Bag-of-Words Analysis Haiku (unsupervised, tf-idf & kmeans, Silhouette).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analysis of Philadelphia Reflections Haiku\n",
    "\n",
    "> # Question: how do the haiku cluster?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------------------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from __future__ import division\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy\n",
    "import matplotlib.pyplot as plt\n",
    "from os.path import expanduser"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.cluster                 import KMeans\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer # combines counting and normalizing\n",
    "from sklearn.pipeline                import Pipeline\n",
    "\n",
    "from sklearn.metrics                 import silhouette_samples, silhouette_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "#%qtconsole"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read in the Philadelphia Reflections haiku"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "data_folder = 'corpus_haiku'\n",
    "\n",
    "# Note the use of read(); readlines() does not work\n",
    "\n",
    "documents = [open(os.path.join(data_folder, filename)).read() \n",
    "             for filename in os.listdir(data_folder)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Unsupervised Feature Selection\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "by choosing a relatively low **max_df** for the vectorizer we remove a majority of the frequently-occuring function words we used for authorship determination and the TF-IDF algorithm will identify the important words.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Find the K with the best silhouette score and greatest separation   \n",
    "\n",
    "for Various K values and distance metrics "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/numpy/core/_methods.py:59: RuntimeWarning: Mean of empty slice.\n",
      "  warnings.warn(\"Mean of empty slice.\", RuntimeWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The silhouette scores in descending order for metric = cosine\n",
      "2 \t0.848970196154\n",
      "3 \t0.806930242385\n",
      "5 \t0.738301152431\n",
      "4 \t0.736364293384\n",
      "6 \t0.727508882456\n",
      "8 \t0.719759729611\n",
      "7 \t0.716139927555\n",
      "9 \t0.71562733291\n",
      "12 \t0.705375793795\n",
      "14 \t0.705203316425\n",
      "13 \t0.701534842864\n",
      "18 \t0.683752829\n",
      "20 \t0.682600485264\n",
      "19 \t0.679222449718\n",
      "11 \t0.673489128659\n",
      "10 \t0.672596251047\n",
      "16 \t0.669573615169\n",
      "17 \t0.667909194446\n",
      "15 \t0.665960262768\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.cm as cm\n",
    "\n",
    "# ========================= PARAMETERS =========================\n",
    "\n",
    "kmeans_kwargs = dict(copy_x = False, # improve numerical accuracy\n",
    "                     n_jobs = -1     # use all available cores\n",
    "                    )\n",
    "\n",
    "tfidf_kwargs  = dict(max_df      = 0.4,   # select only words that appear in\n",
    "                                          # no more than 40% of the documents\n",
    "                     stop_words  = 'english'\n",
    "                    )\n",
    "\n",
    "range_n_clusters = range(2,20+1) # K values\n",
    "\n",
    "#metric  = 'euclidean'\n",
    "#metric  = 'cityblock'\n",
    "metric  = 'cosine'\n",
    "#metric  = 'l1'\n",
    "#metric  = 'l2'\n",
    "#metric  = 'manhattan'\n",
    "\n",
    "verbose = False        # print out lovely graphs and summary?\n",
    "\n",
    "# ========================= PARAMETERS =========================\n",
    "\n",
    "summary = []                     # list of results\n",
    "\n",
    "for n_clusters in range_n_clusters:\n",
    "    # Initialize the clusterer with n_clusters value and a random generator\n",
    "    # seed of 10 for reproducibility.\n",
    "    vectorizer = TfidfVectorizer(**tfidf_kwargs)\n",
    "    clusterer  = KMeans(n_clusters=n_clusters, random_state=10, **kmeans_kwargs)\n",
    "    pipeline   = Pipeline([ ('tfidf',  vectorizer),\n",
    "                            ('kmeans', clusterer)])\n",
    "    \n",
    "    pipeline.fit(documents)\n",
    "    \n",
    "    cluster_labels = pipeline.predict(documents)\n",
    "    X              = pipeline.transform(documents)\n",
    "    \n",
    "    # The silhouette_score gives the average value for all the samples.\n",
    "    # This gives a perspective into the density and separation of the formed\n",
    "    # clusters\n",
    "    \n",
    "    silhouette_avg = silhouette_score(X, cluster_labels, metric=metric)\n",
    "\n",
    "    # Compute the silhouette scores for each sample\n",
    "    sample_silhouette_values = silhouette_samples(X, cluster_labels, metric=metric)\n",
    "    \n",
    "    silhouette_avg_pos = np.array(sample_silhouette_values)[np.array(sample_silhouette_values)>0].mean()\n",
    "    silhouette_avg_neg = np.array(sample_silhouette_values)[np.array(sample_silhouette_values)<0].mean()\n",
    "    if np.isnan(silhouette_avg_pos): silhouette_avg_pos = 0\n",
    "    if np.isnan(silhouette_avg_neg): silhouette_avg_neg = 0\n",
    "        \n",
    "    if verbose: print(\"For n_clusters = {} the average silhouette_score is : {:.4f}; pos:{:.4f}, neg:{:.4f}\".\\\n",
    "                      format(n_clusters, silhouette_avg, silhouette_avg_pos, silhouette_avg_neg))\n",
    "    summary.append((n_clusters, silhouette_avg, silhouette_avg_pos, silhouette_avg_neg))\n",
    "    \n",
    "    if verbose:\n",
    "        # Create a subplot with 1 row and 2 columns\n",
    "        fig, (ax1, ax2) = plt.subplots(1, 2)\n",
    "        fig.set_size_inches(18, 7)\n",
    "\n",
    "        # The 1st subplot is the silhouette plot\n",
    "        # ===============\n",
    "        # The silhouette coefficient can range from -1, 1 but in this example all\n",
    "        # lie within [-0.1, 1]\n",
    "        ax1.set_xlim([-1, 1])\n",
    "\n",
    "        # The (n_clusters+1)*10 is for inserting blank space between silhouette\n",
    "        # plots of individual clusters, to demarcate them clearly.\n",
    "        ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])\n",
    "\n",
    "        y_lower = 10\n",
    "        for i in range(n_clusters):\n",
    "            # Aggregate the silhouette scores for samples belonging to\n",
    "            # cluster i, and sort them\n",
    "            ith_cluster_silhouette_values = \\\n",
    "                sample_silhouette_values[cluster_labels == i]\n",
    "\n",
    "            ith_cluster_silhouette_values.sort()\n",
    "\n",
    "            size_cluster_i = ith_cluster_silhouette_values.shape[0]\n",
    "            y_upper = y_lower + size_cluster_i\n",
    "\n",
    "            color = cm.spectral(float(i) / n_clusters)\n",
    "            ax1.fill_betweenx(np.arange(y_lower, y_upper),\n",
    "                              0, ith_cluster_silhouette_values,\n",
    "                              facecolor=color, edgecolor=color, alpha=0.7)\n",
    "\n",
    "            # Label the silhouette plots with their cluster numbers at the middle\n",
    "            ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))\n",
    "\n",
    "            # Compute the new y_lower for next plot\n",
    "            y_lower = y_upper + 10  # 10 for the 0 samples\n",
    "\n",
    "        ax1.set_title(\"The silhouette plot for the various clusters.\")\n",
    "        ax1.set_xlabel(\"The silhouette coefficient values\")\n",
    "        ax1.set_ylabel(\"Cluster label\")\n",
    "\n",
    "        # The vertical line for average silhoutte score of all the values\n",
    "        ax1.axvline(x=silhouette_avg, color=\"red\", linestyle=\"--\")\n",
    "\n",
    "        if silhouette_avg_pos>0: ax1.axvline(x=silhouette_avg_pos, color=\"blue\", linestyle=\"--\")\n",
    "        if silhouette_avg_neg<0: ax1.axvline(x=silhouette_avg_neg, color=\"blue\", linestyle=\"--\")\n",
    "\n",
    "        ax1.set_yticks([])  # Clear the yaxis labels / ticks\n",
    "        ax1.set_xticks([-1, -0.8, -0.6, -0.4, -0.2, 0, 0.2, 0.4, 0.6, 0.8, 1])\n",
    "\n",
    "        # 2nd Plot showing the actual clusters formed\n",
    "        # ========\n",
    "        colors = cm.spectral(cluster_labels.astype(float) / n_clusters)\n",
    "        ax2.scatter(X[:, 0], X[:, 1], marker='.', s=30, lw=0, alpha=0.7,c=colors)\n",
    "\n",
    "        # Labeling the clusters\n",
    "        centers = clusterer.cluster_centers_\n",
    "        # Draw white circles at cluster centers\n",
    "        ax2.scatter(centers[:, 0], centers[:, 1],\n",
    "                    marker='o', c=\"white\", alpha=1, s=200)\n",
    "\n",
    "        for i, c in enumerate(centers):\n",
    "            ax2.scatter(c[0], c[1], marker='$%d$' % i, alpha=1, s=50)\n",
    "\n",
    "        ax2.set_title(\"The visualization of the clustered data.\")\n",
    "        ax2.set_xlabel(\"Feature space for the 1st feature\")\n",
    "        ax2.set_ylabel(\"Feature space for the 2nd feature\")\n",
    "\n",
    "        plt.suptitle((\"Silhouette analysis for KMeans clustering on sample data \"\n",
    "                      \"with n_clusters = %d (metric=%s)\" % (n_clusters,metric)),\n",
    "                     fontsize=14, fontweight='bold')\n",
    "\n",
    "        plt.show()\n",
    "\n",
    "x_vals = [x[0] for x in summary]\n",
    "y_avg  = [x[1] for x in summary]\n",
    "y_pos  = [x[2] for x in summary]\n",
    "y_neg  = [x[3] for x in summary]\n",
    "if verbose:\n",
    "    plt.plot(x_vals,y_avg, label='avg')\n",
    "    plt.xlim([0,len(x_vals)+2])\n",
    "    plt.axvline(x=np.argmax(y_avg)+2, color=\"red\", linestyle=\"--\")\n",
    "    plt.title('Silhouette Scores by KMeans K value')\n",
    "    plt.xlabel('KMeans K value')\n",
    "    plt.ylabel('Silhoutette Score')\n",
    "    plt.plot(x_vals,y_pos, label='pos', linestyle=\"--\")\n",
    "    # plt.plot(x_vals,y_neg)\n",
    "    plt.legend(loc='best')\n",
    "    plt.show()\n",
    "\n",
    "print(\"The silhouette scores in descending order for metric = {}\".format(metric))\n",
    "import operator\n",
    "ordered_summary = sorted(summary, key = operator.itemgetter(1), reverse=True)\n",
    "for cnum, avg, _, _ in ordered_summary:\n",
    "    print cnum, \"\\t\", avg"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Fit using the top 60%-tile Silhouette-Average-Scored K valued clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEZCAYAAACNebLAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa79168d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 265 documents\n",
      "1) time (score: 0.0155)\n",
      "2) friends (score: 0.0153)\n",
      "3) end (score: 0.0146)\n",
      "4) life (score: 0.0141)\n",
      "5) best (score: 0.0136)\n",
      "Average cluster score 0.0146\n",
      "\n",
      "For cluster 1, 212 documents\n",
      "1) new (score: 0.0270)\n",
      "2) old (score: 0.0225)\n",
      "3) shout (score: 0.0217)\n",
      "4) high (score: 0.0193)\n",
      "5) fast (score: 0.0191)\n",
      "Average cluster score 0.0219\n",
      "\n",
      "The clusters by average score, descending\n",
      "1 \t0.0219447475182\n",
      "0 \t0.0146430109625\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEZCAYAAACNebLAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa76a7a50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 172 documents\n",
      "1) stop (score: 0.0221)\n",
      "2) words (score: 0.0173)\n",
      "3) old (score: 0.0167)\n",
      "4) eyes (score: 0.0160)\n",
      "5) food (score: 0.0159)\n",
      "Average cluster score 0.0176\n",
      "\n",
      "For cluster 1, 139 documents\n",
      "1) new (score: 0.0331)\n",
      "2) change (score: 0.0245)\n",
      "3) old (score: 0.0213)\n",
      "4) prize (score: 0.0199)\n",
      "5) way (score: 0.0196)\n",
      "Average cluster score 0.0237\n",
      "\n",
      "For cluster 2, 166 documents\n",
      "1) time (score: 0.0234)\n",
      "2) friends (score: 0.0222)\n",
      "3) ends (score: 0.0212)\n",
      "4) try (score: 0.0205)\n",
      "5) best (score: 0.0173)\n",
      "Average cluster score 0.0209\n",
      "\n",
      "The clusters by average score, descending\n",
      "1 \t0.023660507073\n",
      "2 \t0.0209234752027\n",
      "0 \t0.0175982687144\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEZCAYAAACNebLAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEvP8DxrBAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefbe4e65d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 64 documents\n",
      "1) like (score: 0.0309)\n",
      "2) bite (score: 0.0234)\n",
      "3) fresh (score: 0.0222)\n",
      "4) wise (score: 0.0217)\n",
      "5) words (score: 0.0195)\n",
      "Average cluster score 0.0235\n",
      "\n",
      "For cluster 1, 104 documents\n",
      "1) check (score: 0.0407)\n",
      "2) good (score: 0.0342)\n",
      "3) win (score: 0.0241)\n",
      "4) old (score: 0.0236)\n",
      "5) new (score: 0.0222)\n",
      "Average cluster score 0.0290\n",
      "\n",
      "For cluster 2, 152 documents\n",
      "1) time (score: 0.0267)\n",
      "2) old (score: 0.0232)\n",
      "3) age (score: 0.0229)\n",
      "4) slow (score: 0.0212)\n",
      "5) years (score: 0.0201)\n",
      "Average cluster score 0.0228\n",
      "\n",
      "For cluster 3, 157 documents\n",
      "1) try (score: 0.0269)\n",
      "2) cash (score: 0.0236)\n",
      "3) shout (score: 0.0189)\n",
      "4) stop (score: 0.0188)\n",
      "5) turn (score: 0.0168)\n",
      "Average cluster score 0.0210\n",
      "\n",
      "The clusters by average score, descending\n",
      "1 \t0.0289506766664\n",
      "0 \t0.0235300188073\n",
      "2 \t0.0228195949642\n",
      "3 \t0.0209846782414\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEZCAYAAAB1mUk3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa79168d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 104 documents\n",
      "1) years (score: 0.0292)\n",
      "2) change (score: 0.0277)\n",
      "3) old (score: 0.0261)\n",
      "4) youth (score: 0.0248)\n",
      "5) truth (score: 0.0212)\n",
      "Average cluster score 0.0258\n",
      "\n",
      "For cluster 1, 115 documents\n",
      "1) game (score: 0.0332)\n",
      "2) fans (score: 0.0254)\n",
      "3) play (score: 0.0238)\n",
      "4) win (score: 0.0218)\n",
      "5) seek (score: 0.0198)\n",
      "Average cluster score 0.0248\n",
      "\n",
      "For cluster 2, 108 documents\n",
      "1) old (score: 0.0270)\n",
      "2) end (score: 0.0258)\n",
      "3) young (score: 0.0207)\n",
      "4) rich (score: 0.0193)\n",
      "5) life (score: 0.0186)\n",
      "Average cluster score 0.0223\n",
      "\n",
      "For cluster 3, 84 documents\n",
      "1) stop (score: 0.0396)\n",
      "2) cash (score: 0.0370)\n",
      "3) quick (score: 0.0282)\n",
      "4) shout (score: 0.0243)\n",
      "5) charge (score: 0.0224)\n",
      "Average cluster score 0.0303\n",
      "\n",
      "For cluster 4, 66 documents\n",
      "1) wait (score: 0.0425)\n",
      "2) voice (score: 0.0406)\n",
      "3) choice (score: 0.0388)\n",
      "4) don (score: 0.0278)\n",
      "5) hear (score: 0.0234)\n",
      "Average cluster score 0.0346\n",
      "\n",
      "The clusters by average score, descending\n",
      "4 \t0.0346152674689\n",
      "3 \t0.030311436122\n",
      "0 \t0.0257990172278\n",
      "1 \t0.0247763234876\n",
      "2 \t0.0223023828006\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEZCAYAAAB1mUk3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa5caba90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 121 documents\n",
      "1) win (score: 0.0310)\n",
      "2) play (score: 0.0261)\n",
      "3) game (score: 0.0222)\n",
      "4) gold (score: 0.0202)\n",
      "5) home (score: 0.0196)\n",
      "Average cluster score 0.0238\n",
      "\n",
      "For cluster 1, 66 documents\n",
      "1) gals (score: 0.0321)\n",
      "2) guys (score: 0.0295)\n",
      "3) soon (score: 0.0279)\n",
      "4) hot (score: 0.0244)\n",
      "5) old (score: 0.0225)\n",
      "Average cluster score 0.0273\n",
      "\n",
      "For cluster 2, 78 documents\n",
      "1) change (score: 0.0395)\n",
      "2) time (score: 0.0319)\n",
      "3) check (score: 0.0311)\n",
      "4) day (score: 0.0233)\n",
      "5) shift (score: 0.0224)\n",
      "Average cluster score 0.0296\n",
      "\n",
      "For cluster 3, 75 documents\n",
      "1) old (score: 0.0365)\n",
      "2) words (score: 0.0306)\n",
      "3) slow (score: 0.0287)\n",
      "4) sweet (score: 0.0249)\n",
      "5) past (score: 0.0238)\n",
      "Average cluster score 0.0289\n",
      "\n",
      "For cluster 4, 74 documents\n",
      "1) don (score: 0.0342)\n",
      "2) head (score: 0.0299)\n",
      "3) try (score: 0.0257)\n",
      "4) choose (score: 0.0246)\n",
      "5) cow (score: 0.0241)\n",
      "Average cluster score 0.0277\n",
      "\n",
      "For cluster 5, 63 documents\n",
      "1) cash (score: 0.0353)\n",
      "2) buy (score: 0.0289)\n",
      "3) shout (score: 0.0284)\n",
      "4) need (score: 0.0250)\n",
      "5) try (score: 0.0244)\n",
      "Average cluster score 0.0284\n",
      "\n",
      "The clusters by average score, descending\n",
      "2 \t0.0296391769374\n",
      "3 \t0.028897037471\n",
      "5 \t0.028405571299\n",
      "4 \t0.027698306971\n",
      "1 \t0.027279910304\n",
      "0 \t0.0238183676903\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEZCAYAAAB1mUk3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDMzMDCcDMzPDycDMzHAyMDMz4P8DgUUe+1MKXacAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa79168d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 76 documents\n",
      "1) head (score: 0.0402)\n",
      "2) hide (score: 0.0318)\n",
      "3) face (score: 0.0294)\n",
      "4) turn (score: 0.0288)\n",
      "5) smile (score: 0.0278)\n",
      "Average cluster score 0.0316\n",
      "\n",
      "For cluster 1, 72 documents\n",
      "1) time (score: 0.0336)\n",
      "2) ends (score: 0.0298)\n",
      "3) shout (score: 0.0266)\n",
      "4) dance (score: 0.0265)\n",
      "5) old (score: 0.0224)\n",
      "Average cluster score 0.0278\n",
      "\n",
      "For cluster 2, 51 documents\n",
      "1) buy (score: 0.0339)\n",
      "2) prize (score: 0.0300)\n",
      "3) wish (score: 0.0296)\n",
      "4) boast (score: 0.0251)\n",
      "5) win (score: 0.0221)\n",
      "Average cluster score 0.0281\n",
      "\n",
      "For cluster 3, 96 documents\n",
      "1) wait (score: 0.0335)\n",
      "2) end (score: 0.0252)\n",
      "3) near (score: 0.0245)\n",
      "4) free (score: 0.0237)\n",
      "5) game (score: 0.0227)\n",
      "Average cluster score 0.0259\n",
      "\n",
      "For cluster 4, 66 documents\n",
      "1) good (score: 0.0416)\n",
      "2) old (score: 0.0404)\n",
      "3) quick (score: 0.0386)\n",
      "4) check (score: 0.0306)\n",
      "5) new (score: 0.0245)\n",
      "Average cluster score 0.0352\n",
      "\n",
      "For cluster 5, 49 documents\n",
      "1) stand (score: 0.0401)\n",
      "2) loud (score: 0.0360)\n",
      "3) words (score: 0.0303)\n",
      "4) beat (score: 0.0239)\n",
      "5) strain (score: 0.0225)\n",
      "Average cluster score 0.0306\n",
      "\n",
      "For cluster 6, 67 documents\n",
      "1) voice (score: 0.0382)\n",
      "2) don (score: 0.0314)\n",
      "3) right (score: 0.0310)\n",
      "4) choice (score: 0.0295)\n",
      "5) choose (score: 0.0249)\n",
      "Average cluster score 0.0310\n",
      "\n",
      "The clusters by average score, descending\n",
      "4 \t0.0351580320794\n",
      "0 \t0.0315965797936\n",
      "6 \t0.0310185904172\n",
      "5 \t0.0305584505484\n",
      "2 \t0.028135666825\n",
      "1 \t0.027786906108\n",
      "3 \t0.0259280843759\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEiCAYAAAD+lrlwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZZz4zcy6jBO/mVmXceI3M+sy/x86AusgFLFyIAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa76130d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 70 documents\n",
      "1) change (score: 0.0376)\n",
      "2) man (score: 0.0290)\n",
      "3) time (score: 0.0285)\n",
      "4) check (score: 0.0262)\n",
      "5) day (score: 0.0245)\n",
      "Average cluster score 0.0291\n",
      "\n",
      "For cluster 1, 51 documents\n",
      "1) game (score: 0.0420)\n",
      "2) play (score: 0.0390)\n",
      "3) best (score: 0.0358)\n",
      "4) gals (score: 0.0353)\n",
      "5) match (score: 0.0301)\n",
      "Average cluster score 0.0364\n",
      "\n",
      "For cluster 2, 58 documents\n",
      "1) gone (score: 0.0335)\n",
      "2) loud (score: 0.0277)\n",
      "3) noise (score: 0.0254)\n",
      "4) stay (score: 0.0242)\n",
      "5) flee (score: 0.0234)\n",
      "Average cluster score 0.0268\n",
      "\n",
      "For cluster 3, 65 documents\n",
      "1) years (score: 0.0393)\n",
      "2) age (score: 0.0311)\n",
      "3) days (score: 0.0297)\n",
      "4) slow (score: 0.0277)\n",
      "5) wait (score: 0.0258)\n",
      "Average cluster score 0.0307\n",
      "\n",
      "For cluster 4, 63 documents\n",
      "1) choice (score: 0.0336)\n",
      "2) voice (score: 0.0320)\n",
      "3) old (score: 0.0316)\n",
      "4) kids (score: 0.0257)\n",
      "5) ll (score: 0.0246)\n",
      "Average cluster score 0.0295\n",
      "\n",
      "For cluster 5, 62 documents\n",
      "1) try (score: 0.0366)\n",
      "2) pride (score: 0.0325)\n",
      "3) cool (score: 0.0320)\n",
      "4) shout (score: 0.0310)\n",
      "5) tongue (score: 0.0276)\n",
      "Average cluster score 0.0320\n",
      "\n",
      "For cluster 6, 43 documents\n",
      "1) cow (score: 0.0374)\n",
      "2) meal (score: 0.0326)\n",
      "3) deal (score: 0.0320)\n",
      "4) beef (score: 0.0304)\n",
      "5) tow (score: 0.0299)\n",
      "Average cluster score 0.0324\n",
      "\n",
      "For cluster 7, 65 documents\n",
      "1) stop (score: 0.0257)\n",
      "2) rules (score: 0.0246)\n",
      "3) bad (score: 0.0244)\n",
      "4) old (score: 0.0217)\n",
      "5) try (score: 0.0216)\n",
      "Average cluster score 0.0236\n",
      "\n",
      "The clusters by average score, descending\n",
      "1 \t0.0364224218074\n",
      "6 \t0.0324448349013\n",
      "5 \t0.0319660527717\n",
      "3 \t0.0307345499124\n",
      "4 \t0.0294860938102\n",
      "0 \t0.029149522268\n",
      "2 \t0.0268447382772\n",
      "7 \t0.0235939506804\n",
      "----------------------------------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEfCAYAAABGcq0DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefe7a6a0d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "For cluster 0, 67 documents\n",
      "1) hide (score: 0.0359)\n",
      "2) pride (score: 0.0314)\n",
      "3) cool (score: 0.0311)\n",
      "4) words (score: 0.0289)\n",
      "5) head (score: 0.0259)\n",
      "Average cluster score 0.0306\n",
      "\n",
      "For cluster 1, 58 documents\n",
      "1) dance (score: 0.0327)\n",
      "2) step (score: 0.0260)\n",
      "3) watch (score: 0.0244)\n",
      "4) craze (score: 0.0241)\n",
      "5) feet (score: 0.0239)\n",
      "Average cluster score 0.0262\n",
      "\n",
      "For cluster 2, 57 documents\n",
      "1) years (score: 0.0419)\n",
      "2) age (score: 0.0404)\n",
      "3) youth (score: 0.0363)\n",
      "4) clocks (score: 0.0300)\n",
      "5) young (score: 0.0295)\n",
      "Average cluster score 0.0356\n",
      "\n",
      "For cluster 3, 44 documents\n",
      "1) checks (score: 0.0365)\n",
      "2) seek (score: 0.0363)\n",
      "3) game (score: 0.0331)\n",
      "4) end (score: 0.0311)\n",
      "5) fools (score: 0.0281)\n",
      "Average cluster score 0.0330\n",
      "\n",
      "For cluster 4, 47 documents\n",
      "1) work (score: 0.0386)\n",
      "2) fat (score: 0.0351)\n",
      "3) waist (score: 0.0313)\n",
      "4) just (score: 0.0311)\n",
      "5) new (score: 0.0306)\n",
      "Average cluster score 0.0333\n",
      "\n",
      "For cluster 5, 39 documents\n",
      "1) near (score: 0.0386)\n",
      "2) stop (score: 0.0357)\n",
      "3) blame (score: 0.0355)\n",
      "4) red (score: 0.0294)\n",
      "5) curse (score: 0.0281)\n",
      "Average cluster score 0.0335\n",
      "\n",
      "For cluster 6, 69 documents\n",
      "1) won (score: 0.0446)\n",
      "2) run (score: 0.0362)\n",
      "3) man (score: 0.0300)\n",
      "4) win (score: 0.0296)\n",
      "5) right (score: 0.0261)\n",
      "Average cluster score 0.0333\n",
      "\n",
      "For cluster 7, 40 documents\n",
      "1) eggs (score: 0.0370)\n",
      "2) end (score: 0.0352)\n",
      "3) fowl (score: 0.0322)\n",
      "4) leave (score: 0.0294)\n",
      "5) need (score: 0.0288)\n",
      "Average cluster score 0.0325\n",
      "\n",
      "For cluster 8, 56 documents\n",
      "1) ends (score: 0.0395)\n",
      "2) friends (score: 0.0327)\n",
      "3) fun (score: 0.0323)\n",
      "4) play (score: 0.0306)\n",
      "5) meet (score: 0.0303)\n",
      "Average cluster score 0.0331\n",
      "\n",
      "The clusters by average score, descending\n",
      "2 \t0.0355985835687\n",
      "5 \t0.0334685632264\n",
      "4 \t0.0333101913555\n",
      "6 \t0.0332775904755\n",
      "8 \t0.0330722575894\n",
      "3 \t0.0330258676369\n",
      "7 \t0.0325220735737\n",
      "0 \t0.0306362510957\n",
      "1 \t0.0262350716323\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "top_four = np.where(y_avg>np.percentile(y_avg,60))[0]+2\n",
    "\n",
    "for n_clusters in top_four:\n",
    "\n",
    "    pipe = Pipeline( [('vec',    TfidfVectorizer(**tfidf_kwargs)),\n",
    "                      ('kmeans', KMeans(n_clusters=n_clusters, **kmeans_kwargs))])\n",
    "\n",
    "    pipe.fit(documents)\n",
    "    labels = pipe.predict(documents)\n",
    "    \n",
    "    # how many documents in each cluster?\n",
    "    # ===================================\n",
    "    from collections import Counter\n",
    "    c = Counter(labels)\n",
    "\n",
    "    plt.bar(c.keys(),c.values())\n",
    "\n",
    "    plt.title('Distribution of Philadelphia Reflections haiku clusters')\n",
    "    plt.xlabel('Cluster Number')\n",
    "    plt.ylabel('Frequency of Haiku in each Cluster')\n",
    "\n",
    "    for k,v in c.iteritems():\n",
    "        plt.text(k+0.25,v+5,v)\n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "    # get the term list extracted by Tfidfvectorizer\n",
    "    # ==============================================\n",
    "    terms = pipe.named_steps['vec'].get_feature_names()\n",
    "    \n",
    "    # find the five most-important terms in each cluster\n",
    "    # ==================================================\n",
    "    cluster_avg5_scores = []\n",
    "    for cluster_number in range(n_clusters):\n",
    "        print(\"\\nFor cluster {}, {:,} documents\".format(cluster_number, c[cluster_number]))\n",
    "        centroid = pipe.named_steps['kmeans'].cluster_centers_[cluster_number]\n",
    "        most_important = centroid.argsort()\n",
    "        total_score = 0\n",
    "        for i in range(5):\n",
    "            term_index = most_important[-(i+1)]\n",
    "            total_score += centroid[term_index]\n",
    "            print(\"{}) {} (score: {:.4f})\".\\\n",
    "                  format(i+1, terms[term_index].encode('utf-8'), centroid[term_index]))\n",
    "        print(\"Average cluster score {:.4f}\".format(total_score/5.0))\n",
    "        cluster_avg5_scores.append((cluster_number, total_score/5.0))\n",
    "        \n",
    "    print(\"\\nThe clusters by average score, descending\")\n",
    "    import operator\n",
    "    cluster_avg5_scores.sort(key = operator.itemgetter(1), reverse=True)\n",
    "    for cnum, avg in cluster_avg5_scores:\n",
    "        print cnum, \"\\t\", avg\n",
    "    print('----------------------------------------------------------------')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}