{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Email Analysis: K-Means Clustering" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Clustering is an 'unsupervised learning' algorithm that has many diverse applications from computer graphics to economics. In Natural Language Processing, clustering has been used to determine authorship of the federalist papers, and analyze hundrends of thousands of documents as they are released in an instant. Clustering is particularly useful when you want to get a snapshot of a large dataset you have at hand." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Pre-analysis Houskeeping - Imports and Defining Methods" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "True" ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" } ], "source": [ "%matplotlib inline\n", "\n", "# Basic Imports\n", "import re, os, logging, sys, random, json, xml.etree.ElementTree \n", "from time import time\n", "from operator import itemgetter\n", "from datetime import date\n", "from datetime import timedelta\n", "from dateutil.parser import parse\n", "from pprint import pprint\n", "\n", "# Sci-kit learn Machine Learning Tookit Imports\n", "from sklearn.cluster import KMeans\n", "from sklearn.feature_extraction.text import TfidfVectorizer\n", "from sklearn.decomposition import PCA\n", "from sklearn.decomposition import TruncatedSVD\n", "from sklearn.pipeline import make_pipeline\n", "from sklearn.preprocessing import Normalizer\n", "from sklearn.model_selection import cross_val_score\n", "from sklearn import metrics\n", "from sklearn.metrics import pairwise_distances\n", "\n", "# Plotting Imports\n", "import numpy as np\n", "from scipy.stats import mode\n", "import matplotlib.pyplot as plt, mpld3\n", "from mpld3 import plugins\n", "from click_info import ClickInfo\n", "# mpld3.enable_notebook()\n", "\n", "# Data Object Related Imports\n", "from peewee import *\n", "from Email import Email\n", "\n", "def make_histogram(innerDict):\n", " x = np.arange(len(innerDict.keys()))\n", " y = innerDict.values()\n", " fig = plt.figure(figsize=(10,10))\n", " ax = fig.add_subplot(1,1,1)\n", " ax.bar(x, y)\n", " ax.set_xticks(x)\n", " ax.set_xticklabels(innerDict.keys(), rotation=70)\n", " plt.show()\n", "\n", "db = SqliteDatabase('emails.db')\n", "db.connect()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Perpare the tools we will use" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "N_CLUSTERS = 3\n", "reduce_dimensionality = True\n", "color_by_party = True\n", "k_means = KMeans(n_clusters=N_CLUSTERS, init='k-means++', max_iter=100, n_init=1, verbose=True)\n", "vectorizer = TfidfVectorizer(max_df=0.5, min_df=0.1, stop_words='english')\n", "lsa = TruncatedSVD(2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Perpare the Data" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "emails = np.array([email for email in Email.select()])\n", "text = np.array([email.text for email in emails])\n", "parties = np.array([email.sender.party if email and email.sender else \"\" for email in emails])\n", "vectors = vectorizer.fit_transform(text)\n", "if reduce_dimensionality == True:\n", " X = lsa.fit_transform(vectors)\n", "else:\n", " X = vectors" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Run K-means Clustering" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Initialization complete\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 0, inertia 28.34377379162939\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 1, inertia 25.984420059959188\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 2, inertia 25.076546105657336\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 3, inertia 23.201610785608267\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 4, inertia 21.628345173350453\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 5, inertia 21.159105434683763\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 6, inertia 21.023185724950125\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 7, inertia 21.0058769997022\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 8, inertia 21.003138810900573\n", "start iteration\n", "done sorting\n", "end inner loop\n", "Iteration 9, inertia 21.00191439314213\n", "center shift 1.762941e-03 within tolerance 3.144098e-06\n" ] } ], "source": [ "km = k_means.fit(X)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "k_means_labels = k_means.labels_\n", "k_means_cluster_centers = k_means.cluster_centers_\n", "k_means_labels_unique = np.unique(k_means_labels)\n", "terms = vectorizer.get_feature_names()" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "if reduce_dimensionality == False:\n", " order_centroids = k_means_cluster_centers.argsort()[:, ::-1]\n", " for i in range(N_CLUSTERS):\n", " print([(terms[ind], k_means_cluster_centers[i][ind]) for ind in order_centroids[i, :100]])\n", " print(\"\")" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[('senator', 81.68354778383616), ('vote', 67.61626487534294), ('message', 60.12901329847169), ('obama', 57.52809911702628), ('today', 55.83791372067094), ('political', 49.889051497995766), ('mitch', 48.85739672414709), ('support', 47.44594993948533), ('box', 45.59186927555744), ('washington', 44.043405060418955), ('future', 42.80345209449893), ('government', 42.32095331527883), ('state', 42.23593442154838), ('make', 41.58965449221981), ('new', 41.052881294920695), ('election', 40.995523166029756), ('united', 39.321101734031856), ('people', 39.06248820824899), ('states', 38.53066302841452), ('receive', 38.286905639561596), ('day', 37.107030805670135), ('just', 36.85562724607295), ('know', 36.31792073996653), ('emails', 34.91940465559045), ('team', 33.99959124888822), ('mcconnell', 33.89535026093225), ('www', 33.24288690891687), ('like', 32.54625077754243), ('time', 32.061142148291665), ('case', 30.967481078321626), ('join', 30.96195386614523), ('receiving', 30.361111047665602), ('country', 30.123240182423494), ('committee', 30.03084568761094), ('stand', 29.66541660348948), ('voters', 29.343287156184083), ('makes', 28.97461490766093), ('friend', 28.882935638134477), ('link', 28.795438426556892), ('right', 28.615736625970236), ('years', 28.56775053233214), ('online', 28.42929753055851), ('families', 28.253511809808117), ('general', 28.20783590373403), ('po', 27.55082537901218), ('sure', 27.183256431820535), ('stop', 27.173182884505998), ('thank', 25.963118944914388), ('work', 25.827251033537358), ('days', 25.362714473394142), ('believe', 24.97687152489734), ('republican', 24.962657394668238), ('let', 24.425145492093872), ('november', 23.181234328427955), ('list', 22.996501333312466), ('want', 22.923034575474805), ('race', 22.506183990460052), ('important', 22.363651145681292), ('forward', 21.974949566633587), ('friends', 21.823330042939887), ('working', 21.639941369068882), ('thanks', 21.510137154506953), ('win', 21.111414030529254), ('way', 20.542174105957095), ('contribution', 20.476620018523203), ('ad', 20.39047574336651), ('send', 20.259128751926863), ('week', 20.239813281238078), ('final', 19.762188111595812), ('record', 19.557350498453065), ('big', 19.498441620191322), ('diaemailid', 19.317128085663242), ('thread', 19.317128085663242), ('going', 19.279383209403086), ('fight', 19.138730559004273), ('voting', 19.100401863194005), ('making', 19.09393600479615), ('agenda', 19.05744677118215), ('standing', 18.81632427540422), ('tomorrow', 18.305840029495833), ('year', 18.289876103138205), ('strong', 18.15884118921512), ('news', 17.728850229672233), ('polls', 17.462647548603915), ('manager', 17.4475851601016), ('2014', 16.956334916408245), ('supporters', 16.83906598226089), ('contribute', 16.645596709050324), ('best', 16.35465288255357), ('majority', 15.713154431317239), ('left', 15.235405283084313), ('got', 15.22913184491973), ('momentum', 14.959971749013866), ('money', 14.932020951959293), ('tonight', 14.834256677169758), ('october', 14.632339934830133), ('don', 14.208454385160097), ('weeks', 14.068151492206454), ('ll', 14.061780274324388), ('running', 13.744268901703856)]\n", "\n", "[('express', 177.24459019441468), ('actblue', 67.99374401126678), ('immediately', 64.66050789225854), ('donation', 62.45493276105552), ('payment', 62.20157634730501), ('saved', 61.78756104592887), ('25', 60.26997635311869), ('_if', 58.2420832759923), ('mark', 57.561093917323646), ('goal', 52.93683682777798), ('contribute', 47.79439042468887), ('race', 47.645799100835205), ('10', 47.46560940354138), ('koch', 47.16598324245259), ('right', 46.87594034556449), ('win', 46.63316250236415), ('50', 46.08728736857387), ('money', 45.20071679953026), ('information', 44.453784873792515), ('don', 42.00702247678514), ('election', 40.708245091365015), ('day', 40.31196563776257), ('democratic', 40.3078406463049), ('000', 40.126894088799475), ('friends', 39.57001502408208), ('rove', 38.827960527163924), ('make', 38.3148276168308), ('midnight', 37.88349859656095), ('majority', 37.396715767301785), ('days', 37.114413961143384), ('friend', 36.660872829510716), ('brothers', 36.128379883399546), ('info', 34.68020257340684), ('lose', 34.31858927288748), ('ll', 33.327609375110626), ('contributions', 32.49816306597632), ('polls', 32.28388358065796), ('just', 31.97740344090822), ('gifts', 31.335258615418777), ('deadline', 31.01306726050012), ('fight', 30.900260995960167), ('million', 30.876601165806242), ('tax', 30.863607607943997), ('fec', 30.827574045585955), ('going', 30.554177825435854), ('deductible', 30.553968017525335), ('field', 30.512592686717426), ('thanks', 29.168970738120425), ('republicans', 29.123228100589905), ('dear', 28.87627638964626), ('new', 28.72782372482291), ('team', 28.727189545167516), ('diaemailid', 28.660055412142608), ('thread', 28.660055412142608), ('know', 28.501910941944608), ('list', 28.416030587122524), ('voters', 28.08463125140844), ('special', 27.47861178964736), ('spending', 26.93999798741874), ('grassroots', 26.735261487964138), ('democrats', 26.230643667219887), ('thank', 24.845449007463046), ('time', 24.707983971142788), ('hit', 24.62163577513254), ('support', 24.473119487375314), ('close', 24.17952363720731), ('final', 23.870267196100535), ('ads', 23.770044854421762), ('reach', 23.652888060211417), ('contribution', 23.234204075653057), ('raise', 22.594281233988255), ('away', 22.452055671376044), ('let', 21.452170771442063), ('tonight', 21.099528114753227), ('got', 20.147765565973124), ('vote', 19.818215380126162), ('today', 19.755605633483356), ('http', 19.507274461496223), ('buy', 19.390856606413656), ('big', 19.287228274861587), ('tomorrow', 19.083518350473007), ('hours', 19.061134336235895), ('left', 18.736524900755636), ('week', 18.686213025655), ('attacks', 18.663193016493544), ('fund', 18.597131300018958), ('like', 18.232331405044285), ('november', 18.190898347719106), ('committee', 17.907359628149443), ('interests', 17.88807350462523), ('people', 17.5078846750848), ('attack', 17.27804635610164), ('sure', 17.003123251516143), ('stop', 16.55099752341225), ('way', 16.535007434577054), ('work', 16.46599341457255), ('come', 16.463260947778153), ('running', 15.95816987145392), ('2014', 15.81031053743898), ('won', 15.580512019424111)]\n", "\n", "[('mark', 48.367412949737286), ('friends', 40.04805949550476), ('election', 33.97346899555415), ('make', 33.51095673420504), ('contributions', 33.16763323391554), ('just', 31.562825961787684), ('support', 31.337267421792188), ('right', 30.87169647074232), ('today', 30.517221865479833), ('race', 28.828813000624795), ('contribute', 27.849053436146907), ('win', 27.41700285727893), ('day', 27.127339916991996), ('like', 27.059046364283873), ('know', 26.713868848523205), ('deadline', 26.337383382955757), ('friend', 26.190034642808364), ('vote', 25.227628562017138), ('thank', 25.087250434442616), ('going', 24.53204120585876), ('goal', 23.978501639873066), ('voters', 22.880007063582482), ('new', 22.826970767179574), ('list', 22.796657925012305), ('team', 22.37119702746338), ('time', 22.192796625408096), ('days', 21.750314781942464), ('ad', 21.733489585652535), ('thanks', 21.696160056923624), ('2014', 21.083688827921613), ('ads', 20.559560935740755), ('let', 20.219533652691972), ('ll', 20.12855530650213), ('tax', 19.910610519941475), ('sure', 19.494179653445123), ('tonight', 19.457500037509448), ('join', 19.147428331695377), ('deductible', 18.9461283739399), ('final', 18.931800349659216), ('box', 18.258957114157923), ('polls', 18.188967854523405), ('000', 17.853123443345318), ('10', 17.831196496702418), ('rights', 17.58488530013367), ('big', 17.481660919198223), ('midnight', 17.394042953919964), ('diaemailid', 17.184944011218562), ('thread', 17.184944011218562), ('reserved', 16.722208041910328), ('hours', 16.69659443398428), ('policy', 16.42715958198113), ('people', 16.40294678828414), ('close', 16.36697418076768), ('content', 16.205254142105566), ('best', 16.147141328048267), ('working', 16.074992991872605), ('senator', 15.96148551486747), ('important', 15.915630560742208), ('gifts', 15.705887550160035), ('strong', 15.53944938358869), ('don', 15.176733945941272), ('po', 15.12055718954509), ('privacy', 15.116092944816886), ('want', 15.10451874013279), ('money', 14.917741702546607), ('country', 14.838220550833046), ('fight', 14.74555605003162), ('grassroots', 14.68549670748008), ('reach', 14.666449506082703), ('way', 14.4866714665071), ('political', 14.273222514080256), ('work', 14.231673896069092), ('stand', 13.865488422860999), ('supporters', 13.830104095391059), ('away', 13.765298081800484), ('contribution', 13.658118790607395), ('send', 13.650099744198863), ('week', 13.547446669477804), ('hard', 13.532178660430953), ('left', 13.48863642331085), ('50', 13.21163768385512), ('message', 13.047826199940348), ('momentum', 13.037631399460514), ('washington', 12.880875885159655), ('http', 12.765409462803884), ('special', 12.538607319522455), ('news', 12.527471815528827), ('critical', 12.469594439580476), ('state', 12.463056087488997), ('victory', 12.173095799490483), ('tomorrow', 12.160773771299343), ('committee', 12.052313741460289), ('families', 12.02928824534972), ('donation', 12.010668971384025), ('got', 11.945237244281532), ('raise', 11.923879832497796), ('fund', 11.876911546282354), ('record', 11.842699715759325), ('field', 11.74785636064783), ('standing', 11.721358137144357)]\n", "\n" ] } ], "source": [ "for k in range(N_CLUSTERS):\n", " z = vectors.toarray()[k_means_labels == k]\n", " wordz_tfidf = [(terms[i], z[:,i].sum()) for i in range(z.shape[1])]\n", " wordz_tfidf = sorted(wordz_tfidf, key=lambda x: x[1], reverse=True )\n", " print(wordz_tfidf[:100])\n", " print(\"\")" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/html": [ "\n", "\n", "\n", "\n", "
\n", "" ], "text/plain": [ "