " graci segu\n",
      " segu\n",
      " tod\n",
      " graci tod\n",
      " brev devuelv\n",
      " devuelv follow\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " ver si\n",
      " ver\n",
      " pued\n",
      " si quier\n",
      " par\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " dias amig\n",
      " dias feliz\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "\n",
      "Número de clusters: 10\n",
      "1    13880\n",
      "7     5347\n",
      "8     2300\n",
      "6     1451\n",
      "9      741\n",
      "4      718\n",
      "5      703\n",
      "3      590\n",
      "2      528\n",
      "0      105\n",
      "Name: Cluster_unibi_10, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " buen\n",
      " com\n",
      " si\n",
      " dias\n",
      " via\n",
      " tod\n",
      " buen dias\n",
      " mejor\n",
      " dia\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " madr spain\n",
      " spain\n",
      " madr\n",
      " madr madr\n",
      " buen\n",
      " fuenlabr madr\n",
      " mostol madr\n",
      " fuenlabr\n",
      " dias\n",
      " tod\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much graci\n",
      " much\n",
      " graci segu\n",
      " segu\n",
      " tod\n",
      " graci tod\n",
      " brev devuelv\n",
      " devuelv follow\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " much mas\n",
      " par\n",
      " much\n",
      " madr\n",
      " cad\n",
      " si\n",
      " mas …\n",
      " mas bonit\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " hoy\n",
      " tod\n",
      " com\n",
      " si\n",
      " dia\n",
      " buen\n",
      " nuev\n",
      " hac\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " madr\n",
      " tod\n",
      " mejor\n",
      " com\n",
      " dia\n",
      " hac\n",
      " buen\n",
      " hoy\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " in\n",
      " at\n",
      " in madr\n",
      " im at\n",
      " im\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madridbaraj mad\n",
      "\n",
      "\n",
      "Número de clusters: 11\n",
      "0     12919\n",
      "1      5761\n",
      "8      1548\n",
      "9      1231\n",
      "3      1148\n",
      "7      1042\n",
      "4       703\n",
      "6       668\n",
      "5       648\n",
      "2       590\n",
      "10      105\n",
      "Name: Cluster_unibi_11, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " via\n",
      " madr\n",
      " mejor\n",
      " dia\n",
      " per\n",
      " hoy\n",
      " nuev\n",
      " hac\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " buen dias\n",
      " par\n",
      " mas\n",
      " hoy\n",
      " madr\n",
      " dia\n",
      " nuestr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " par\n",
      " hac\n",
      " com …\n",
      " si\n",
      " hoy\n",
      " dia\n",
      " per\n",
      " buen\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " in\n",
      " at\n",
      " im at\n",
      " im\n",
      " in madr\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madridbaraj mad\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " graci\n",
      " much graci\n",
      " …\n",
      " much\n",
      " graci segu\n",
      " segu\n",
      " follow tuitutil\n",
      " devuelv follow\n",
      " brev devuelv\n",
      " devuelv\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " per\n",
      " ver si\n",
      " si quier\n",
      " ver\n",
      " pued\n",
      " par\n",
      " mas\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " madr spain\n",
      " spain\n",
      " retir\n",
      " in\n",
      " españ\n",
      " in madr\n",
      " madr españ\n",
      " retir madr\n",
      " gran via\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " tod …\n",
      " dias\n",
      " mas\n",
      " par tod\n",
      " per\n",
      " buen\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "\n",
      "Número de clusters: 12\n",
      "4     12229\n",
      "6      4365\n",
      "1      2177\n",
      "5      1691\n",
      "3      1372\n",
      "10     1111\n",
      "2       757\n",
      "9       703\n",
      "8       694\n",
      "7       590\n",
      "11      569\n",
      "0       105\n",
      "Name: Cluster_unibi_12, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " mejor\n",
      " madr\n",
      " dia\n",
      " hac\n",
      " com\n",
      " nuestr\n",
      " list\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " im at\n",
      " im\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madridbaraj mad\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much mas\n",
      " much\n",
      " par\n",
      " cad\n",
      " pued\n",
      " mas bonit\n",
      " men\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " spain\n",
      " via\n",
      " tod\n",
      " madr spain\n",
      " mejor\n",
      " españ\n",
      " per\n",
      " dia\n",
      " sol\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " com\n",
      " hoy\n",
      " …\n",
      " dia\n",
      " madr\n",
      " tod\n",
      " hac\n",
      " hoy toc\n",
      " toc\n",
      " per\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " tod\n",
      " dia\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      " hac\n",
      " sol\n",
      " vid\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " much graci\n",
      " graci segu\n",
      " much\n",
      " segu\n",
      " tod\n",
      " graci tod\n",
      " brev devuelv\n",
      " devuelv follow\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " per\n",
      " ver si\n",
      " ver\n",
      " pued\n",
      " si quier\n",
      " com\n",
      " par\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " retir\n",
      " viern\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " jardin\n",
      " retir madr\n",
      "\n",
      "\n",
      "Número de clusters: 13\n",
      "0     12206\n",
      "12     4527\n",
      "11     2403\n",
      "3      2168\n",
      "1       839\n",
      "8       795\n",
      "10      733\n",
      "9       703\n",
      "6       613\n",
      "5       590\n",
      "2       564\n",
      "4       117\n",
      "7       105\n",
      "Name: Cluster_unibi_13, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " mas\n",
      " spain\n",
      " tod\n",
      " madr spain\n",
      " via\n",
      " per\n",
      " sol\n",
      " hoy\n",
      " hac\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " madr\n",
      " tierr\n",
      " feliz dia\n",
      " feliz\n",
      " buen dia\n",
      " dia tierr\n",
      " par\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " retir\n",
      " viern\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " jardin\n",
      " retir madr\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " mas\n",
      " madr\n",
      " hac\n",
      " com\n",
      " nuestr\n",
      " par …\n",
      " hoy\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " gran via\n",
      " via\n",
      " gran\n",
      " via madr\n",
      " madr españ\n",
      " españ\n",
      " madr\n",
      " madr gran\n",
      " primark\n",
      " phot\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " quier\n",
      " …\n",
      " si\n",
      " si quier\n",
      " quier …\n",
      " mas\n",
      " com\n",
      " per\n",
      " ser\n",
      " tod\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " mejor\n",
      " …\n",
      " list\n",
      " madr\n",
      " par\n",
      " sal\n",
      " tod\n",
      " mas\n",
      " sab\n",
      " reserv\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " in\n",
      " at\n",
      " in madr\n",
      " im at\n",
      " im\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madridbaraj mad\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " com\n",
      " si\n",
      " graci\n",
      " …\n",
      " tod\n",
      " much\n",
      " per\n",
      " madr\n",
      " hac\n",
      " much graci\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " tod\n",
      " hoy\n",
      " nuev\n",
      " nuestr\n",
      " hac\n",
      " vid\n",
      " sol\n",
      "\n",
      "\n",
      "Número de clusters: 14\n",
      "0     12352\n",
      "8      4972\n",
      "1      2242\n",
      "9      1218\n",
      "2      1128\n",
      "7       773\n",
      "10      703\n",
      "6       638\n",
      "4       590\n",
      "3       506\n",
      "11      504\n",
      "5       500\n",
      "12      125\n",
      "13      112\n",
      "Name: Cluster_unibi_14, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " si\n",
      " graci\n",
      " nuev\n",
      " mejor\n",
      " per\n",
      " dia\n",
      " hoy\n",
      " quier\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " …\n",
      " in\n",
      " in madr\n",
      " españ\n",
      " madr …\n",
      " palaci\n",
      " real\n",
      " sol\n",
      " real madr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " si\n",
      " par\n",
      " com …\n",
      " per\n",
      " hoy\n",
      " dia\n",
      " com si\n",
      " quier\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " at\n",
      " im at\n",
      " im\n",
      " in\n",
      " madr\n",
      " in madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " cosl\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " dias amig\n",
      " dias feliz\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " via\n",
      " gran via\n",
      " gran\n",
      " via madr\n",
      " madr\n",
      " madr españ\n",
      " españ\n",
      " via youtub\n",
      " via elmund\n",
      " par\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " mas\n",
      " si\n",
      " com hac\n",
      " par hac\n",
      " tod\n",
      " hac …\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " hoy\n",
      " dia\n",
      " si\n",
      " nuestr\n",
      " nuev\n",
      " mejor\n",
      " graci\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " graci\n",
      " tod …\n",
      " sab\n",
      " graci tod\n",
      " per\n",
      " tod mund\n",
      " par tod\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " madr spain\n",
      " spain\n",
      " madr\n",
      " fuenlabr madr\n",
      " madr madr\n",
      " mostol madr\n",
      " fuenlabr\n",
      " mostol\n",
      " dia\n",
      " pabloalbor\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " retir\n",
      " retir madr\n",
      " parqu retir\n",
      " parqu\n",
      " buen retir\n",
      " jardin buen\n",
      " retir jardin\n",
      " jardin\n",
      " madr\n",
      " buen\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " aeropuert adolf\n",
      " suarez madridbaraj\n",
      " madridbaraj\n",
      " madridbaraj mad\n",
      " adolf suarez\n",
      " adolf\n",
      " suarez\n",
      " mad\n",
      " mad aenaaeropuert\n",
      " aenaaeropuert\n",
      "\n",
      "\n",
      "Número de clusters: 15\n",
      "7     12494\n",
      "10     4762\n",
      "3      1297\n",
      "11     1273\n",
      "4      1114\n",
      "13      967\n",
      "9       755\n",
      "6       755\n",
      "0       628\n",
      "2       590\n",
      "14      561\n",
      "12      367\n",
      "8       360\n",
      "5       335\n",
      "1       105\n",
      "Name: Cluster_unibi_15, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " via\n",
      " gran via\n",
      " gran\n",
      " via madr\n",
      " madr\n",
      " madr españ\n",
      " españ\n",
      " via youtub\n",
      " via elmund\n",
      " par\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " unfollowers tuitutil\n",
      " nuev seguidor\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " ey\n",
      " facult cc\n",
      " factor\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " much mas\n",
      " much\n",
      " par\n",
      " pued\n",
      " cad\n",
      " mas bonit\n",
      " mejor\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " par\n",
      " si\n",
      " madr\n",
      " com …\n",
      " tod\n",
      " hoy\n",
      " per\n",
      " gust\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " tendenci madr\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " madr\n",
      " ana\n",
      " dasoul\n",
      " visavis ahor\n",
      " tiempodejueg\n",
      " tiempodejueg ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " im at\n",
      " im\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " aeropuert adolf\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " spain\n",
      " tod\n",
      " madr spain\n",
      " mejor\n",
      " sol\n",
      " per\n",
      " hoy\n",
      " nuev\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " sobr\n",
      " …\n",
      " rey juan\n",
      " univers rey\n",
      " rey\n",
      " juan carl\n",
      " juan\n",
      " univers\n",
      " carl\n",
      " par\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " com hac\n",
      " mas\n",
      " tod\n",
      " par hac\n",
      " falt\n",
      " si\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " madr\n",
      " tod\n",
      " hoy\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      " noch\n",
      " sab\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " graci\n",
      " dia\n",
      " …\n",
      " much graci\n",
      " tod\n",
      " much\n",
      " hoy\n",
      " graci segu\n",
      " madr\n",
      " segu\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " tendenci spain\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " sv ahor\n",
      " visavis ahor\n",
      " tiempodejueg ahor\n",
      " tiempodejueg\n",
      " axe\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " per\n",
      " ver si\n",
      " si quier\n",
      " pued\n",
      " ver\n",
      " par\n",
      " ser\n",
      "\n",
      "Cluster 14 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " retir\n",
      " viern\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " jardin\n",
      " retir madr\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Realización del clustering con k-means\n",
    "kmeans_function(16, mx_tfidf_unibi, tweets, vectorizer_unibi, \"unibi\", 3333)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### K-means con unigramas, bigramas y trigramas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Convertimos la lista en la matriz Tf-idf\n",
    "vectorizer_unibitri = TfidfVectorizer(\n",
    "                max_df = 0.95,\n",
    "                min_df = 0.00005,\n",
    "                max_features = 30000,\n",
    "                analyzer = 'word',\n",
    "                tokenizer = tokenize,\n",
    "                lowercase = True,\n",
    "                use_idf = True,\n",
    "                stop_words = spanish_stopwords,\n",
    "                ngram_range=(1,3))\n",
    "mx_tfidf_unibitri = vectorizer_unibitri.fit_transform(tweets_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26363, 30000)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mx_tfidf_unibitri.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de clusters: 5\n",
      "1    17647\n",
      "0     6642\n",
      "3      791\n",
      "2      696\n",
      "4      587\n",
      "Name: Cluster_unibitri_5, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " madr\n",
      " mas\n",
      " tod\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " si\n",
      " nuestr\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " par\n",
      " buen\n",
      " mas\n",
      " com\n",
      " si\n",
      " tod\n",
      " spain\n",
      " dias\n",
      " via\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " madr\n",
      " at\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " madr comun\n",
      " madr comun madr\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "\n",
      "Número de clusters: 6\n",
      "0    20706\n",
      "1     2400\n",
      "4     1470\n",
      "5      696\n",
      "3      587\n",
      "2      504\n",
      "Name: Cluster_unibitri_6, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " com\n",
      " si\n",
      " in\n",
      " hoy\n",
      " graci\n",
      " spain\n",
      " dia\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tuitutil\n",
      " nuev\n",
      " seguidor\n",
      " nuev seguidor\n",
      " nuev seguidor unfollowers\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " dias amig\n",
      " buen dias amig\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " retir\n",
      " madr\n",
      " retir madr\n",
      " par\n",
      " graci\n",
      " parqu retir\n",
      " buen\n",
      " tod …\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "\n",
      "Número de clusters: 7\n",
      "0    15068\n",
      "6     5635\n",
      "2     2415\n",
      "4     1459\n",
      "5      696\n",
      "1      587\n",
      "3      503\n",
      "Name: Cluster_unibitri_7, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " mas\n",
      " com\n",
      " si\n",
      " tod\n",
      " graci\n",
      " via\n",
      " per\n",
      " nuev\n",
      " mejor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " mas\n",
      " mejor\n",
      " com\n",
      " hac\n",
      " madr\n",
      " dia\n",
      " hoy\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " buen dias amig\n",
      " dias amig\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " madr spain\n",
      " in madr\n",
      " spain\n",
      " at\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " tod\n",
      " hoy\n",
      " com\n",
      " si\n",
      " dia\n",
      " graci\n",
      " nuestr\n",
      "\n",
      "\n",
      "Número de clusters: 8\n",
      "6    15042\n",
      "7     5519\n",
      "1     2353\n",
      "0     1263\n",
      "4      800\n",
      "2      694\n",
      "5      587\n",
      "3      105\n",
      "Name: Cluster_unibitri_8, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " par\n",
      " si\n",
      " per\n",
      " hac\n",
      " tod\n",
      " com …\n",
      " madr\n",
      " hoy\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " madr\n",
      " mas\n",
      " mejor\n",
      " dia\n",
      " hoy\n",
      " buen\n",
      " hac\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " madr\n",
      " at\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " in madr comun\n",
      " madr comun madr\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " madr\n",
      " buen\n",
      " mas\n",
      " spain\n",
      " si\n",
      " dias\n",
      " tod\n",
      " graci\n",
      " via\n",
      " madr spain\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " tod\n",
      " hoy\n",
      " dia\n",
      " si\n",
      " graci\n",
      " buen\n",
      " nuestr\n",
      "\n",
      "\n",
      "Número de clusters: 9\n",
      "2    15903\n",
      "7     6093\n",
      "6      884\n",
      "4      816\n",
      "1      778\n",
      "5      696\n",
      "3      587\n",
      "8      501\n",
      "0      105\n",
      "Name: Cluster_unibitri_9, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " madr\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " in madr comun\n",
      " madr comun\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " par\n",
      " madr\n",
      " mas\n",
      " com\n",
      " si\n",
      " tod\n",
      " graci\n",
      " via\n",
      " mejor\n",
      " per\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " madr\n",
      " madr spain\n",
      " spain\n",
      " retir\n",
      " retir madr\n",
      " gran via\n",
      " españ\n",
      " madr españ\n",
      " via madr\n",
      " gran via madr\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " ser\n",
      " gan\n",
      " …\n",
      " par\n",
      " pued\n",
      " pued ser\n",
      " mas\n",
      " si\n",
      " per\n",
      " quier\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " tod\n",
      " madr\n",
      " mas\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " si\n",
      " nuestr\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " dias amig\n",
      " buen dias amig\n",
      "\n",
      "\n",
      "Número de clusters: 10\n",
      "1    13694\n",
      "2     5016\n",
      "8     2181\n",
      "6     1249\n",
      "7      807\n",
      "3      777\n",
      "4      772\n",
      "5      694\n",
      "9      587\n",
      "0      586\n",
      "Name: Cluster_unibitri_10, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " buen dias\n",
      " dias\n",
      " buen\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " seguidor unfollowers\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " madr\n",
      " mas\n",
      " com\n",
      " si\n",
      " graci\n",
      " via\n",
      " dia\n",
      " mejor\n",
      " quier\n",
      " hoy\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " mas\n",
      " hoy\n",
      " com\n",
      " dia\n",
      " si\n",
      " nuestr\n",
      " mejor\n",
      " hac\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " per\n",
      " …\n",
      " si\n",
      " com\n",
      " tod\n",
      " per si\n",
      " per …\n",
      " mas\n",
      " par\n",
      " ser\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " madr\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " madr comun madr\n",
      " madr comun\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " graci\n",
      " tod …\n",
      " par\n",
      " sab\n",
      " graci tod\n",
      " mas\n",
      " com\n",
      " si\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " madr\n",
      " madr spain\n",
      " spain\n",
      " retir\n",
      " retir madr\n",
      " gran via\n",
      " via madr\n",
      " gran via madr\n",
      " españ\n",
      " madr españ\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " mejor\n",
      " mas\n",
      " com\n",
      " hac\n",
      " hoy\n",
      " dia\n",
      " madr\n",
      " nuestr\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "\n",
      "Número de clusters: 11\n",
      "1     16462\n",
      "5      3174\n",
      "2      1381\n",
      "4      1206\n",
      "7      1150\n",
      "6       877\n",
      "8       724\n",
      "3       587\n",
      "10      363\n",
      "9       334\n",
      "0       105\n",
      "Name: Cluster_unibitri_11, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " madr\n",
      " buen\n",
      " tod\n",
      " com\n",
      " graci\n",
      " dias\n",
      " mejor\n",
      " hac\n",
      " per\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " in madr\n",
      " madr spain\n",
      " at\n",
      " spain\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hoy\n",
      " pas\n",
      " …\n",
      " hoy toc\n",
      " toc\n",
      " buen\n",
      " com\n",
      " dias\n",
      " madr\n",
      " tod\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " …\n",
      " tod\n",
      " madr\n",
      " mejor\n",
      " hac\n",
      " much\n",
      " com\n",
      " nuestr\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " buen\n",
      " feliz dia\n",
      " feliz\n",
      " tierr\n",
      " buen dia\n",
      " par\n",
      " madr\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " retir\n",
      " quier\n",
      " buen\n",
      " per\n",
      " com\n",
      " jardin buen retir\n",
      " jardin buen\n",
      " buen retir\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " sol\n",
      " madr\n",
      " gran via\n",
      " via\n",
      " gran\n",
      " gran via madr\n",
      " via madr\n",
      " españ\n",
      " via madr españ\n",
      " sol madr\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " madr\n",
      " ana\n",
      " dasoul\n",
      " coron\n",
      " valverd\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " ahor tendenci spain\n",
      " tendenci spain\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " axe\n",
      " jorg\n",
      " txelui\n",
      " sv ahor\n",
      "\n",
      "\n",
      "Número de clusters: 12\n",
      "0     11648\n",
      "5      4359\n",
      "1      2298\n",
      "6      2235\n",
      "9      2130\n",
      "4       695\n",
      "3       653\n",
      "2       653\n",
      "7       587\n",
      "10      509\n",
      "8       491\n",
      "11      105\n",
      "Name: Cluster_unibitri_12, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " si\n",
      " graci\n",
      " tod\n",
      " mejor\n",
      " per\n",
      " quier\n",
      " hoy\n",
      " dia\n",
      " hac\n",
      " buen\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " hac\n",
      " mas\n",
      " com\n",
      " hoy\n",
      " dia\n",
      " mejor\n",
      " nuestr\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " in\n",
      " at\n",
      " im at\n",
      " im\n",
      " in madr\n",
      " madr\n",
      " comun madr\n",
      " comun\n",
      " in madr comun\n",
      " madr comun madr\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " via\n",
      " gran via\n",
      " gran\n",
      " via madr\n",
      " gran via madr\n",
      " via madr españ\n",
      " madr\n",
      " madr españ\n",
      " españ\n",
      " via youtub\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " tendenci spain\n",
      " ahor tendenci spain\n",
      " tendenci madr\n",
      " ahor tendenci madr\n",
      " spain\n",
      " madr\n",
      " visavis ahor\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " …\n",
      " tod\n",
      " hoy\n",
      " si\n",
      " dia\n",
      " graci\n",
      " nuestr\n",
      " hac\n",
      " nuev\n",
      " mejor\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " …\n",
      " si\n",
      " tod\n",
      " much\n",
      " hac\n",
      " per\n",
      " dia\n",
      " quier\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " buen dias amig\n",
      " dias amig\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " madr\n",
      " …\n",
      " retir\n",
      " in\n",
      " in madr\n",
      " retir madr\n",
      " madr …\n",
      " españ\n",
      " sol\n",
      " palaci\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " madr spain\n",
      " spain\n",
      " madr\n",
      " madr madr spain\n",
      " fuenlabr madr spain\n",
      " fuenlabr madr\n",
      " mostol madr spain\n",
      " in madr spain\n",
      " mostol madr\n",
      " fuenlabr\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "\n",
      "Número de clusters: 13\n",
      "0     13715\n",
      "9      4818\n",
      "11     1782\n",
      "4      1401\n",
      "8       800\n",
      "12      784\n",
      "7       767\n",
      "2       587\n",
      "1       496\n",
      "6       412\n",
      "10      363\n",
      "5       333\n",
      "3       105\n",
      "Name: Cluster_unibitri_13, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " par\n",
      " madr\n",
      " com\n",
      " graci\n",
      " tod\n",
      " via\n",
      " mejor\n",
      " per\n",
      " hac\n",
      " nuev\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dias\n",
      " buen dias\n",
      " buen\n",
      " viern\n",
      " feliz\n",
      " hoy\n",
      " …\n",
      " felizviern\n",
      " dias amig\n",
      " buen dias amig\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " par\n",
      " much\n",
      " much mas\n",
      " madr\n",
      " cad\n",
      " pued\n",
      " mas …\n",
      " mas bonit\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " ahor tendenci madr\n",
      " tendenci madr\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " madr\n",
      " dasoul\n",
      " coron\n",
      " ana\n",
      " valverd\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " cas\n",
      " puert\n",
      " sol madr\n",
      " puert sol\n",
      " sol\n",
      " puert sol madr\n",
      " madr\n",
      " …\n",
      " vuelv\n",
      " cas camp\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " in\n",
      " in madr\n",
      " at\n",
      " madr\n",
      " im at\n",
      " im\n",
      " comun madr\n",
      " comun\n",
      " in madr comun\n",
      " madr comun\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " madr\n",
      " madr spain\n",
      " spain\n",
      " retir\n",
      " retir madr\n",
      " gran via\n",
      " via madr\n",
      " gran via madr\n",
      " españ\n",
      " madr españ\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " tod\n",
      " madr\n",
      " com\n",
      " nuestr\n",
      " dia\n",
      " graci\n",
      " nuev\n",
      " hac\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " ahor tendenci spain\n",
      " tendenci spain\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " axe\n",
      " jorg\n",
      " txelui\n",
      " sv ahor\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " si\n",
      " hoy\n",
      " …\n",
      " dia\n",
      " com\n",
      " par\n",
      " quier\n",
      " per\n",
      " tod\n",
      " hac\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " esper\n",
      " felic\n",
      " disfrut\n",
      " …\n",
      " par\n",
      " tod\n",
      " much\n",
      " mejor\n",
      " list\n",
      " madr\n",
      "\n",
      "\n",
      "Número de clusters: 14\n",
      "11    11741\n",
      "1      4366\n",
      "5      2916\n",
      "2      1290\n",
      "4      1112\n",
      "7      1067\n",
      "6       843\n",
      "13      745\n",
      "8       628\n",
      "3       587\n",
      "10      363\n",
      "9       333\n",
      "0       267\n",
      "12      105\n",
      "Name: Cluster_unibitri_14, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " cre\n",
      " …\n",
      " mas\n",
      " pued\n",
      " par\n",
      " si\n",
      " per\n",
      " tod\n",
      " nunc\n",
      " cre si\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " buen dias\n",
      " tod\n",
      " madr\n",
      " graci\n",
      " com\n",
      " noch\n",
      " mejor\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " in madr\n",
      " at\n",
      " madr spain\n",
      " im at\n",
      " im\n",
      " spain\n",
      " comun madr\n",
      " comun\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hoy\n",
      " pas\n",
      " …\n",
      " hoy toc\n",
      " toc\n",
      " com\n",
      " madr\n",
      " par\n",
      " tod\n",
      " aqu\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " …\n",
      " madr\n",
      " tod\n",
      " mejor\n",
      " much\n",
      " com\n",
      " nuestr\n",
      " sab\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " feliz dia\n",
      " buen\n",
      " tierr\n",
      " feliz\n",
      " buen dia\n",
      " madr\n",
      " par\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " retir\n",
      " quier\n",
      " jardin buen retir\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " retir jardin buen\n",
      " jardin\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " sol\n",
      " gran via\n",
      " madr\n",
      " via\n",
      " gran\n",
      " gran via madr\n",
      " via madr\n",
      " via madr españ\n",
      " sol madr\n",
      " puert sol\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " ahor tendenci madr\n",
      " tendenci madr\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " madr\n",
      " dasoul\n",
      " coron\n",
      " ana\n",
      " valverd\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " ahor tendenci spain\n",
      " tendenci spain\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " axe\n",
      " jorg\n",
      " txelui\n",
      " sv ahor\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " madr\n",
      " com\n",
      " graci\n",
      " tod\n",
      " via\n",
      " mejor\n",
      " per\n",
      " nuev\n",
      " quier\n",
      " sab\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " mas\n",
      " si\n",
      " tod\n",
      " com hac\n",
      " par hac\n",
      " per\n",
      "\n",
      "\n",
      "Número de clusters: 15\n",
      "11    11268\n",
      "1      4300\n",
      "5      2812\n",
      "2      1283\n",
      "4      1103\n",
      "7      1050\n",
      "6       837\n",
      "13      737\n",
      "8       676\n",
      "14      642\n",
      "3       587\n",
      "10      363\n",
      "9       333\n",
      "0       267\n",
      "12      105\n",
      "Name: Cluster_unibitri_15, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " cre\n",
      " …\n",
      " mas\n",
      " pued\n",
      " par\n",
      " si\n",
      " per\n",
      " tod\n",
      " nunc\n",
      " cre si\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " buen\n",
      " dias\n",
      " buen dias\n",
      " tod\n",
      " madr\n",
      " graci\n",
      " com\n",
      " noch\n",
      " mejor\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " in madr\n",
      " at\n",
      " madr spain\n",
      " im at\n",
      " im\n",
      " spain\n",
      " comun madr\n",
      " comun\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " autop autov\n",
      " autov\n",
      " autop\n",
      " dgtmadr dgt\n",
      " dgtmadr\n",
      " dgt\n",
      " retencion nivel\n",
      " retencion\n",
      " nivel\n",
      " pk\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hoy\n",
      " pas\n",
      " …\n",
      " hoy toc\n",
      " toc\n",
      " com\n",
      " madr\n",
      " par\n",
      " tod\n",
      " aqu\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " …\n",
      " tod\n",
      " madr\n",
      " mejor\n",
      " much\n",
      " com\n",
      " nuestr\n",
      " ser\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " feliz dia\n",
      " buen\n",
      " tierr\n",
      " feliz\n",
      " buen dia\n",
      " madr\n",
      " par\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " retir\n",
      " quier\n",
      " jardin buen retir\n",
      " jardin buen\n",
      " buen retir\n",
      " retir jardin\n",
      " retir jardin buen\n",
      " jardin\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " sol\n",
      " madr\n",
      " españ\n",
      " madr españ\n",
      " sol madr\n",
      " via madr españ\n",
      " gran via madr\n",
      " via madr\n",
      " puert sol\n",
      " gran via\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " ahor tendenci madr\n",
      " tendenci madr\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " madr\n",
      " dasoul\n",
      " coron\n",
      " ana\n",
      " valverd\n",
      "\n",
      "Cluster 10 : Palabras más frecuentes :\n",
      " ahor tendenci spain\n",
      " tendenci spain\n",
      " ahor tendenci\n",
      " tendenci\n",
      " ahor\n",
      " spain\n",
      " axe\n",
      " jorg\n",
      " txelui\n",
      " sv ahor\n",
      "\n",
      "Cluster 11 : Palabras más frecuentes :\n",
      " madr\n",
      " com\n",
      " graci\n",
      " tod\n",
      " mejor\n",
      " per\n",
      " quier\n",
      " nuev\n",
      " sab\n",
      " tien\n",
      "\n",
      "Cluster 12 : Palabras más frecuentes :\n",
      " seguidor unfollowers\n",
      " unfollowers\n",
      " unfollowers tuitutil\n",
      " nuev seguidor unfollowers\n",
      " nuev seguidor\n",
      " seguidor unfollowers tuitutil\n",
      " seguidor\n",
      " tuitutil\n",
      " nuev\n",
      " eugelopezmaur\n",
      "\n",
      "Cluster 13 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " mas\n",
      " si\n",
      " tod\n",
      " com hac\n",
      " par hac\n",
      " per\n",
      "\n",
      "Cluster 14 : Palabras más frecuentes :\n",
      " via\n",
      " gran via\n",
      " gran\n",
      " via youtub\n",
      " par\n",
      " via elmund\n",
      " elmund\n",
      " madr\n",
      " youtub\n",
      " elpais\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Realización del clustering con k-means\n",
    "kmeans_function(16, mx_tfidf_unibitri, tweets, vectorizer_unibitri, \"unibitri\", 3333)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* ### Asignación de temas/topics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Por defecto, asignamos el tema 'otros'\n",
    "tweets['Topic'] = 'otros'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'sitios-madrid'\n",
    "tweets.loc[(tweets[\"Cluster_unig_8\"] == 4) & (tweets[\"Cluster_unig_11\"] == 7) & \n",
    "           (tweets[\"Cluster_unig_12\"] == 7) & (tweets[\"Cluster_unig_13\"] == 1) & \n",
    "           (tweets[\"Cluster_unig_15\"] == 12) & (tweets[\"Cluster_unig2_10\"] == 5) &\n",
    "           (tweets[\"Cluster_unig2_11\"] == 8) & (tweets[\"Cluster_unig2_12\"] == 8) &\n",
    "           (tweets[\"Cluster_unig2_13\"] == 5) & (tweets[\"Cluster_unig2_14\"] == 8) &\n",
    "           (tweets[\"Cluster_unig2_15\"] == 4) & (tweets[\"Cluster_unibi_11\"] == 8) & \n",
    "           (tweets[\"Cluster_unibi_14\"] == 1) & (tweets[\"Cluster_unibitri_12\"] == 9) &\n",
    "           (tweets[\"Cluster_unibitri_13\"] == 6),'Topic'] = 'sitios-madrid'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            26307\n",
       "sitios-madrid       56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Se incluye un nuevo subconjunto a 'sitios-madrid', que teóricamente incluye pueblos de la Comunidad\n",
    "tweets.loc[(tweets[\"Cluster_unig_14\"] == 4) | (tweets[\"Cluster_unig_15\"] == 0) | \n",
    "           (tweets[\"Cluster_unibi_10\"] == 2) | (tweets[\"Cluster_unibitri_12\"] == 10),\n",
    "           'Topic'] = 'sitios-madrid'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            25269\n",
       "sitios-madrid     1094\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2781</th>\n",
       "      <td>I'm at @Facilisimo.com in Villaviciosa de Odón...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3828</th>\n",
       "      <td>I'm at AC Hotel La Finca in Pozuelo de Alarcón...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4824</th>\n",
       "      <td>¡La cocina es el corazón de la casa! Refórmala...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5845</th>\n",
       "      <td>I'm at Madrid in Madrid, España</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6208</th>\n",
       "      <td>I'm at Viva Gym embajadores</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "2781  I'm at @Facilisimo.com in Villaviciosa de Odón...\n",
       "3828  I'm at AC Hotel La Finca in Pozuelo de Alarcón...\n",
       "4824  ¡La cocina es el corazón de la casa! Refórmala...\n",
       "5845                   I'm at Madrid in Madrid, España \n",
       "6208                       I'm at Viva Gym embajadores "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'sitios-madrid',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'buenos-dias'\n",
    "tweets.loc[(tweets[\"Cluster_unig_7\"] == 1) & (tweets[\"Cluster_unig_11\"] == 5) & \n",
    "           (tweets[\"Cluster_unig_13\"] == 8) & (tweets[\"Cluster_unig_14\"] == 13) & \n",
    "           (tweets[\"Cluster_unig_15\"] == 10) & (tweets[\"Cluster_unig2_6\"] == 5) &\n",
    "           (tweets[\"Cluster_unig2_8\"] == 1) & (tweets[\"Cluster_unig2_9\"] == 5) &\n",
    "           (tweets[\"Cluster_unig2_10\"] == 8) & (tweets[\"Cluster_unig2_11\"] == 10) &\n",
    "           (tweets[\"Cluster_unig2_12\"] == 10) & (tweets[\"Cluster_unig2_14\"] == 10) &\n",
    "           (tweets[\"Cluster_unig2_15\"] == 8) & (tweets[\"Cluster_unibi_7\"] == 2) & \n",
    "           (tweets[\"Cluster_unibi_9\"] == 7) & (tweets[\"Cluster_unibi_12\"] == 11) & \n",
    "           (tweets[\"Cluster_unibi_13\"] == 2) & (tweets[\"Cluster_unibi_14\"] == 5) & \n",
    "           (tweets[\"Cluster_unibi_15\"] == 14) & (tweets[\"Cluster_unibitri_6\"] == 2) &\n",
    "           (tweets[\"Cluster_unibitri_7\"] == 3) & (tweets[\"Cluster_unibitri_9\"] == 8) &\n",
    "           (tweets[\"Cluster_unibitri_12\"] == 8) & (tweets[\"Cluster_unibitri_13\"] == 1),\n",
    "           'Topic'] = 'buenos-dias'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24865\n",
       "sitios-madrid     1091\n",
       "buenos-dias        407\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>413</th>\n",
       "      <td>@kardiez2000 Buenos días, no sé si conocías la...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>687</th>\n",
       "      <td>@margarita_5239 Buenos días cielete y feliz fi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>699</th>\n",
       "      <td>@maricarmengar15 @Pedramuz @Mariquinaperez @ma...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>845</th>\n",
       "      <td>@raulitopro11 buenos dias  Raul feliz Viernes ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1066</th>\n",
       "      <td>@Vermutinn jajajajjaja Se lo merecía,  es insu...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "413   @kardiez2000 Buenos días, no sé si conocías la...\n",
       "687   @margarita_5239 Buenos días cielete y feliz fi...\n",
       "699   @maricarmengar15 @Pedramuz @Mariquinaperez @ma...\n",
       "845   @raulitopro11 buenos dias  Raul feliz Viernes ...\n",
       "1066  @Vermutinn jajajajjaja Se lo merecía,  es insu..."
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'buenos-dias',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'retiro'\n",
    "tweets.loc[(tweets[\"Cluster_unibi_14\"] == 12) & (tweets[\"Cluster_unibitri_9\"] == 4) & \n",
    "           (tweets[\"Cluster_unibitri_10\"] == 7) & (tweets[\"Cluster_unibitri_11\"] == 7) & \n",
    "           (tweets[\"Cluster_unibitri_14\"] == 7) & (tweets[\"Cluster_unibitri_15\"] == 7), \n",
    "           'Topic'] = 'retiro'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24805\n",
       "sitios-madrid     1090\n",
       "buenos-dias        407\n",
       "retiro              61\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Es muy restrictivo, cuando había un grupo claro en Cluster_unibi_14. Usaremos solo ese\n",
    "tweets.loc[(tweets[\"Cluster_unibi_14\"] == 12), 'Topic'] = 'retiro'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24743\n",
       "sitios-madrid     1088\n",
       "buenos-dias        407\n",
       "retiro             125\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9832</th>\n",
       "      <td>\"Claro que lo bueno cuesta que no lo ganas si ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10039</th>\n",
       "      <td>ayer eramos rookies hoy somos all-stars @ Parq...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10330</th>\n",
       "      <td>De Jardines del Buen Retiro \\n#mrcade #reproch...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10438</th>\n",
       "      <td>#Beerlover #Instabeer - #Madrid, #Spain. 🇪🇸 @ ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10500</th>\n",
       "      <td>#TBT Mayo 15. Palacio de Cristal en el Parque ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 content\n",
       "9832   \"Claro que lo bueno cuesta que no lo ganas si ...\n",
       "10039  ayer eramos rookies hoy somos all-stars @ Parq...\n",
       "10330  De Jardines del Buen Retiro \\n#mrcade #reproch...\n",
       "10438  #Beerlover #Instabeer - #Madrid, #Spain. 🇪🇸 @ ...\n",
       "10500  #TBT Mayo 15. Palacio de Cristal en el Parque ..."
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'retiro',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'gran-via'\n",
    "tweets.loc[(tweets[\"Cluster_unibi_13\"] == 4) & (tweets[\"Cluster_unibi_14\"] == 6) &\n",
    "           (tweets[\"Cluster_unibi_15\"] == 0) & (tweets[\"Cluster_unibitri_12\"] == 3), 'Topic'] = 'gran-via'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24633\n",
       "sitios-madrid     1083\n",
       "buenos-dias        407\n",
       "retiro             125\n",
       "gran-via           115\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>Shredded Chicken Salpicón  via @Leslie_Limon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>728</th>\n",
       "      <td>vía @giphy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1719</th>\n",
       "      <td>Alimentaria celebra en el recinto Gran Vía de ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3717</th>\n",
       "      <td>Ayer me compré de oferta estos 3 libros en las...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5531</th>\n",
       "      <td>Vía @LaCastaSusana</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "72         Shredded Chicken Salpicón  via @Leslie_Limon\n",
       "728                                          vía @giphy\n",
       "1719  Alimentaria celebra en el recinto Gran Vía de ...\n",
       "3717  Ayer me compré de oferta estos 3 libros en las...\n",
       "5531                                Vía @LaCastaSusana "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'gran-via',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'gracias'\n",
    "tweets.loc[(tweets[\"Cluster_unig_14\"] == 11) & (tweets[\"Cluster_unig2_9\"] == 6) &\n",
    "           (tweets[\"Cluster_unig2_14\"] == 12) & (tweets[\"Cluster_unibi_13\"] == 11) & \n",
    "           (tweets[\"Cluster_unibi_15\"] == 11), 'Topic'] = 'gracias'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24577\n",
       "sitios-madrid     1083\n",
       "buenos-dias        407\n",
       "retiro             125\n",
       "gran-via           115\n",
       "gracias             56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>233</th>\n",
       "      <td>@GlamourSpain correspondiente porque no lo he ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>Y se cumple un año de #SeisHermanas muchas gra...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>449</th>\n",
       "      <td>Mi principe, muchas gracias, siempre apoyando,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>643</th>\n",
       "      <td>@Tri_finisher @totibest Muchas gracias por com...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1287</th>\n",
       "      <td>@stephipugliese @BrunoVD @FerRubioA @soyAlicia...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "233   @GlamourSpain correspondiente porque no lo he ...\n",
       "270   Y se cumple un año de #SeisHermanas muchas gra...\n",
       "449   Mi principe, muchas gracias, siempre apoyando,...\n",
       "643   @Tri_finisher @totibest Muchas gracias por com...\n",
       "1287  @stephipugliese @BrunoVD @FerRubioA @soyAlicia..."
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'gracias',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'followers'\n",
    "tweets.loc[(tweets[\"Cluster_unibi_5\"] == 2) & (tweets[\"Cluster_unibi_8\"] == 0) & \n",
    "           (tweets[\"Cluster_unibi_10\"] == 0) & (tweets[\"Cluster_unibi_11\"] == 10) & \n",
    "           (tweets[\"Cluster_unibi_12\"] == 0) & (tweets[\"Cluster_unibi_13\"] == 7) & \n",
    "           (tweets[\"Cluster_unibi_15\"] == 1) & (tweets[\"Cluster_unibitri_6\"] == 1) & \n",
    "           (tweets[\"Cluster_unibitri_8\"] == 3) & (tweets[\"Cluster_unibitri_9\"] == 0) & \n",
    "           (tweets[\"Cluster_unibitri_10\"] == 0) & (tweets[\"Cluster_unibitri_11\"] == 0) & \n",
    "           (tweets[\"Cluster_unibitri_12\"] == 11) & (tweets[\"Cluster_unibitri_9\"] == 0) & \n",
    "           (tweets[\"Cluster_unibitri_13\"] == 3) & (tweets[\"Cluster_unibitri_14\"] == 12) & \n",
    "           (tweets[\"Cluster_unibitri_15\"] == 12), 'Topic'] = 'followers'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros            24472\n",
       "sitios-madrid     1083\n",
       "buenos-dias        407\n",
       "retiro             125\n",
       "gran-via           115\n",
       "followers          105\n",
       "gracias             56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>418</th>\n",
       "      <td>Nuevos seguidores: 2, unfollowers: 1 (14:45) #...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>562</th>\n",
       "      <td>Nuevos seguidores: 1, unfollowers: 1 (13:37) #...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1003</th>\n",
       "      <td>Nuevos seguidores: 1, unfollowers: 0 (11:20) #...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1116</th>\n",
       "      <td>Nuevos seguidores: 0, unfollowers: 1 (11:19) #...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1170</th>\n",
       "      <td>Nuevos seguidores: 1, unfollowers: 0 (11:19) #...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "418   Nuevos seguidores: 2, unfollowers: 1 (14:45) #...\n",
       "562   Nuevos seguidores: 1, unfollowers: 1 (13:37) #...\n",
       "1003  Nuevos seguidores: 1, unfollowers: 0 (11:20) #...\n",
       "1116  Nuevos seguidores: 0, unfollowers: 1 (11:19) #...\n",
       "1170  Nuevos seguidores: 1, unfollowers: 0 (11:19) #..."
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'followers',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'tendencias-twitter'\n",
    "tweets.loc[(tweets[\"Cluster_unig_5\"] == 2) & (tweets[\"Cluster_unig_6\"] == 4) & \n",
    "           (tweets[\"Cluster_unig_7\"] == 5) & (tweets[\"Cluster_unig_8\"] == 7) & \n",
    "           (tweets[\"Cluster_unig_9\"] == 8) & (tweets[\"Cluster_unig_10\"] == 9) &\n",
    "           (tweets[\"Cluster_unig_11\"] == 8) & (tweets[\"Cluster_unig_12\"] == 10) &\n",
    "           (tweets[\"Cluster_unig_13\"] == 9) & (tweets[\"Cluster_unig_14\"] == 12) &\n",
    "           (tweets[\"Cluster_unig_15\"] == 4) &  (tweets[\"Cluster_unig2_5\"] == 4) &\n",
    "           (tweets[\"Cluster_unig2_7\"] == 5) & (tweets[\"Cluster_unig2_8\"] == 5) &\n",
    "           (tweets[\"Cluster_unig2_9\"] == 1) & (tweets[\"Cluster_unig2_10\"] == 7) &\n",
    "           (tweets[\"Cluster_unig2_11\"] == 4) & (tweets[\"Cluster_unig2_12\"] == 4) &\n",
    "           (tweets[\"Cluster_unig2_13\"] == 7) & (tweets[\"Cluster_unig2_14\"] == 4) &\n",
    "           (tweets[\"Cluster_unig2_15\"] == 6),\n",
    "           'Topic'] = 'tendencias-twitter'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros                 23762\n",
       "sitios-madrid          1083\n",
       "tendencias-twitter      710\n",
       "buenos-dias             407\n",
       "retiro                  125\n",
       "gran-via                115\n",
       "followers               105\n",
       "gracias                  56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>'nicholson' es ahora una tendencia en #Madrid</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6992</th>\n",
       "      <td>8 Tendencias de Email marketing  |</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7734</th>\n",
       "      <td>Seguro de Vida; ahora!</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8896</th>\n",
       "      <td>8 Tendencias de Email marketing  |</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9693</th>\n",
       "      <td>#cuentamenoche es ahora una tendencia en #Madrid</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "6        'nicholson' es ahora una tendencia en #Madrid \n",
       "6992                8 Tendencias de Email marketing  | \n",
       "7734                            Seguro de Vida; ahora! \n",
       "8896                8 Tendencias de Email marketing  | \n",
       "9693  #cuentamenoche es ahora una tendencia en #Madrid "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'tendencias-twitter',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'discoteca'\n",
    "tweets.loc[(tweets[\"Cluster_unig_9\"] == 7) & (tweets[\"Cluster_unig_11\"] == 0) &\n",
    "           (tweets[\"Cluster_unig2_6\"] == 2) & (tweets[\"Cluster_unig2_8\"] == 0) &\n",
    "           (tweets[\"Cluster_unig2_10\"] == 3) & (tweets[\"Cluster_unig2_12\"] == 11) &\n",
    "           (tweets[\"Cluster_unig2_13\"] == 4) & (tweets[\"Cluster_unig2_14\"] == 11) &\n",
    "           (tweets[\"Cluster_unig2_15\"] == 13),'Topic'] = 'discoteca'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros                 23591\n",
       "sitios-madrid          1083\n",
       "tendencias-twitter      710\n",
       "buenos-dias             407\n",
       "discoteca               171\n",
       "retiro                  125\n",
       "gran-via                115\n",
       "followers               105\n",
       "gracias                  56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>327</th>\n",
       "      <td>Listas #show meclub! Te esperamos con la mejor...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>328</th>\n",
       "      <td>Zoe-  RESERVADOS y Listas! - SÁBADO- 100 eu re...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>332</th>\n",
       "      <td>No te pierdas este finde en en Riviera Madrid ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>347</th>\n",
       "      <td>Este sabado la mejor fiesta con mejor espectac...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>360</th>\n",
       "      <td>Este sabado SHOW ME' en La Tartufo!! TRAP, #FU...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               content\n",
       "327  Listas #show meclub! Te esperamos con la mejor...\n",
       "328  Zoe-  RESERVADOS y Listas! - SÁBADO- 100 eu re...\n",
       "332  No te pierdas este finde en en Riviera Madrid ...\n",
       "347  Este sabado la mejor fiesta con mejor espectac...\n",
       "360  Este sabado SHOW ME' en La Tartufo!! TRAP, #FU..."
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'discoteca',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'trafico' (como aparece en todos, cogeremos el primero)\n",
    "tweets.loc[(tweets[\"Cluster_unig_5\"] == 1),'Topic'] = 'trafico'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros                 22992\n",
       "sitios-madrid          1083\n",
       "tendencias-twitter      710\n",
       "trafico                 599\n",
       "buenos-dias             407\n",
       "discoteca               171\n",
       "retiro                  125\n",
       "gran-via                115\n",
       "followers               105\n",
       "gracias                  56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2549</th>\n",
       "      <td>#RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2568</th>\n",
       "      <td>#RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2869</th>\n",
       "      <td>#RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2873</th>\n",
       "      <td>#RETENCIÓN nivel ROJO en AUTOPISTA / AUTOVÍA #...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3148</th>\n",
       "      <td>#RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                content\n",
       "2549  #RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...\n",
       "2568  #RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...\n",
       "2869  #RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV...\n",
       "2873  #RETENCIÓN nivel ROJO en AUTOPISTA / AUTOVÍA #...\n",
       "3148  #RETENCIÓN nivel AMARILLO en AUTOPISTA / AUTOV..."
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'trafico',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Asignación de tema 'aeropuerto'\n",
    "tweets.loc[(tweets[\"Cluster_unig_14\"] == 1) & (tweets[\"Cluster_unig_15\"] == 5) &\n",
    "           (tweets[\"Cluster_unibi_7\"] == 3) & (tweets[\"Cluster_unibi_14\"] == 13),\n",
    "           'Topic'] = 'aeropuerto'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "otros                 22880\n",
       "sitios-madrid          1083\n",
       "tendencias-twitter      710\n",
       "trafico                 599\n",
       "buenos-dias             407\n",
       "discoteca               171\n",
       "retiro                  125\n",
       "gran-via                115\n",
       "aeropuerto              112\n",
       "followers               105\n",
       "gracias                  56\n",
       "Name: Topic, dtype: int64"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets['Topic'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8034</th>\n",
       "      <td>I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9904</th>\n",
       "      <td>Pues ya estamos aquí (@ Aeropuerto Adolfo Suár...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10525</th>\n",
       "      <td>I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10621</th>\n",
       "      <td>I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10973</th>\n",
       "      <td>Back (@ Aeropuerto Adolfo Suárez Madrid-Baraja...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 content\n",
       "8034   I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...\n",
       "9904   Pues ya estamos aquí (@ Aeropuerto Adolfo Suár...\n",
       "10525  I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...\n",
       "10621  I'm at Aeropuerto Adolfo Suárez Madrid-Barajas...\n",
       "10973  Back (@ Aeropuerto Adolfo Suárez Madrid-Baraja..."
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets.loc[tweets[\"Topic\"] == 'aeropuerto',[\"content\"]].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(22880, 3)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Repetiremos el ejercicio de clustering con el subconjunto catalogado como 'otros', para ver que\n",
    "# no hay más clústeres triviales\n",
    "ot_tweets = tweets.loc[tweets[\"Topic\"] == 'otros',[\"content\",\"Latitude\",\"Longitude\"]]\n",
    "ot_tweets.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>content</th>\n",
       "      <th>Latitude</th>\n",
       "      <th>Longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Regalo para socios: 'Talento a la fuga', el li...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>La recuperación del ‘ladrillo’ dispara el 144,...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TELEVISIÓN SERIES - Cuenta atrás para conocer ...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Nuestra taza de hoy es un homenaje a uno de lo...</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>@MCadepe nos vemos el 19 de mayooo!!!</td>\n",
       "      <td>40.416</td>\n",
       "      <td>-3.703</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             content  Latitude  Longitude\n",
       "0  Regalo para socios: 'Talento a la fuga', el li...    40.416     -3.703\n",
       "1  La recuperación del ‘ladrillo’ dispara el 144,...    40.416     -3.703\n",
       "2  TELEVISIÓN SERIES - Cuenta atrás para conocer ...    40.416     -3.703\n",
       "3  Nuestra taza de hoy es un homenaje a uno de lo...    40.416     -3.703\n",
       "4              @MCadepe nos vemos el 19 de mayooo!!!    40.416     -3.703"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ot_tweets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Extraemos los tweets y los guardamos en una lista\n",
    "ot_tweets_text = ot_tweets[\"content\"].tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Convertimos la lista en la matriz Tf-idf\n",
    "vectorizer_unig_ot = TfidfVectorizer(\n",
    "                max_df = 0.95,\n",
    "                max_features = 30000,\n",
    "                analyzer = 'word',\n",
    "                tokenizer = tokenize,\n",
    "                lowercase = True,\n",
    "                use_idf = True,\n",
    "                stop_words = spanish_stopwords,\n",
    "                ngram_range=(1,1))\n",
    "mx_tfidf_unig_ot = vectorizer_unig_ot.fit_transform(ot_tweets_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(22880, 29442)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mx_tfidf_unig_ot.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Número de clusters: 5\n",
      "0    12166\n",
      "1     4933\n",
      "3     2255\n",
      "2     2206\n",
      "4     1320\n",
      "Name: Cluster_uni_ot_5, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " si\n",
      " via\n",
      " per\n",
      " graci\n",
      " dia\n",
      " buen\n",
      " quier\n",
      " hac\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " mas\n",
      " hoy\n",
      " com\n",
      " si\n",
      " dia\n",
      " hac\n",
      " mejor\n",
      " nuestr\n",
      " vid\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " mas\n",
      " mejor\n",
      " dia\n",
      " hoy\n",
      " com\n",
      " hac\n",
      " nuestr\n",
      " madr\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " real\n",
      " palaci\n",
      " plaz\n",
      " mas\n",
      " adecc\n",
      " dia\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " graci\n",
      " dias\n",
      " buen\n",
      " mejor\n",
      " per\n",
      " madr\n",
      "\n",
      "\n",
      "Número de clusters: 6\n",
      "5    11906\n",
      "1     4479\n",
      "2     2296\n",
      "4     1850\n",
      "3     1245\n",
      "0     1104\n",
      "Name: Cluster_uni_ot_6, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " com\n",
      " per\n",
      " par\n",
      " ver\n",
      " pued\n",
      " mas\n",
      " hac\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " mas\n",
      " com\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      " vid\n",
      " stigmab\n",
      " buen\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " real\n",
      " palaci\n",
      " plaz\n",
      " mas\n",
      " adecc\n",
      " par\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " graci\n",
      " dias\n",
      " buen\n",
      " mejor\n",
      " mas\n",
      " madr\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hoy\n",
      " dia\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " buen\n",
      " mas\n",
      " feliz\n",
      " tierr\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " mas\n",
      " com\n",
      " via\n",
      " graci\n",
      " mejor\n",
      " per\n",
      " buen\n",
      " nuev\n",
      " quier\n",
      "\n",
      "\n",
      "Número de clusters: 7\n",
      "1    11509\n",
      "2     4570\n",
      "3     2182\n",
      "5     2122\n",
      "6     1000\n",
      "4      904\n",
      "0      593\n",
      "Name: Cluster_uni_ot_7, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " quier\n",
      " …\n",
      " si\n",
      " mas\n",
      " com\n",
      " per\n",
      " hac\n",
      " cuand\n",
      " tod\n",
      " ser\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " tod\n",
      " via\n",
      " graci\n",
      " per\n",
      " buen\n",
      " mejor\n",
      " dia\n",
      " hac\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " …\n",
      " tod\n",
      " mas\n",
      " com\n",
      " dia\n",
      " vid\n",
      " graci\n",
      " hac\n",
      " nuev\n",
      " mejor\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " real\n",
      " palaci\n",
      " plaz\n",
      " adecc\n",
      " mas\n",
      " mejor\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hoy\n",
      " …\n",
      " dia\n",
      " par\n",
      " toc\n",
      " com\n",
      " tod\n",
      " buen\n",
      " h\n",
      " madr\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " mas\n",
      " mejor\n",
      " nuestr\n",
      " com\n",
      " hac\n",
      " madr\n",
      " dia\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " com\n",
      " per\n",
      " pued\n",
      " ver\n",
      " mas\n",
      " tien\n",
      " tod\n",
      " ser\n",
      "\n",
      "\n",
      "Número de clusters: 8\n",
      "6    11688\n",
      "5     4480\n",
      "3     2077\n",
      "2     1282\n",
      "7     1126\n",
      "1      814\n",
      "4      761\n",
      "0      652\n",
      "Name: Cluster_uni_ot_8, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " buen\n",
      " …\n",
      " dias\n",
      " noch\n",
      " par\n",
      " tod\n",
      " com\n",
      " viern\n",
      " hoy\n",
      " graci\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " tierr\n",
      " par\n",
      " feliz\n",
      " madr\n",
      " buen\n",
      " libr\n",
      " com\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " madr\n",
      " par\n",
      " much\n",
      " pued\n",
      " si\n",
      " mejor\n",
      " bonit\n",
      " person\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " palaci\n",
      " adecc\n",
      " real\n",
      " plaz\n",
      " par\n",
      " mejor\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " hac\n",
      " …\n",
      " par\n",
      " com\n",
      " si\n",
      " mas\n",
      " tod\n",
      " años\n",
      " cuand\n",
      " falt\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " …\n",
      " par\n",
      " com\n",
      " hoy\n",
      " si\n",
      " nuev\n",
      " nuestr\n",
      " mejor\n",
      " vid\n",
      " sab\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " par\n",
      " com\n",
      " si\n",
      " via\n",
      " per\n",
      " graci\n",
      " mejor\n",
      " quier\n",
      " nuev\n",
      " hoy\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " graci\n",
      " list\n",
      " per\n",
      " mejor\n",
      " madr\n",
      " si\n",
      "\n",
      "\n",
      "Número de clusters: 9\n",
      "2    9759\n",
      "8    3651\n",
      "0    1978\n",
      "1    1921\n",
      "5    1657\n",
      "4    1299\n",
      "6    1110\n",
      "3     816\n",
      "7     689\n",
      "Name: Cluster_uni_ot_9, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " palaci\n",
      " real\n",
      " adecc\n",
      " plaz\n",
      " estacion\n",
      " estadi\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " tod\n",
      " mejor\n",
      " nuestr\n",
      " hac\n",
      " madr\n",
      " hoy\n",
      " via\n",
      " nuev\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " via\n",
      " graci\n",
      " per\n",
      " mejor\n",
      " hac\n",
      " nuev\n",
      " quier\n",
      " hoy\n",
      " cuand\n",
      " tien\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " dia\n",
      " …\n",
      " hoy\n",
      " tierr\n",
      " feliz\n",
      " madr\n",
      " par\n",
      " buen\n",
      " libr\n",
      " mas\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " mas\n",
      " …\n",
      " much\n",
      " madr\n",
      " par\n",
      " pued\n",
      " mejor\n",
      " person\n",
      " bonit\n",
      " quier\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " tod\n",
      " si\n",
      " …\n",
      " quier\n",
      " sab\n",
      " per\n",
      " pued\n",
      " tien\n",
      " hac\n",
      " porqu\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " com\n",
      " …\n",
      " par\n",
      " si\n",
      " hac\n",
      " per\n",
      " quier\n",
      " tod\n",
      " hoy\n",
      " madr\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " buen\n",
      " …\n",
      " dias\n",
      " noch\n",
      " tod\n",
      " par\n",
      " com\n",
      " viern\n",
      " hoy\n",
      " graci\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " …\n",
      " hoy\n",
      " nuev\n",
      " mejor\n",
      " nuestr\n",
      " hac\n",
      " stigmab\n",
      " sol\n",
      " graci\n",
      " vid\n",
      "\n",
      "\n",
      "Número de clusters: 10\n",
      "0    9475\n",
      "7    3561\n",
      "3    2003\n",
      "6    1888\n",
      "8    1804\n",
      "1    1085\n",
      "4    1030\n",
      "2     745\n",
      "5     673\n",
      "9     616\n",
      "Name: Cluster_uni_ot_10, dtype: int64\n",
      "Cluster 0 : Palabras más frecuentes :\n",
      " via\n",
      " dia\n",
      " buen\n",
      " per\n",
      " nuev\n",
      " quier\n",
      " hac\n",
      " hoy\n",
      " tien\n",
      " cuand\n",
      "\n",
      "Cluster 1 : Palabras más frecuentes :\n",
      " tod\n",
      " …\n",
      " par\n",
      " sab\n",
      " dias\n",
      " buen\n",
      " per\n",
      " madr\n",
      " mund\n",
      " hoy\n",
      "\n",
      "Cluster 2 : Palabras más frecuentes :\n",
      " mejor\n",
      " …\n",
      " madr\n",
      " par\n",
      " sal\n",
      " mas\n",
      " tod\n",
      " sab\n",
      " mund\n",
      " list\n",
      "\n",
      "Cluster 3 : Palabras más frecuentes :\n",
      " madr\n",
      " in\n",
      " …\n",
      " españ\n",
      " palaci\n",
      " real\n",
      " plaz\n",
      " adecc\n",
      " dia\n",
      " estadi\n",
      "\n",
      "Cluster 4 : Palabras más frecuentes :\n",
      " si\n",
      " …\n",
      " quier\n",
      " per\n",
      " com\n",
      " ver\n",
      " pued\n",
      " tien\n",
      " hac\n",
      " ser\n",
      "\n",
      "Cluster 5 : Palabras más frecuentes :\n",
      " graci\n",
      " …\n",
      " segu\n",
      " tod\n",
      " tuitutil\n",
      " devuelv\n",
      " follow\n",
      " brev\n",
      " buen\n",
      " mil\n",
      "\n",
      "Cluster 6 : Palabras más frecuentes :\n",
      " par\n",
      " …\n",
      " dia\n",
      " hoy\n",
      " mas\n",
      " hac\n",
      " nuestr\n",
      " com\n",
      " nuev\n",
      " via\n",
      "\n",
      "Cluster 7 : Palabras más frecuentes :\n",
      " …\n",
      " hoy\n",
      " dia\n",
      " nuev\n",
      " hac\n",
      " nuestr\n",
      " stigmab\n",
      " vid\n",
      " sol\n",
      " buen\n",
      "\n",
      "Cluster 8 : Palabras más frecuentes :\n",
      " mas\n",
      " com\n",
      " …\n",
      " hac\n",
      " dia\n",
      " much\n",
      " pued\n",
      " quier\n",
      " per\n",
      " hoy\n",
      "\n",
      "Cluster 9 : Palabras más frecuentes :\n",
      " ten\n",
      " mañan\n",
      " …\n",
      " par\n",
      " hoy\n",
      " per\n",
      " com\n",
      " mas\n",
      " dia\n",
      " sol\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Realización del clustering con k-means. No se detectan nuevas agrupaciones relevantes\n",
    "kmeans_function(11, mx_tfidf_unig_ot, ot_tweets, vectorizer_unig_ot, \"uni_ot\", 3333)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Se exportan los tweets con la nueva columna a un fichero Excel\n",
    "tweets[['content','Latitude','Longitude','Topic']].to_excel('ficheros/TweetsConTopic/tweets_spainGeo_topic.xlsx', header=True, index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Topic modeling con LDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Se redefine la función de tokenización para tener en cuenta las palabras de parada, sin hacer stemming\n",
    "# (para tener más términos para buscar los temas)\n",
    "def tokenize_stop(text):\n",
    "    # Eliminamos lo que no sean palabras\n",
    "    text = ''.join([c for c in text if c not in non_words])\n",
    "    # Eliminamos las palabras de parada\n",
    "    text = ''.join([c for c in text if c not in spanish_stopwords])\n",
    "\n",
    "    # Tokenización\n",
    "    tokens = tknzr.tokenize(text)\n",
    "    \n",
    "    return tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Topic 0: Words: \n",
      "['d', 'l', 'dl', 'Mdrid', '…', 'ls', 'El', 'n', 'cn', 'un']\n",
      "Topic 1: Words: \n",
      "['d', 'l', 'n', '…', 'pr', 'qu', 'ls', 'un', 'nivl', 's']\n",
      "Topic 2: Words: \n",
      "['Mdrid', 'in', 'd', 'n', 'l', 'D', 'ls', 'cn', '…', 'Sl']\n",
      "Topic 3: Words: \n",
      "['l', 'qu', 's', 'n', '…', 'd', 'pr', 'un', 'ls', 'm']\n",
      "Topic 4: Words: \n",
      "['…', 'd', 'l', 'n', 'pr', 'cn', 'ES', 'Stigmbs', 'Mdrid', 'L']\n",
      "Topic 5: Words: \n",
      "['n', 'l', 'DGT', 'pk', 'DGTMdrid', 'RETENCIÓN', 'AUTOPISTA', 'AUTOVÍA', 'AMARILLO', 'A']\n",
      "Topic 6: Words: \n",
      "['…', 'd', 'dís', 'Buns', 'mdrid', 'pr', 'fin', 'smn', 'ls', 'virns']\n",
      "Topic 7: Words: \n",
      "['Mdrid', 'n', 'un', 's', 'hr', 'tndnci', 't', 'Im', 'd', 'Arpurt']\n"
     ]
    }
   ],
   "source": [
    "# Generamos una lista de tweets donde cada tweet sea una lista de palabras\n",
    "words_tweet = [tokenize_stop(tweet) for tweet in tweets_text]\n",
    "# Se genera el diccionario\n",
    "dictionary = corpora.Dictionary(words_tweet)\n",
    "# Se convierte el diccionario en BOW (Bag of Words)\n",
    "tweet_term_matrix = [dictionary.doc2bow(doc) for doc in words_tweet]\n",
    "# Lanzamos el modelo (tarda bastante, más de 1 hora)\n",
    "ldamodel = models.ldamodel.LdaModel(tweet_term_matrix, num_topics=8, id2word = dictionary, passes=100)\n",
    "\n",
    "# Impresión de resultados (no hay mucha precisión en los resultados; sólo está claro el tema de 'trafico',\n",
    "# que es el Topic 5, y hay ciertos indicios del tema 'tendencias' y 'aeropuerto' en el Topic 7, pero no\n",
    "# se distinguen).\n",
    "for topic in ldamodel.show_topics(num_topics=10, formatted=False, num_words=10):\n",
    "    print(\"Topic {}: Words: \".format(topic[0]))\n",
    "    topicwords = [w for (w, val) in topic[1]]\n",
    "    print(topicwords)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Topic 0: Words: \n",
      "['de', 'buen', 'a', 'dias', 'palaci', 'el', 'un', 'los', 'graci', 'adolf']\n",
      "Topic 1: Words: \n",
      "['que', 'no', '…', 'lo', 'y', 'a', 'te', 'es', 'spain', 'si']\n",
      "Topic 2: Words: \n",
      "['…', 'y', 'de', 'con', 'dia', 'del', 'el', 'madr', 'par', 'un']\n",
      "Topic 3: Words: \n",
      "['de', 'la', 'el', 'en', '…', 'a', 'y', 'los', 'que', 'las']\n",
      "Topic 4: Words: \n",
      "['de', '…', 'la', 'y', 'que', 'a', 'el', 'me', 'en', 'mi']\n",
      "Topic 5: Words: \n",
      "['de', 'madr', 'la', 'el', 'en', 'in', '…', 'del', 'españ', 'las']\n",
      "Topic 6: Words: \n",
      "['en', 'es', 'una', 'ahor', 'madr', 'tendenci', 'nuev', 'estadi', 'aeropuert', 'suarez']\n",
      "Topic 7: Words: \n",
      "['en', 'al', 'nivel', 'pk', 'dgt', 'retencion', 'dgtmadr', 'autop', 'autov', 'amarill']\n"
     ]
    }
   ],
   "source": [
    "# Se repetirá el proceso haciendo también stemming.\n",
    "words_tweet_stem = [tokenize(tweet) for tweet in tweets_text]\n",
    "dictionary_stem = corpora.Dictionary(words_tweet_stem)\n",
    "tweet_term_matrix_stem = [dictionary_stem.doc2bow(doc) for doc in words_tweet_stem]\n",
    "ldamodel_stem = models.ldamodel.LdaModel(tweet_term_matrix_stem, num_topics=8, id2word = dictionary_stem, passes=100)\n",
    "\n",
    "# Impresión de resultados (no hay mucha precisión en los resultados nuevamente; el único claro vuelve a ser\n",
    "# 'trafico', que es el Topic 7).\n",
    "for topic_stem in ldamodel_stem.show_topics(num_topics=10, formatted=False, num_words=10):\n",
    "    print(\"Topic {}: Words: \".format(topic_stem[0]))\n",
    "    topicwords_stem = [w for (w, val) in topic_stem[1]]\n",
    "    print(topicwords_stem)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Licencia"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "El notebook está licenciado libremente bajo la licencia [Creative Commons Attribution Share-Alike](https://creativecommons.org/licenses/by/2.0/).\n",
    "\n",
    "La base del código empleado procede del trabajo de Yogesh Kulkarni llamado [Sentiment Analysis of Twitter Posts on Chennai Floods using Python](https://www.analyticsvidhya.com/blog/2017/01/sentiment-analysis-of-twitter-posts-on-chennai-floods-using-python/) y del trabajo de Brandon Rose llamado [Document Clustering with Python](http://brandonrose.org/clustering)\n",
    "\n",
    "© 2017 - Juan Bermudo Mera, Margarita Bolívar Jiménez, Lourdes Fernández Nieto, Ramón Pérez Hernández.\n",
    "\n",
    "Universidad Politécnica de Madrid."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}