{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When it comes to consuming real-time content, Twitter is the place to be; Be it sending out 'Tweets' real-time, or discovering latest online 'Trends' anywhere, or ability to begin a 'conversation' with anyone, Twitter does it all. In fact, Twitter Management wrote this in their [letter to shareholders][1] last year. \n",
    "[1]: http://files.shareholder.com/downloads/AMDA-2F526X/106725176x0x874459/8A4D1A1D-D184-4AFE-9AC1-F880C5EA06F1/Q415_Shareholder_Letter.pdf\n",
    "\n",
    "> _We’re focused now on what Twitter does best: live. Twitter is live: live commentary, live connections, live conversations. Whether it’s breaking news, entertainment, sports, or everyday topics, hearing about and watching a live event unfold is the fastest way to understand the power of Twitter._ \n",
    "\n",
    "> _Twitter has always been considered a “second screen” for what’s happening in the world and we believe we can become the first screen for everything that’s happening now. And by doing so, we believe we can build the planet’s largest daily connected audience. A connected audience is one that watches together, and can talk with one another in real-time. It’s what Twitter has provided for close to 10 years, and it’s what we will continue to drive in the future_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Embedded in a Twitter User's Social-graph* is a wealth of information on User's likes and interests. Unlike Facebook or LinkedIn, the magic of Twitter is in its 'Follow' structure - where any can follow any without they knowing each other. This directed social-graph, when methodologically summarized, can reveal interesting information on the most influential/central friends in the network and also help personalize/enrich one's Twitter experience by unearthing implicit-clusters in the network.\n",
    "\n",
    "_*Social-graph: [User, 1st degree Connections, and the links between]_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook, we'll look into these: \n",
    "1. Extract the ego-network of 'self' using Twitter API \n",
    "+ Identify the influential nodes using 'Page Rank' formulation \n",
    "+ Identify implicit clusters formed\n",
    "+ Recommend new friends to follow on the basis of cluster of interest "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: This study is limited to  Ego network: the focal node (\"ego\": here the self-node) and the nodes to whom ego is directly connected to (\"alters\") plus the ties, if any, among the alters. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Extract social-graph strucutre using Twitter API "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from collections import Counter, defaultdict\n",
    "from datetime import datetime\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "import csv\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os.path\n",
    "import pandas as pd\n",
    "import re\n",
    "import seaborn as sns;  sns.set()\n",
    "import time\n",
    "import twitter\n",
    "\n",
    "% matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (10,7)\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "#print twitter.__path__  \n",
    "\n",
    "import random\n",
    "random.seed(1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>body {\n",
       "    margin: 0;\n",
       "    font-family: Helvetica;\n",
       "}\n",
       "table.dataframe {\n",
       "    border-collapse: collapse;\n",
       "    border: none;\n",
       "}\n",
       "table.dataframe tr {\n",
       "    border: none;\n",
       "}\n",
       "table.dataframe td, table.dataframe th {\n",
       "    margin: 0;\n",
       "    border: 1px solid white;\n",
       "    padding-left: 0.25em;\n",
       "    padding-right: 0.25em;\n",
       "}\n",
       "table.dataframe th:not(:empty) {\n",
       "    background-color: #fec;\n",
       "    text-align: left;\n",
       "    font-weight: normal;\n",
       "}\n",
       "table.dataframe tr:nth-child(2) th:empty {\n",
       "    border-left: none;\n",
       "    border-right: 1px dashed #888;\n",
       "}\n",
       "table.dataframe td {\n",
       "    border: 2px solid #ccf;\n",
       "    background-color: #f4f4ff;\n",
       "}\n",
       "h3 {\n",
       "    color: white;\n",
       "    background-color: black;\n",
       "    padding: 0.5em;\n",
       "}\n",
       "</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Pandas data-frame print formatting\n",
    "from IPython.core.display import HTML\n",
    "css = open('style-table.css').read() + open('style-notebook.css').read()\n",
    "HTML('<style>{}</style>'.format(css))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "self_screen_name = 'bala_io'            # Self\n",
    "\n",
    "# Keep appending data\n",
    "fof_filename     = \"edges.csv\"          # 'Alters' and their Source->Sink Edges. \n",
    "cache_filename   = \"cache.csv\"          # local cache of (TwitterId, FullName, UserName)\n",
    "\n",
    "# One-time-use files\n",
    "binaryMap_filename = \"binaryMap.csv\"   # Directed Graph. Adjacencies as 0/1.RowFollowCol\n",
    "cluster_filename   = \"results.csv\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<twitter.api.Api at 0x7f1c2f6bb898>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Twitter auth. https://dev.twitter.com/oauth/overview/application-owner-access-tokens\n",
    "\n",
    "with open(\"../../passwd/credentials.txt\", \"r\") as f:\n",
    "    reader = csv.reader(f )\n",
    "    login_dict = {line[0]: line[1]                       \n",
    "                    for line in reader}        \n",
    "\n",
    "api = twitter.Api(consumer_key=login_dict.get('consumer_key') ,\n",
    "                  consumer_secret=login_dict.get('consumer_secret'),\n",
    "                  access_token_key=login_dict.get('access_token_key'),\n",
    "                  access_token_secret=login_dict.get('access_token_secret'))\n",
    "api"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 'Self' and Friends of Self\n",
    "\n",
    "self_node    =  api.GetUser(screen_name = self_screen_name)    \n",
    "self_node_id =  str(self_node.id)                      # Twitter Id of Self\n",
    "\n",
    "friends_of_self = api.GetFriendIDs(user_id = self_node_id, \n",
    "                                   screen_name = self_screen_name , \n",
    "                                   stringify_ids = True)\n",
    "index = [self_node_id] + friends_of_self"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# GetFriendIDs() API call is rate-limited at 15 req / 15 min \n",
    "# https://dev.twitter.com/rest/public/rate-limiting\n",
    "\n",
    "# For each of the list of nodes, fetch the list of nodes it follows, append to file\n",
    "\n",
    "def update_FoF_File(fileName, to_fetch_list):\n",
    "    \n",
    "    with open(fileName, 'a') as f:\n",
    "        apiReqCount = 0\n",
    "        \n",
    "        for node in to_fetch_list:\n",
    "            friends_of_node = api.GetFriendIDs(user_id = node,  stringify_ids = True) \n",
    "            row = ','.join([str(i) for i in [node] +  friends_of_node ]) + \"\\n\"\n",
    "            f.write(row) \n",
    "            \n",
    "            apiReqCount += 1\n",
    "            \n",
    "            if (apiReqCount == 15): \n",
    "                apiReqCount = 0\n",
    "                \n",
    "                print(\"Off to Sleep :)\")\n",
    "                time.sleep(15*60 + 10)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# parse FoF file and return list of nodes, for whom source->sink Edges are already there.\n",
    "\n",
    "def getFinishList(fileName):\n",
    "        \n",
    "    if not os.path.isfile(fileName):\n",
    "        return [] \n",
    "    with open(fileName, 'r') as f:\n",
    "        return [ row.strip().split(',')[0] for row in f ]  # 1st entry is a user"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Ego-network as adjacency-matrix\n",
    "# Parses FoF file in order of index, create list of adjacencies as 0 | 1\n",
    "# Writes to File. Adjacency-matrix in Row_follows_Column format\n",
    "\n",
    "def updateBinaryMapFile(fof_filename, binaryMap_filename, index):\n",
    "    \n",
    "    with open(fof_filename, \"r\") as f:\n",
    "        stripped_f = (line.replace('\\r', '') for line in f)\n",
    "        reader = csv.reader(stripped_f)\n",
    "        fof_dict = {line[0]: line[1:] for line in reader \n",
    "                    if line[0] in index}   # dict of node:his_followers\n",
    "        if self_node_id not in fof_dict:\n",
    "            fof_dict[self_node_id] = index[1:]             # dict of Self\n",
    "    \n",
    "    bool_list = []\n",
    "    \n",
    "    for user in index:\n",
    "        user_friends = set( fof_dict[user] )  \n",
    "        bool_row = [item in user_friends for item in index]  # for each, fill T/F \n",
    "        bool_list.append(bool_row)\n",
    "    \n",
    "    int_nparray = np.array(bool_list) + 0                    # Bool to int\n",
    "\n",
    "    binaryMap_rfc = pd.DataFrame(data = int_nparray, columns= index, index = index)\n",
    "    binaryMap_rfc.to_csv(binaryMap_filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# For list of Ids, fetch Profile details. If not in Offline file, make an API call\n",
    "# Returns ['UserName', 'FullName', 'Followers_count', 'Friends_count', 'Location', 'Created_at']\n",
    "# UsersLookup API 100 Ids/request\n",
    "\n",
    "def lookup_in_cache(friendsIdsList):\n",
    "\n",
    "    cache, delta_cache = pd.DataFrame(), pd.DataFrame()\n",
    "    UserNameList, namesList     = [], []\n",
    "    followers_count, friends_count = [], []\n",
    "    location, created_at           = [], []\n",
    "    \n",
    "    if os.path.isfile(cache_filename):\n",
    "        cache = pd.read_csv(cache_filename, skip_blank_lines=True,\n",
    "                        dtype={'Ids':str, 'Friends_count':int, 'Followers_count':int})\n",
    "        cache.set_index('Ids', inplace=True)\n",
    "        to_fetch_list = list ( set (friendsIdsList) - set(cache.index) )\n",
    "    else :        \n",
    "        to_fetch_list = friendsIdsList   \n",
    "    \n",
    "    i = 0\n",
    "    while (i < len(to_fetch_list) * 1./100):   \n",
    "        print(\"... Cache-Miss for \" + str(len(to_fetch_list)) +  \" nodes. Updating Cache...\")\n",
    "        low, high = i * 100, min( len(to_fetch_list), (i+1)*100 )  # UsersLookup api\n",
    "        twitterObjectsList = api.UsersLookup(user_id = to_fetch_list[low:high])    \n",
    "        \n",
    "        temp = zip(*[( tempObject.screen_name,        #ScreenName\n",
    "                      tempObject.name,                #Name\n",
    "                      tempObject.followers_count,     #Followers\n",
    "                      tempObject.friends_count,       #Friends\n",
    "                      tempObject.location,            #Location\n",
    "                      tempObject.created_at           #CreatedAt\n",
    "                     ) for tempObject in twitterObjectsList])\n",
    "        temp = list(temp)\n",
    "        \n",
    "        UserNameList    += list(temp[0])\n",
    "        namesList       += list(temp[1])\n",
    "        followers_count += list(temp[2])\n",
    "        friends_count   += list(temp[3])\n",
    "        location        += list(temp[4])\n",
    "        created_at      += list(temp[5])\n",
    "        \n",
    "        i = i + 1       \n",
    "    \n",
    "    if len(to_fetch_list) > 0:\n",
    "        delta_cache  = pd.DataFrame({'UserName':UserNameList,\n",
    "                                     'FullName':namesList,\n",
    "                                     'Ids': to_fetch_list,\n",
    "                                     'Followers':followers_count,\n",
    "                                     'Friends': friends_count,\n",
    "                                     'Location':location,\n",
    "                                     'Created':created_at})\n",
    "        \n",
    "        delta_cache['Created'] = delta_cache['Created'].apply(lambda x: \n",
    "                                                datetime.strptime(\n",
    "                                                    re.sub(r\"\\+[0-9]* \", \"\",x),'%c').\n",
    "                                                strftime(\"%b-%Y\"))\n",
    "    \n",
    "        delta_cache.set_index('Ids', inplace=True, drop = True)\n",
    "        cache = cache.append(delta_cache)\n",
    "        cache.to_csv(cache_filename)        \n",
    "\n",
    "    return cache.loc[friendsIdsList]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Display cluster-wise most-influential users, for the given clustering algo\n",
    "\n",
    "def top_nodes_in_cluster(df, cluster_algo, n_clusters):\n",
    "    dummy_df = pd.DataFrame()\n",
    "\n",
    "    for i in range(n_clusters):\n",
    "        nodes_in_cluster = list( df [df[cluster_algo] == i ]['FullName'] )     \n",
    "        if len(nodes_in_cluster) >= 10:            # show only clusters of size > 10        \n",
    "            col_name           = str(i) + \" : \" + str(len(nodes_in_cluster)) + \" Ids\"\n",
    "            dummy_df[col_name] = nodes_in_cluster[:10]      \n",
    "    \n",
    "    return dummy_df   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# identify 20 friends to follow after aggregating friends followed by top 50% in list\n",
    "\n",
    "def discover_Friends_toFollow(ids_of_interest, friend_list, prop = .5, count = 20):    \n",
    "    \n",
    "    ids_of_interest  = ids_of_interest[:int(len(ids_of_interest) * prop)]\n",
    "    \n",
    "    if self_node_id in ids_of_interest:\n",
    "        ids_of_interest.remove(self_node_id)\n",
    "\n",
    "    print(\"'Who-To-Follow' reco after looking at %3d friends' friends:\" %(len(ids_of_interest)))\n",
    "    \n",
    "    with open(fof_filename) as f:\n",
    "        reader = csv.reader(f)\n",
    "        fof_dict = {row[0]:row[0:] for row in reader}  # dict of node:her_followers\n",
    "\n",
    "    friendsToFollow = []\n",
    "    for id in ids_of_interest:\n",
    "        friendsToFollow += list (set(fof_dict[str(id)])  - set(friend_list) ) \n",
    "        \n",
    "\n",
    "    friendsToFollow = Counter(friendsToFollow).most_common(count)    \n",
    "    tuples_list = list(zip(*friendsToFollow) )\n",
    "           \n",
    "    topFriendsToFollowDF = pd.DataFrame()\n",
    "    topFriendsToFollowDF['Ids'] = list(tuples_list[0])\n",
    "    topFriendsToFollowDF['Freq'] = list(tuples_list[1])\n",
    "    topFriendsToFollowDF.set_index('Ids', drop = True, inplace = True)              \n",
    "    index = topFriendsToFollowDF.index\n",
    "    \n",
    "    topFriendsToFollowDF = topFriendsToFollowDF.merge(lookup_in_cache(index), copy = False,\n",
    "                              left_index = True, right_index = True)\n",
    "    \n",
    "    return topFriendsToFollowDF    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 out of 376 Friends details to be fetched\n"
     ]
    }
   ],
   "source": [
    "# For the list of nodes I follow, fetch their friends-list\n",
    "\n",
    "fof_finish_list   = getFinishList(fof_filename )        # Completed nodes\n",
    "fof_to_fetch_list = list ( set(friends_of_self) - set(fof_finish_list) )  # Pending nodes\n",
    "\n",
    "print( str(len(fof_to_fetch_list)) + \" out of \" + str(len(index)  - 1) + \n",
    "      \" Friends details to be fetched\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# For the remaining nodes, populate their details in fof_file\n",
    "update_FoF_File(fof_filename, fof_to_fetch_list) \n",
    "\n",
    "# Build the adjacency matrix in terms of 0 and 1 (if there is an edge)\n",
    "updateBinaryMapFile(fof_filename, binaryMap_filename, index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(377, 377)\n"
     ]
    }
   ],
   "source": [
    "# Read adj-matrix into df. Cell MxN is 1 iff node in Mth row follows node in Nth column \n",
    "binaryMap_rfc = pd.read_csv(binaryMap_filename, skip_blank_lines=True, index_col = 0)\n",
    "print(binaryMap_rfc.shape)\n",
    "\n",
    "\n",
    "outlinks_count = binaryMap_rfc.sum(axis = 1)   # horizontal-sum to count outlinks\n",
    "inlinks_count = binaryMap_rfc.sum(axis = 0)   # vertical-sum to count inlinks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAGbCAYAAAALJa6vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAE6tJREFUeJzt3X+I5Pd93/GXeuv6otxCTsn2EGeXa4j4gGrqMzWuS0y5\nWHFQW1OpEA67NBzUTfuH09YQaBX3D5tCQP80jf9IS93E+NImsUQSVyI0Cc7VhxMorqtwOAnymxzp\nmeY46Y7Lhuwh5CJl+8eOvLeq9nbfuzu3szOPB4idnzufefNl76nvzHzngfX19QAAsHt/4bAXAABw\n1AgoAIAmAQUA0CSgAACaBBQAQNPS/XywW7fWpv6Rv5MnH8zq6ivTfpgjwzw2mcVW5rHJLLYyj01m\nsdWizWNlZfmB7a6buz1QS0vHDnsJM8U8NpnFVuaxySy2Mo9NZrGVeWyau4ACAJg2AQUA0CSgAACa\nBBQAQJOAAgBoElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElA\nAQA0LR32Ag7ab/yPa1m782r7fufOnj74xQAAc8keKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABN\nAgoAoElAAQA0CSgAgCYBBQDQtKuvchljXEuyluT1JK9V1XvHGA8leSbJmSTXkpyvqtWprBIAYIZ0\n9kD9QFWdrar3Ts4/leRSVT2S5NLkPADA3NvPS3hPJLk4OX0xyZP7Xw4AwOzb1Ut4SdaT/NYY4/Uk\n/7GqPpvkVFXdmFz/UpJTO/2SkycfzNLSsb2tdLeu3s7yiePtu62sLE9hMbNhnp9bl1lsZR6bzGIr\n89hkFluZx4bdBtQHqur6GOMvJfnSGOMbd19ZVetjjPWdfsnq6it7WWPb2p1X2/e5dWttCis5fCsr\ny3P73LrMYivz2GQWW5nHJrPYatHmca9Y3NVLeFV1ffLzZpIvJnlfkpfHGA8nyeTnzX2vFADgCNgx\noMYY3znGWH7jdJIfSvL7SZ5PcmFyswtJnpvWIgEAZsluXsI7leSLY4w3bv+LVfUbY4yvJXl2jPGx\nJN9Mcn56ywQAmB07BlRV/VGSd7/F5beTPDaNRQEAzDJHIgcAaNrtp/CYAZevXG/fZ/nE8fz17/vu\nKawGABaXPVAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmhzGYGIvhwhIknNnTx/wSgCAWWcPFABA\nk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBoElAAAE0CCgCgSUABADQJ\nKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgAgCYBBQDQJKAAAJoEFABAk4AC\nAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBoElAAAE0CCgCgSUABADQJKACA\nJgEFANAkoAAAmgQUAEDT0mEv4Ki7fOV6+z7nzp6ewkoAgPvFHigAgCYBBQDQJKAAAJoEFABAk4AC\nAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0LTrr3IZYxxL8r+SXK+qD48xHkryTJIzSa4lOV9Vq9NY\nJADALOnsgfoXSV686/xTSS5V1SNJLk3OAwDMvV0F1BjjHUn+bpKfveviJ5JcnJy+mOTJg10aAMBs\n2u1LeD+d5F8mWb7rslNVdWNy+qUkp3b6JSdPPpilpWO9FXZdvZ3lE8en+xj79MLV23u6316f18rK\n8s43WhBmsZV5bDKLrcxjk1lsZR4bdgyoMcaHk9ysqhfGGOfe6jZVtT7GWN/pd62uvtJf4R6s3Xn1\nvjzOUbB84nhu3Vo77GXMhJWVZbO4i3lsMoutzGOTWWy1aPO4Vyzu5iW870/y98YY15J8IckHxxj/\nJcnLY4yHk2Ty8+a+VwoAcATsGFBV9RNV9Y6qOpPkI0n+e1X9wyTPJ7kwudmFJM9NbZUAADNkP8eB\nejrJh8YYf5jkByfnAQDm3q6PA5UkVXU5yeXJ6dtJHjv4JQEAzDZHIgcAaBJQAABNAgoAoElAAQA0\nCSgAgCYBBQDQJKAAAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOA\nAgBoElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgA\ngCYBBQDQJKAAAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBo\nElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgAgCYB\nBQDQJKAAAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAICmpZ1uMMY4nuQrSd4+uf0vV9WnxhgP\nJXkmyZkk15Kcr6rV6S0VAGA27GYP1LeSfLCq3p3kbJLHxxjvT/JUkktV9UiSS5PzAABzb8c9UFW1\nnuTO5OzbJv+tJ3kiybnJ5ReTXE7yrw58hQAAM2bHgEqSMcaxJC8k+b4kP1NVXx1jnKqqG5ObvJTk\n1E6/5+TJB7O0dGzPi92Vq7ezfOL4dB/jiFlZWT7sJcwMs9jKPDaZxVbmsckstjKPDbsKqKp6PcnZ\nMcZ3JfniGONdb7p+fYyxvtPvWV19ZW+rbFq78+p9eZyjYPnE8dy6tXbYy5gJKyvLZnEX89hkFluZ\nxyaz2GrR5nGvWGx9Cq+q/jTJl5M8nuTlMcbDSTL5eXMfawQAODJ2DKgxxspkz1PGGN+R5ENJvpHk\n+SQXJje7kOS5aS0SAGCW7GYP1MNJvjzG+HqSryX5UlX9WpKnk3xojPGHSX5wch4AYO7t5lN4X0/y\nnre4/HaSx6axKACAWeZI5AAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBoElAAAE0C\nCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgAgCYBBQDQJKAA\nAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0LR02Atgdl2+cr19n3NnT09hJQAwW+yB\nAgBoElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgA\ngKalw14A03f5yvXDXgIAzBV7oAAAmgQUAECTgAIAaBJQAABNAgoAoMmn8DhQe/3E37mzpw94JQAw\nPfZAAQA0CSgAgCYBBQDQJKAAAJoEFABAk4ACAGhyGAOOtM5hE5ZPHM/anVcdMgGAfbMHCgCgSUAB\nADQJKACAph3fAzXGeGeSn09yKsl6ks9W1WfGGA8leSbJmSTXkpyvqtXpLRUAYDbsZg/Ua0l+vKoe\nTfL+JB8fYzya5Kkkl6rqkSSXJucBAObejgFVVTeq6ncnp9eSvJjkdJInklyc3OxikientUgAgFnS\nOozBGONMkvck+WqSU1V1Y3LVS9l4ie+eTp58MEtLx7pr7Ll6O8snjk/3MY6YozCPlZXlPd2v+9yW\nTxzPC1dv7+mxHv+bZ/Z0v1m217nPI7PYyjw2mcVW5rFh1wE1xjiR5FeSfKKq/myM8e3rqmp9jLG+\n0+9YXX1lT4vsWrvz6n15nKPgjWMfzbpbt9b2dL/Oc9vvLPa6xlm1srI8d89pr8xiK/PYZBZbLdo8\n7hWLu/oU3hjjbdmIp1+oql+dXPzyGOPhyfUPJ7m5z3UCABwJOwbUGOOBJD+X5MWq+qm7rno+yYXJ\n6QtJnjv45QEAzJ7dvIT3/Ul+JMnvjTGuTC77ZJKnkzw7xvhYkm8mOT+dJQIAzJYdA6qqfifJA9tc\n/djBLgcAYPY5EjkAQJOAAgBoElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQ\nAABNAgoAoElAAQA0CSgAgCYBBQDQJKAAAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA\n0CSgAACaBBQAQJOAAgBoElAAAE0CCgCgSUABADQtHfYCIEkuX7l+2EsAgF2zBwoAoElAAQA0CSgA\ngCYBBQDQJKAAAJoEFABAk4ACAGgSUAAATQIKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBo\nElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgAgCYB\nBQDQJKAAAJoEFABAk4ACAGha2ukGY4zPJflwkptV9a7JZQ8leSbJmSTXkpyvqtXpLRMAYHbsZg/U\n55M8/qbLnkpyqaoeSXJpch4AYCHsGFBV9ZUkf/Kmi59IcnFy+mKSJw94XQAAM2vHl/C2caqqbkxO\nv5Tk1G7udPLkg1laOrbHh9ylq7ezfOL4dB/jiDGPTfuZxcrK8gGuZDbM43PaK7PYyjw2mcVW5rFh\nrwH1bVW1PsZY381tV1df2e/D7cranVfvy+McBcsnjpvHxH5ncevW2gGu5vCtrCzP3XPaK7PYyjw2\nmcVWizaPe8XiXj+F9/IY4+Ekmfy8ucffAwBw5Ow1oJ5PcmFy+kKS5w5mOQAAs283hzH4pSTnknzP\nGOOPk3wqydNJnh1jfCzJN5Ocn+YiAQBmyY4BVVUf3eaqxw54LQAAR4IjkQMANAkoAIAmAQUA0CSg\nAACaBBQAQJOAAgBoElAAAE0CCgCgSUABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaBJQAABNS4e9\nADgqLl+5vqf7nTt7+oBXAsBhswcKAKBJQAEANAkoAIAmAQUA0CSgAACaBBQAQJOAAgBoElAAAE0C\nCgCgSUABADQJKACAJgEFANAkoAAAmpYOewEw7y5fub6n+507e/qAVwLAQbEHCgCgSUABADQJKACA\nJgEFANAkoAAAmgQUAECTgAIAaBJQAABNAgoAoElAAQA0CSgAgCYBBQDQJKAAAJoEFABAk4ACAGgS\nUAAATQIKAKBJQAEANAkoAIAmAQUA0LR02AsA3trlK9fb9zl39vQUVnKw9vK8kqPx3IDFYQ8UAECT\ngAIAaBJQAABNAgoAoElAAQA0CSgAgCaHMYA50jlEwPKJ41m782qSo3GIgPt9+IN5PYwEzJqjemgT\ne6AAAJoEFABAk4ACAGja13ugxhiPJ/lMkmNJfraqnj6QVQEAzLA974EaYxxL8jNJ/naSR5N8dIzx\n6EEtDABgVu3nJbz3JblaVX9UVf83yReSPHEwywIAmF0PrK+v7+mOY4wfTvJ4Vf3jyfkfSfI3qurH\nDnB9AAAzx5vIAQCa9hNQ15O8867z75hcBgAw1/bzKbyvJXlkjPFXshFOH0nyDw5kVQAAM2zPe6Cq\n6rUkP5bkN5O8mOTZqvqDg1oYAMCs2vObyAEAFpU3kQMANAkoAICmfX2Vy6xZ9K+WGWNcS7KW5PUk\nr1XVe8cYDyV5JsmZJNeSnK+q1UNa4lSNMT6X5MNJblbVuyaXbfv8xxg/keRj2ZjXP6+q3zyEZU/F\nNrP4dJIfTXJrcrNPVtV/m1w3z7N4Z5KfT3IqyXqSz1bVZxZ429huHp/Ogm0fY4zjSb6S5O3Z+Pfw\nl6vqUwu8bWw3j09nwbaN3ZibPVC+WubbfqCqzlbVeyfnn0pyqaoeSXJpcn5efT7J42+67C2f/2Tb\n+EiSvzq5z7+fbEPz4vP5/2eRJP9usn2cvesP4LzP4rUkP15VjyZ5f5KPT57zom4b280jWbzt41tJ\nPlhV705yNsnjY4z3Z3G3je3mkSzetrGjuQmo+GqZ7TyR5OLk9MUkTx7iWqaqqr6S5E/edPF2z/+J\nJF+oqm9V1f9OcjUb29Bc2GYW25n3Wdyoqt+dnF7LxqeGT2dxt43t5rGduZ1HVa1X1Z3J2bdN/lvP\n4m4b281jO3M9j53MU0CdTvJ/7jr/x7n3H4V5tJ7kt8YYL4wx/snkslNVdWNy+qVs7LZfJNs9/0Xd\nXv7ZGOPrY4zPjTFOTi5bmFmMMc4keU+Sr8a28eZ5JAu4fYwxjo0xriS5meRLVbXQ28Y280gWcNvY\nyTwFFMkHqupsNl7G/PgY42/dfWVVrefe/zcx1xb9+Sf5D0m+Nxu75m8k+beHu5z7a4xxIsmvJPlE\nVf3Z3dct4rbxFvNYyO2jql6f/N18R5L3jTHe9abrF2rb2GYeC7lt7GSeAmrhv1qmqq5Pft5M8sVs\n7Ep9eYzxcJJMft48vBUeiu2e/8JtL1X18uSP458n+U/Z3NU+97MYY7wtG7HwC1X1q5OLF3bbeKt5\nLPL2kSRV9adJvpyN9/Is7LbxhrvnsejbxnbmKaC+/dUyY4y/mI03tj1/yGu6b8YY3znGWH7jdJIf\nSvL72ZjBhcnNLiR57nBWeGi2e/7PJ/nIGOPtk68jeiTJ/zyE9d03b/yDMPH3s7F9JHM+izHGA0l+\nLsmLVfVTd121kNvGdvNYxO1jjLEyxviuyenvSPKhJN/I4m4bbzmPRdw2dmOujkQ+xvg7SX46G4cx\n+FxV/eQhL+m+GWN8bzb2OiUbHz/9xar6yTHGdyd5NslfTvLNbHwcd7dvLj5Sxhi/lORcku9J8nKS\nTyX5r9nm+Y8x/nWSf5SNTyV9oqp+/RCWPRXbzOJcNnbBr2fjo9n/9I33ecz5LD6Q5LeT/F6SP59c\n/MlsvO9nEbeN7ebx0SzY9jHG+GvZeJP4sWzsUHi2qv7Nvf5uzussknvO4z9nwbaN3ZirgAIAuB/m\n6SU8AID7QkABADQJKACAJgEFANAkoAAAmgQUAECTgAIAaPp/uCRdqN+IxcEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c2f67fc88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Histogram of number of OutLinks per node, within ego-network\n",
    "sns.distplot(outlinks_count, bins = len(index)//10, kde=False);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAGbCAYAAAALJa6vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c692f31d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Histogram of number of InLinks per node, within ego-network\n",
    "sns.distplot(inlinks_count, bins = len(index)//10, kde=False);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Identify influential friends using 'Page Rank' formulation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the adjacency matrix generated above, we can construct a column-stochastic matrix  (also called a transition matrix) such that, a column with `m` outlinks will have `1/m` as value in respective `m` cells. Conceptually, a value in the cell `a x b` gives the probability of a random-surfer in node B jumping to node A. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "binaryMap_cfr      = binaryMap_rfc.transpose()            # column-values: Outlinks\n",
    "binaryMap_cfr_norm = binaryMap_cfr / binaryMap_cfr.sum(axis = 0)\n",
    "colStochMatrix = np.matrix( binaryMap_cfr_norm.fillna(0)) # column-stochastic-matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_Initialize PageRank vector, such that all the nodes have equal PageRank score adding upto 1._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pageRankVector = np.matrix([1.0/len(index)] *  len(index)) # iniitialize page-rank-vector \n",
    "pageRankVector = pageRankVector.transpose()                # transpose to column-vector"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On applying the above Transition-matrix transformation iteratively on the PageRank vector, the vector will eventully converge such that:\n",
    "\n",
    "`Matrix.Vector = Vector`\n",
    "\n",
    "Equivalently, this is Eigen-Vector formulation with PageRank vector being the principal eigen-vector of matrix corresponding to eigen-value `1` [Since M is column-stochastic matrix, principal eigen-value is 1]\n",
    "\n",
    "Here we use Power Iteration method to solve for the PageRank vector. Inorder to handle  the nodes which have zero out-links (dead-ends) and nodes with periodic-loops (spider-traps) in the ego-network, we introduce some randomness through parameter `beta` such that a random-surfer at any node picks a random path approx. every 1 out of 6 times ( 1 - beta = 0.15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sum of Page-Rank Scores: 1.0\n",
      "Converged in 58 iterations\n"
     ]
    }
   ],
   "source": [
    "# PageRank algo: Power Iteration to solve Markov transition matrix \n",
    "# refer this     : http://setosa.io/blog/2014/07/26/markov-chains/index.html\n",
    "\n",
    "beta = 0.85\n",
    "epsilon = 999\n",
    "iteration = 0\n",
    "while epsilon > (1.0/(10**16)):\n",
    "    pageRankVectorUpdating = colStochMatrix * pageRankVector * beta \n",
    "    \n",
    "    # re-insert leaked page-ranks\n",
    "    S = np.array(pageRankVectorUpdating).sum()                      \n",
    "    pageRankVectorUpdated = pageRankVectorUpdating + (\n",
    "        1 - S) * (1.0/len(index)) * np.ones_like(len(index))\n",
    "    \n",
    "    # compute the squared-difference and check for convergence\n",
    "    error = np.array(pageRankVectorUpdated - pageRankVector)\n",
    "    epsilon = np.sqrt((error *  error).sum())   \n",
    "    iteration = iteration + 1\n",
    "    pageRankVector = pageRankVectorUpdated    \n",
    "    \n",
    "print( \"Sum of Page-Rank Scores: \" + str(pageRankVector.sum()) + \n",
    "      \"\\nConverged in \" + str(iteration) + \" iterations\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "... Cache-Miss for 1 nodes. Updating Cache...\n"
     ]
    }
   ],
   "source": [
    "# Collect the results\n",
    "\n",
    "results_df = pd.DataFrame()\n",
    "results_df['Ids'], results_df['PageRank'] = index, pageRankVector\n",
    "results_df['Inlinks'], results_df['Outlinks'] = list(inlinks_count), list(outlinks_count)\n",
    "results_df = results_df.set_index('Ids', drop = True )\n",
    "\n",
    "results_df = results_df.merge(lookup_in_cache(index), copy = False,\n",
    "                              left_index = True, right_index = True)\n",
    "results_df = results_df[['PageRank','UserName', 'FullName', 'Inlinks' , 'Outlinks',\n",
    "                         'Followers','Friends', 'Location', 'Created' ]] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2a. Nodes with high PageRank scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>PageRank</th>\n",
       "      <th>UserName</th>\n",
       "      <th>Inlinks</th>\n",
       "      <th>Outlinks</th>\n",
       "      <th>Followers</th>\n",
       "      <th>Friends</th>\n",
       "      <th>Location</th>\n",
       "      <th>Created</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FullName</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Elon Musk</th>\n",
       "      <td>0.014158</td>\n",
       "      <td>elonmusk</td>\n",
       "      <td>201</td>\n",
       "      <td>4</td>\n",
       "      <td>8845530</td>\n",
       "      <td>41</td>\n",
       "      <td>Boring</td>\n",
       "      <td>Jun-2009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bill Gates</th>\n",
       "      <td>0.011921</td>\n",
       "      <td>BillGates</td>\n",
       "      <td>168</td>\n",
       "      <td>16</td>\n",
       "      <td>34940895</td>\n",
       "      <td>183</td>\n",
       "      <td>Seattle, WA</td>\n",
       "      <td>Jun-2009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Marc Andreessen</th>\n",
       "      <td>0.011725</td>\n",
       "      <td>pmarca</td>\n",
       "      <td>192</td>\n",
       "      <td>24</td>\n",
       "      <td>645085</td>\n",
       "      <td>13952</td>\n",
       "      <td>Menlo Park, CA</td>\n",
       "      <td>May-2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Edward Snowden</th>\n",
       "      <td>0.010114</td>\n",
       "      <td>Snowden</td>\n",
       "      <td>97</td>\n",
       "      <td>0</td>\n",
       "      <td>3150406</td>\n",
       "      <td>1</td>\n",
       "      <td></td>\n",
       "      <td>Dec-2014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tim O'Reilly</th>\n",
       "      <td>0.009973</td>\n",
       "      <td>timoreilly</td>\n",
       "      <td>178</td>\n",
       "      <td>132</td>\n",
       "      <td>1987340</td>\n",
       "      <td>1709</td>\n",
       "      <td>Oakland, CA</td>\n",
       "      <td>Mar-2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WIRED</th>\n",
       "      <td>0.009110</td>\n",
       "      <td>WIRED</td>\n",
       "      <td>103</td>\n",
       "      <td>10</td>\n",
       "      <td>8974304</td>\n",
       "      <td>279</td>\n",
       "      <td>San Francisco/New York</td>\n",
       "      <td>Mar-2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nate Silver</th>\n",
       "      <td>0.008281</td>\n",
       "      <td>NateSilver538</td>\n",
       "      <td>145</td>\n",
       "      <td>26</td>\n",
       "      <td>2276360</td>\n",
       "      <td>992</td>\n",
       "      <td>New York</td>\n",
       "      <td>Aug-2008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chris Dixon</th>\n",
       "      <td>0.008217</td>\n",
       "      <td>cdixon</td>\n",
       "      <td>162</td>\n",
       "      <td>184</td>\n",
       "      <td>545467</td>\n",
       "      <td>3054</td>\n",
       "      <td>CA &amp; NYC</td>\n",
       "      <td>Mar-2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fred Wilson</th>\n",
       "      <td>0.008190</td>\n",
       "      <td>fredwilson</td>\n",
       "      <td>158</td>\n",
       "      <td>75</td>\n",
       "      <td>617149</td>\n",
       "      <td>1160</td>\n",
       "      <td>New York City</td>\n",
       "      <td>Mar-2007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ev Williams</th>\n",
       "      <td>0.008037</td>\n",
       "      <td>ev</td>\n",
       "      <td>141</td>\n",
       "      <td>104</td>\n",
       "      <td>1979132</td>\n",
       "      <td>1781</td>\n",
       "      <td>San Francisco, CA, US</td>\n",
       "      <td>Mar-2006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 PageRank       UserName  Inlinks  Outlinks  Followers  \\\n",
       "FullName                                                                 \n",
       "Elon Musk        0.014158       elonmusk      201         4    8845530   \n",
       "Bill Gates       0.011921      BillGates      168        16   34940895   \n",
       "Marc Andreessen  0.011725         pmarca      192        24     645085   \n",
       "Edward Snowden   0.010114        Snowden       97         0    3150406   \n",
       "Tim O'Reilly     0.009973     timoreilly      178       132    1987340   \n",
       "WIRED            0.009110          WIRED      103        10    8974304   \n",
       "Nate Silver      0.008281  NateSilver538      145        26    2276360   \n",
       "Chris Dixon      0.008217         cdixon      162       184     545467   \n",
       "Fred Wilson      0.008190     fredwilson      158        75     617149   \n",
       "Ev Williams      0.008037             ev      141       104    1979132   \n",
       "\n",
       "                 Friends                Location   Created  \n",
       "FullName                                                    \n",
       "Elon Musk             41                  Boring  Jun-2009  \n",
       "Bill Gates           183             Seattle, WA  Jun-2009  \n",
       "Marc Andreessen    13952          Menlo Park, CA  May-2007  \n",
       "Edward Snowden         1                          Dec-2014  \n",
       "Tim O'Reilly        1709             Oakland, CA  Mar-2007  \n",
       "WIRED                279  San Francisco/New York  Mar-2007  \n",
       "Nate Silver          992                New York  Aug-2008  \n",
       "Chris Dixon         3054                CA & NYC  Mar-2007  \n",
       "Fred Wilson         1160           New York City  Mar-2007  \n",
       "Ev Williams         1781   San Francisco, CA, US  Mar-2006  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df.fillna('').sort_values(by = 'PageRank', ascending =False).set_index('FullName').head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2b. Top 100 influential-nodes in Ego-Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Elon Musk</td>\n",
       "      <td>Reid Hoffman</td>\n",
       "      <td>Sam Altman</td>\n",
       "      <td>Tim Urban</td>\n",
       "      <td>Microsoft Research</td>\n",
       "      <td>Brian Chesky</td>\n",
       "      <td>Tony Fadell</td>\n",
       "      <td>John D. Cook</td>\n",
       "      <td>David Sacks</td>\n",
       "      <td>Clayton Christensen</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Bill Gates</td>\n",
       "      <td>Bill Gurley</td>\n",
       "      <td>Steven Levy</td>\n",
       "      <td>Joi Ito</td>\n",
       "      <td>John Markoff</td>\n",
       "      <td>Max Roser</td>\n",
       "      <td>Matt Cutts</td>\n",
       "      <td>Naval Ravikant</td>\n",
       "      <td>Tim Berners-Lee</td>\n",
       "      <td>GV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Marc Andreessen</td>\n",
       "      <td>Eric Schmidt</td>\n",
       "      <td>Chris Anderson</td>\n",
       "      <td>Benedict Evans</td>\n",
       "      <td>Neil deGrasse Tyson</td>\n",
       "      <td>Bradley Horowitz</td>\n",
       "      <td>Yann LeCun</td>\n",
       "      <td>Tim Ferriss</td>\n",
       "      <td>Recode</td>\n",
       "      <td>Mark Suster</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Edward Snowden</td>\n",
       "      <td>Paul Graham</td>\n",
       "      <td>Hilary Mason</td>\n",
       "      <td>Brad Stone</td>\n",
       "      <td>Charlie Rose Show</td>\n",
       "      <td>🍪Steven Sinofsky ॐ</td>\n",
       "      <td>Atul Gawande</td>\n",
       "      <td>John Lilly</td>\n",
       "      <td>Sarah Lacy</td>\n",
       "      <td>Drew Conway</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Tim O'Reilly</td>\n",
       "      <td>Vinod Khosla</td>\n",
       "      <td>OpenAI</td>\n",
       "      <td>NassimNicholasTaleb</td>\n",
       "      <td>Bill Gross</td>\n",
       "      <td>Gabe Rivera</td>\n",
       "      <td>Jessica Lessin</td>\n",
       "      <td>Dan Primack</td>\n",
       "      <td>Edward Tufte</td>\n",
       "      <td>Fei-Fei Li</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>WIRED</td>\n",
       "      <td>Om Malik</td>\n",
       "      <td>Steve Case</td>\n",
       "      <td>Josh Kopelman</td>\n",
       "      <td>danah boyd</td>\n",
       "      <td>Walt Mossberg</td>\n",
       "      <td>Google Research</td>\n",
       "      <td>Jimmy Wales</td>\n",
       "      <td>Sebastian Thrun</td>\n",
       "      <td>Mike Bloomberg</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Nate Silver</td>\n",
       "      <td>TechCrunch</td>\n",
       "      <td>Sundar Pichai</td>\n",
       "      <td>DAVE MORIN</td>\n",
       "      <td>Pierre Omidyar</td>\n",
       "      <td>Steven Pinker</td>\n",
       "      <td>Jonah Peretti</td>\n",
       "      <td>Paul Kedrosky</td>\n",
       "      <td>Josh Elman</td>\n",
       "      <td>a16z</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Chris Dixon</td>\n",
       "      <td>Ben Horowitz</td>\n",
       "      <td>Patrick Collison</td>\n",
       "      <td>Hunter Walk</td>\n",
       "      <td>Emily Chang</td>\n",
       "      <td>Keith Rabois</td>\n",
       "      <td>jason</td>\n",
       "      <td>Startup L. Jackson</td>\n",
       "      <td>Jeff Hammerbacher</td>\n",
       "      <td>Kevin Weil</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Fred Wilson</td>\n",
       "      <td>Kara Swisher</td>\n",
       "      <td>Max Levchin</td>\n",
       "      <td>Y Combinator</td>\n",
       "      <td>dj patil</td>\n",
       "      <td>Andrew Ng</td>\n",
       "      <td>EFF</td>\n",
       "      <td>Kevin Kelly</td>\n",
       "      <td>Bret Taylor</td>\n",
       "      <td>Mike Bostock</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Ev Williams</td>\n",
       "      <td>Chris Sacca</td>\n",
       "      <td>Nick Bilton</td>\n",
       "      <td>Lessig</td>\n",
       "      <td>Ben Thompson</td>\n",
       "      <td>Glenn Greenwald</td>\n",
       "      <td>Chamath Palihapitiya</td>\n",
       "      <td>John Collison</td>\n",
       "      <td>Paul Buchheit</td>\n",
       "      <td>Jeff Jordan</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0             1                 2                    3  \\\n",
       "0        Elon Musk  Reid Hoffman        Sam Altman            Tim Urban   \n",
       "1       Bill Gates   Bill Gurley       Steven Levy              Joi Ito   \n",
       "2  Marc Andreessen  Eric Schmidt    Chris Anderson       Benedict Evans   \n",
       "3   Edward Snowden   Paul Graham      Hilary Mason           Brad Stone   \n",
       "4     Tim O'Reilly  Vinod Khosla            OpenAI  NassimNicholasTaleb   \n",
       "5            WIRED      Om Malik        Steve Case        Josh Kopelman   \n",
       "6      Nate Silver    TechCrunch     Sundar Pichai           DAVE MORIN   \n",
       "7      Chris Dixon  Ben Horowitz  Patrick Collison          Hunter Walk   \n",
       "8      Fred Wilson  Kara Swisher       Max Levchin         Y Combinator   \n",
       "9      Ev Williams   Chris Sacca       Nick Bilton               Lessig   \n",
       "\n",
       "                     4                   5                     6  \\\n",
       "0   Microsoft Research        Brian Chesky           Tony Fadell   \n",
       "1         John Markoff           Max Roser            Matt Cutts   \n",
       "2  Neil deGrasse Tyson    Bradley Horowitz            Yann LeCun   \n",
       "3    Charlie Rose Show  🍪Steven Sinofsky ॐ          Atul Gawande   \n",
       "4           Bill Gross         Gabe Rivera        Jessica Lessin   \n",
       "5           danah boyd       Walt Mossberg       Google Research   \n",
       "6       Pierre Omidyar       Steven Pinker         Jonah Peretti   \n",
       "7          Emily Chang        Keith Rabois                 jason   \n",
       "8             dj patil           Andrew Ng                   EFF   \n",
       "9         Ben Thompson     Glenn Greenwald  Chamath Palihapitiya   \n",
       "\n",
       "                    7                  8                    9  \n",
       "0        John D. Cook        David Sacks  Clayton Christensen  \n",
       "1      Naval Ravikant    Tim Berners-Lee                   GV  \n",
       "2         Tim Ferriss             Recode          Mark Suster  \n",
       "3          John Lilly         Sarah Lacy          Drew Conway  \n",
       "4         Dan Primack       Edward Tufte           Fei-Fei Li  \n",
       "5         Jimmy Wales    Sebastian Thrun       Mike Bloomberg  \n",
       "6       Paul Kedrosky         Josh Elman                 a16z  \n",
       "7  Startup L. Jackson  Jeff Hammerbacher           Kevin Weil  \n",
       "8         Kevin Kelly        Bret Taylor         Mike Bostock  \n",
       "9       John Collison      Paul Buchheit          Jeff Jordan  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_df = pd.DataFrame()\n",
    "temp_df  = results_df.sort_values( by = 'PageRank', ascending =False)\n",
    "\n",
    "for i in range(10):\n",
    "    dummy_df[i] = list (temp_df [10*i : 10* i + 10]['FullName'])\n",
    "dummy_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2c. Histogram of PageRank scores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "PageRank scores are scaled such that nodes have an average-score of 1. So the scores below give an idea of how influential are the nodes, with respect to an average node. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlAAAAGpCAYAAABClwgJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAACdBCgAgE4CFABAp/8HuZiOTVp4yqMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c2ec9be80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pageRank_to_plot = len(index) * results_df[\"PageRank\"]\n",
    "sns.distplot(pageRank_to_plot, kde=False, rug=True, bins = len(index)//10);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A joint-plot showing how Inlinks and Outlinks of the nodes are distributed (within the ego-network)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAasAAAGoCAYAAAD4hcrDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XmcXFWd///Xrep9S3pLJ52ts56QhJCwhIRNMCxhRxyR\nwSUq48w4jI7i6Ah+cR0QFRV/gqiDKCiCICAIQgQEQSEEQiBkOyH71kl3J51e0mstvz+quukkvVR1\n13Kr6v18PJqqunVv1ed2k3rXOffcc51gMIiIiIibeZJdgIiIyFAUViIi4noKKxERcT2FlYiIuJ7C\nSkREXC8r2QWMRH19S1KGMpaWFtDY2JaMt06aTNxnyMz91j4nV2VlsZPsGtxILathyMryJruEXsFg\nkEScfuCmfU6kTNxv7bO4UUq3rDJRa3s3b26q5+3NDdQ1ttPQ1EG3P0BxQTajCnKYNn4Us2tKmV1T\nRn6u/rwikh70aZYiGpraeeylbby2YT+BQKgllZvjpaQwm2yvh7ZOP7UH2thZ18oLq/eQm+1h8dxx\nLDlxPOMri5JcvYjIyCisXK7bF+CJf2xj+cqd+PxBKkblMaemDDNpNKOLco9Y1x8Isu/AYbbta2Ht\n1gO8uHoPL67ew6mzq/jg+6ZSMSo/SXshIjIyCisXazjUzl2Pr2VbbQvFBdmcNa+a2TWlOE7/x1+9\nHofxlUWMryzitDlj2bK3iVfW7eO19ftZZetYeupkLj2thuwsHaoUkdSisHKpDdsPcucf19LW4WPu\nlDLOPXkCOVEcBPZ4HGZMGM308aNYv6ORl97ey5OvbOdNW8cnLz6OadWj4li9iEhs6Su2C729uYEf\nPfw2nV1+li6cxIWnTooqqPpyHIc5NWV86qLjWDCjgr0H2rjlN6v40z+29R77EhFxO4WVy7yxsY47\nHn0HgA+eNZV508oH7PaLRm62l/NOnsjVS6ZTlJ/NYy9v44cPvUXT4a4Rv7aISLwprFxk3baD/OyJ\ndXg9Dh86ezo140pi/h6TxhTziaWzmFpdwvrtjXzjnpVs2H4w5u8jIhJLCiuX2Lm/hTsfewcH+OD7\npjFxTPyGm+fnZvHBs6Zy9vxqWtq6uO3Bt3j87+oWFBH3Uli5wMHmDn788Bo6uvxcvHhyXIOqh+M4\nLDyuimvOnUlxQTaP/30bP3r4bVra1C0oIu6jsEqybl+An/5xLY2tnZw9v5pZk0oT+v7VFYUsC3cL\nrtt2kG/86nW27GlKaA0iIkNRWCXZg399l617m5ldU8ops8YkpYaebsEz543jUGsnt97/Js+9sSsh\ncw6KiERCYZVEr6yt5YU391A5Oo8LTpkUk1F/w+U4DovnjOWqs6eTm+3ld8+9y8+fWEdHly9pNYmI\n9FBYJUntgcPc94wlN9vD5WdMcc2sEpPHFrNsqWF8RSErN9Tx7XvfYG/D4WSXJSIZzh2fkBnG5w/w\niz+tp8sX4IKFkygrzkt2SUcoLsjh6iUzOMlUUnugjW/f+zovrd6d7LJEJIMprJLgsZe2smNfC8dP\nKUv4gIpIeT0OS06cwGWn1xAMwvd/u4r7/7IJnz+Q7NJEJANpbsAE27ijkWde20lpcS5LTpqQ7HKG\nNGtSKZWj8/nTqzt4/s3dbK1t4t8um8OY0oJklyYiGUQtqwRq7/Rxz583gAOXLJ5MTnZqXJ20vCSP\n/7hyHnNqythW28I3fvU6r67bl+yyRCSDKKwS6OEXNtPQ1MGi2VWMKy9MdjlRycn2cvHiyVy8aDKB\nQJD/+9N67n5yPe2dGi0oIvGnbsAEWbvtAC++tZfK0XmcNmdssssZtjlTyqiuKOTJV7bzytp9bN7T\nxL9fPoeasbGfx1BEpIdaVgnQ1tHNr/68EY8DFy2ajNeb2r/20uJcrjl3BguPG0NdYzs337eKZ17b\nSUAnEYtInKT2p2aKeOD5d2ls6WTx3LFUpcnABK/Xw9nzx/Ohs6eRl+PloRc2c/tDb3OotTPZpYlI\nGlJYxdlbmxv4xzv7qCrNZ9Hs1O3+G8iUcSV84sJZTB1XwtptB/naL19jla1LdlkikmYUVnHU2t7N\nvU9vxOtxQt1/nuRNpxRPhXnZfPB9Uzn3pAl0dPm587G1/FKDL0QkhhRWcXT/s5toOtzF6cePpXJ0\nfrLLiSvHcThxZiXLls5ibFk+/1i7j6/9ciV2Z2OySxORNKCwipPXN9bx2vr9VJcXsHBWVbLLSZjy\nkjw+cp5h8ZwqDrZ08N3frea+5Za2DrWyRGT4FFZx0NTayW+WW7K8oe4/T5p2/w3E63E4c141Hzl3\nJhWj8nhx9R6+evcKHcsSkWFTWMVYMBjk3mcsre3dvG/+eMpK3DVJbSJVVxSy7ALDGcePo7Wtmzsf\nW8tPHlnDweaOZJcmIilGJwXH2Ctr9/HW5gYmjSnixBkVyS4n6bxeD6fNHYuZNJq/vL6L1e82sHbb\nQS5YOIkLT51Efq7+FxSRoallFUMHmzu4/9lN5GR7uPDU5F5M0W3KS/K4+v3TufDUSeRme3nyle38\nz89e5clXttPW0Z3s8kTE5fS1NkYCwSD3/HkDHV1+li6cxKii3GSX5DqO43D81HLMpNG8sbGe1zfW\n8ehLW/nzih2cOa+as04Yx/jKomSXKSIupLCKkeUrd7J+eyNTq0s4fmpZsstxtZwsL6fNHctJMyt5\na3MDb9g6nn1jF8++sYsp40o4yVSyYEYFY8sK1DoVEUBhFRNb9zbzyN+2UpiXpe6/KOTmeDl1dhUn\nm0o2721mzeYGtu9rZlttM394cQulxbnMnDiaqeNKqK4spLq8kNFFOfr9imQghdUItXX4+NnjawkE\nglyyuIbCvOxkl5RyvF4PZuJozMTRtHX62Lqnic17m9ld18pr6/fz2vr9vevm53gZW17I2LICqsry\nqSotYExp6LYgT/87i6Qr/esegUAwyC+fWt97jarJY4uTXVLKK8jNYu7UcuZOLScYDHKwpZP9B9s4\n0NzJweYOGpo62LG/hW21zcdsW5SfTVVZPmPLCqgZW8KUcSVMqioiK8VnuRcRhdWIPPXKdla/Gxqm\nfsbx45JdTtpxHIfykjzKjzpXLRAI0nS4i8aWThpbO2ls6eRQSycHWzrZureZLXua+cc7oSsZ52Z7\nmDFxNHNryjjJjKF8VOae9yaSyhRWw/T25gb++PI2Sgqyuez0moybpSKZPB6H0uJcSouPHXHpDwRp\nbOlg34E29h5oY1ddK2u3HmTt1oM8+NfNTKsu4ZRZYzh51piMPmFbJNUorIZh654mfv7EOrxehyvO\nnEqBjlO5htfjUDEqn4pR+cydWg5AS1s3W/Y2YXc2srW2mS17m3nwr5uZOWEUZ55QzclmDLk53iRX\nLiKDUVhFqeFQO9+5/006uvxcdnoNY8vS42KK6ay4IJv50yuYP72Cwx3dbNrVxMadjWza3cSm3U3c\n/+wmFs0Zy5nzxlEztlijDUVcSGEVhea2Ln740Ns0tnTy/hPHM2tSabJLkigV5mWzYEYFC2ZUcKi1\nk3e2HuSdrQd4cfUeXly9h4ljijjrhGoWzanSyE4RF3GCwWCyaxi2+vqWhBV/qLWT7z+wmtoDbZw5\nfzyLjxuTqLd2hdGjCzh0qC3ZZcRFIBBk275m1mw5wJY9zQSCQbK8DifPGsNlZ01nTEkOngxqbVVW\nFlNf35LsMhLKTftcWVmcOf+zRUEtqwgcbO7gew+spq6xnZNNJUsXTaapqT3ZZUmMeDwO06pHMa16\nFK3t3azbdpA1Ww+wYt1+VqzbT2lRLvNnhlpjsyaVaii8SBKoZTWELXuauOPRd2g63MXiOVWccfw4\nSksL07aVMZB0bln1JxgMsrv+MHZ3E+u3HaCjyw+ETko+flo5s2vKmDlxNFWl+Wl3jMtNrYxEcdM+\nq2XVP7WsBhAMBnl5TS2//YvFHwhy9vxqFh6XOVf8zXSO4zBxTBHHzxzDwYPV7KpvZfPuJjbvaWLl\nhjpWbghdSLIgN4vqikKqKwqoLi+kuqKQytH5lBTmkJfjTbsgE0kWhVU/Gls6+e1fLKvfbSAvx8uV\nZ9ZQM64k2WVJkng8DpOriplcVcz7TxxPQ1MHu+tb2V1/mP2NbWzZGwqxo+VkexhVmENJYQ6Fednk\n5XjJz80iL8dLXk4W+Tle8vo8zsvxUpCbRV7ue+t5PepyFAGF1RG6uv28+NZeHv/7Vto7/UwcU8SF\np05itC73IWGO41A5Op/K0fksmFEJgN8foLG1k4am0HRQLYe7ONzp43B7N4c7fDQ0NTPc3vacLE8o\nuHK95Odk9YZYfm4W+Tmh5QW5oeVF+dkUF2RTVJATus3L1snqkjYUVkBLWxevrN3HM6/tpOlwFznZ\nHs4/ZSInTCtXN44Myev19J6I3J9gMEiXL0BXt/+92+4AXb7QbWe3v/d+V7efzvBznV1+OsPLW9u6\nOdDUgc8feeo5QEFeFsUFORQVZFOcn01xQQ4lheHbghxKCrIpLgzdL8pXuIl7ZWxYNbZ0sn77Qd56\nt4G3NjfgDwTJzvKwaHYVJ88aQ4Euty4x4jgOudlecrNHPkuGPxAMB5r/vWDr9tPR5ae900d7l4+2\nTh/tHT7aOkPLmtu62N/YNmTrzgGKCrIZXZxHTpbT2zXZ002Zk+Uhu+fHG7rNCj/OyfIesfyI9bK9\n763vdfQFUIYlbT+RA8Egre3dNLV2cai1k0OtnTQ2d7K7vpWdda3UNb439LxyVB5zp5YzZ0qZQkpc\nzetxQl2AUf5/GgwG6ejy09bpo63DF77t7nPfx+EOH22d3Rw41E57l2/YXZdD6Rt0RwRgb9h5yfI6\n4XAL/by3jYODQ9+8cxwHB8AhfBt63LOOxwkFpMcBx+PgOeq+48DoUfkcPtxJlqfnPR28Hgdv+P29\nHocsr9N73xte573nPGqVxllKfzLf8ttVdHT68fkDdPsCoVt/AJ8vdDvQP7a8HC9TxhVTM7aEmrHF\nVIzK07c9SWuO817IlQ8xVmj06AIaGw/3tuJCXZcB/IHQvyu/P4jPH8DnD+L3B/AFQo/9/mD4+dBz\nvsCR6/ZuEwj03u/q9tPW4Qu/dpBAIHVPpXEc8Hr6hJjXIaufwPN6POFUPVaWx+H7//W+xBaeIlI6\nrHbXteI49H77ys3xUujNDn0r83ooLshhVGEOo4pCt6OLchlXXkBpce6IwqmispiGgpT+1UUtE/cZ\nMnO/k7nPgXDwdYfDrOdLaM9Pty/0JTQI9HwbDYb/EyTUgjxyWTC0fjBIIBDqcQkGQ7e99wNBCoty\naWpqD4VuINgbsv5A8MhlfQL4iGWBYG9I+wM9gRxa1hvI4WD3D3LcUYM/B5bS/wp/en1yvoHkZnvJ\nicHxh1SSifsMmbnfmbjPbjopWPqnHBcREddL6emWREQkM6hlJSIirqewEhER11NYiYiI6ymsRETE\n9RRWIiLiegorERFxPYWViIi4nsJKRERcT2ElIiKul9JzAzY0dWj6DRFJKxWj8iKaZbu+viXoDwT4\nzA9eYu6UMj73T/PiXVpCVFYW97v/almJiKSonlnqff5AskuJO4WViEiKCoQzKhOux6ewEhFJUT0Z\nFST9j4gorEREUpTHE0qrVL7CcqQUViIiKUphJSIirudxHBwyI6ziPnTdGOMF3gD2WGsvMcaUAb8H\naoDtwFXW2sbwujcA1wJ+4HPW2uXxrk9EJJV5PA7+DLiIbiJaVv8FbOjz+CvA89baGcDz4ccYY2YD\nVwNzgKXAT8NBJyIiA/B4nIxoWcU1rIwxE4CLgbv7LL4cuDd8/17gij7LH7TWdlprtwGbgYXxrE9E\nJNV5HKd3CHs6i3c34O3Al4HiPsuqrLW14fv7gKrw/fHAij7r7Q4vE5EEWvHqP7j9B98lEAhw6eUf\n4GPLrj1mneXPPMX99/2KYDBIQUEh//0/X2XGTAPAg7/7DX96/FEcx2Ha9BnceNO3yM3NTeg+DFZf\nX5/59Cdoa2sDoLHxILNnz+XW226P+H2am5q46atfZl/tXsaOq+bbt3yfkpISfL5uvvO/32ST3YDf\n72fpRZfy8U8c+3scidLSArKyvGRleXA8DpWVxUNvlMLiFlbGmEuAOmvtKmPM2f2tY60NGmPSv/0q\n0g+fz0dWVuz/Cfr9frze4fWg+/1+fvC9W7j9jp8zZkwV/7LsGs4482ymTJ12xHrV1eO542f3UFJS\nwquv/J3vfedb/N+v7qe+bj9/+P3vuP/3j5Gbl8dNN3yJ5559hosvuTwWuxaxgeo72l3/9+ve+zf+\nz/WcedY5Ub3Pb+69h5NPWcjHll3Lb+79Jb+995f8x2e/wF+fe5bu7i5+88AjdHS085EPX8l55y9l\nXHXsvn83NraF7wXp6vZTX98Ss9dOpoFCN54tq9OBy4wxFwF5QIkx5rfAfmPMOGttrTFmHFAXXn8P\nMLHP9hPCy0SSqnbvHq7/r//AzJrNJruBKVOncdM3/pe8vHw2bljPT26/jfb2NkaNGs1Xv/5tKioq\neeKPj/D4Y4/g6+5m/MSJfO2bN5OXl8//fvMmcnJyeHfTRo6fN58z33cOt//ge0DoBM87f/4rCgoK\nuPMnP2LFK3/HcRyWferTnHveUt5c9Tr3/N/PGDV6NFu3bMbMms3Xv3ULjuPwwcsvZMm55/P6yhV8\n5GOf4NzzLxzWvm5Yt5YJEyYyfvwEAJacv5SXX3rxmLA6ft783vtz5s6jrm5/72O/309nZyferCw6\nOtqpqKgE4LFHHgLgAx+86ojXeurJx3npxb9yuLWV+vo6Llh6MZ/69L8Pq/5I6uvP4dZW3nxjJV+9\n6VsAtLe38aPbbmXrls34fD6u/fRnOPN9xwbZyy+9wB0/+yUAF158Gf/579fyH5/9Ao7j0NHejs/n\no7Ojk+ysLAoLi0a0TwPxOplxzCpuYWWtvQG4ASDcsvpva+1HjTHfB5YBt4ZvHw9v8gTwO2PMD4Fq\nYAawMl71iURj547t3PD/vsG8ExZwy7e/xqN/eIirrr6GH912K7fedjulpWU89+wz/OKun3DjTd/i\nfWcv4bIrPgjAL+66gz89/hgf+vA1ANTX7ednd9+H1+vly9d/li9++QbmnbCAtrY2cnJyePGF53l3\nk+Xe+x+m6dAh/uUT1zB/wUkAbLIb+e2Dj1JRWcm//8sy1ry9mhPmnwjAqFGj+dVvfn9M7cufeYrf\n/ebeY5ZPmDiRm2/9wRHL6uvrGFM1tvfxmDFjWLfunUF/N08+8RiLFp8BQOWYKv75o8u48rILyM3N\n45RTF3PqotOAY0Oqrw3r1vKbBx4hLy+Paz9xDYtPP5PjZs85Yp2bbvwSO3fsOGbbq6/5GBdefGlE\n9Q3kpb+9wEmnnEphUShQ7v3V3Zx08kJuvOlbtLQ08+lPfISTF55Kfn7BEds1HjzYG8bl5RU0HjwI\nwDlLzuXll17g8ovOpaOjnc994UuUjBo1aA3DlSkDLJIx6/qtwEPGmGuBHcBVANbadcaYh4D1gA+4\nzlrrT0J9IseoqhrLvBMWAHDB0ot5+PcPsGjxaWzdupnP/2eoFRAI+CmvqABg69bN/OKuO2htbaG9\nrY2F4Q9sgHOWnN/bTXf8CfP5ye23cf7Si3jf2ecypqqKNW+t5rzzl+L1eikrL2f+iSexYf06CgsL\nOW7OXMZUhQ7zzphp2Fe7tzeslpx3Qb+1X7D0Yi5YenFcfi+r3ljJk088xl2/+DUAzc3NvPy3F3j4\nj3+muLiY//eVL7H86Se54MJLBn2dUxYuYtTo0QC875wlrHl79TFh9e1bvj/i+gby3F+e5pLLr+x9\nvHLFq/z9pRf53W/vA6Crq4v9+/ZRM2XqgK/hOE7v9Efr163F4/Hy+J+fpaW5mc/86yc5eeGi3hZr\nLHk8Dn6FVWxYa18EXgzfPwAsGWC9m4GbE1GTSFSOmijUcSAYhClTpvGLe35zzOo3f/MmvvP925kx\n0/DUk4+zetUbvc/l5+f33v/Ysms57fSzePUfL/Pvn17Gj/6/uwYtIyc7u/e+x+vB53/v+1xen9ft\nK5qWVWXlGOr27+t9XFdXR2Vl1dGbArD53U3cevM3+cHtd/YGzRsrV1BdPZ7S0jIgFDzvrHl7yLA6\n9vd77MSs0bas+quvP4cONbJ+3Vpu+d6P+iwNcvN3f8jkyTVHrHvzt25ik7VUVFbyg9vvpLSsjIaG\neioqKmloqGd0eL+fXf40ixafRlZWNqVl5cw7YT4b16+LT1g5CisRCdu/r5a1a95m7rwT+Mvyp5k3\nfwGTJtdw6FBj73Kfr5udO3Ywddp02traqKiowOfr5i/P/JnKyjH9vu7u3buYNn0G06bPYMOGdezY\nvo0TFizg8Uf/wIUXX0ZzcxNvrX6T6z57PTt2bBtW7dG0rGbNnsPuXTvZu2c3lWOqeP4vz/D1b3/n\nmPX27avlxv+5nq9982Ym9flArxo7lrVr19DR0U5ubh5vvP4as46bDcAfHnoAgH+66p+Peb3XV66g\nuamJ3NxcXv7bC9zw/755zDrRtKwGqq8/Lzz/LKedcdYRIxYXLjqNPzz0O67/7xtwHIdNdgMzzXF8\n9WvfPmLbM846m6efeoKPLbuWp596oneARlXVWFa9sZKlF11Ke3sb69a+w1VXfzTi+qPh8Th0+9J/\n7LrCSiQCkybX8MgfHuSW//06NVOm8oEPXkV2djb/e+tt3H7bdznc2orP7+PD//xRpk6bzqf/7To+\n/cmPMrq0lNlzju8dHn20hx74LW+ueh2Px8OUqdNYdNoZZGdns/adNSz7yIdwHIf/+OznKa+oGHZY\nRSMrK4svfOkGrv/cZ/AHAlxy6RVMnTYdOHKAxK/u/jnNTYe47bu3AOD1ernnvgeYM3ce5yw5j09+\n7Gq8Xi8zzSwu/8A/AbBjxzbmzVvQ7/vOnj2XG79yPfV1oQEWR3cBRmug+gC++Pnr+MpXv977BeL5\nZ5fz0WWfOmL7T37qX/nxj77Hx6/5JwKBANXV4/n+j+445n0+9vFPcdONX+LJJ/7I2LHjegP1yg9d\nzS3f+hof+fAHALjoksuZPmPmiPZpIN4M6QZ0gik8TYeuFCyJULt3D1+6/rP89sFHk11KSvvSF/6T\nW773I7L7dGVCaDTgxg3r+OKXbkxSZe4SzZWCAW66+zUaWzq54wtnxbewBBnoSsFqWYlIQvTXMpGR\n0wALEQFgXPV4tari6OJLLk/4ScPpxOM4BFK4hyxSukSIiEgKy5TzrBRWIiIpzOPJjOtZKaxERFJZ\nEIhoSEZqU1iJiKSwQDCI15P+aaWwEhFJYf5AsN8ZP9KNwkpEJIUFg6ERgelOYSUiksICgSCeDPgk\nz4BdFBFJX4FgUC0rERFxN4WViIi4XjAAjkYDioiIm4VaVsmuIv4UViIiKSwQDOLJgLRSWImIpLBA\nQMesRETE5XSelYiIuJ66AUVExPVCJwUrrERExMV0npWIiLheIICmWxIREXdTy0pERFwvqAEWIiLi\ndjrPSkREXC8YhAzIKoWViEgqC4KuFCwiIuIGCisREXE9hZWIiLiewkpEJMWl/xErhZWIiKQAhZWI\nSIoKBoPJLiFhFFYiIikuA0auK6xERMT9FFYiIuJ6CisRkRSVOUesFFYiIpICFFYiIuJ6CisRkVSV\nQf2ACisRkRSnWddFRERcQGElIiKup7ASEUlRwQw6aKWwEhER11NYiYiI6ymsRERSVAZNuq6wEhFJ\ndRkwcl1hJSIi7qewEhER11NYiYiI6ymsRERSXAYcslJYiYiI+ymsRERSlIaui4hI6siAsesKKxER\ncT2FlYiIuJ7CSkQkZWXOQSuFlYhIikv/I1YKKxERSQFZ8XphY0we8BKQG36fP1hrv26MKQN+D9QA\n24GrrLWN4W1uAK4F/MDnrLXL41WfiEiq09D12OgE3m+tPQGYDyw1xiwCvgI8b62dATwffowxZjZw\nNTAHWAr81BjjjWN9IiLpIQP6AeMWVtbaoLW2NfwwO/wTBC4H7g0vvxe4Inz/cuBBa22ntXYbsBlY\nGK/6REQkdcStGxAg3DJaBUwH7rTWvmaMqbLW1oZX2QdUhe+PB1b02Xx3eJmIiByltLQAf7gbMCcn\ni8rK4uQWFGdxDStrrR+Yb4wZDTxmjJl71PNBY0wG9bqKiMRGY2Mbnd1+ALq6fNTXtyS5otgYKHQT\nMhrQWnsIeIHQsaj9xphxAOHbuvBqe4CJfTabEF4mIiKDcDLgoFXcwsoYUxluUWGMyQfOAzYCTwDL\nwqstAx4P338CuNoYk2uMmQLMAFbGqz4REUkd8WxZjQNeMMasAV4HnrXWPgncCpxnjHkXODf8GGvt\nOuAhYD3wDHBduBtRRET6k0EHUeJ2zMpauwZY0M/yA8CSAba5Gbg5XjWJiKSjDJh0XTNYiIiI+yms\nRETE9RRWIiIpKphBB60UViIi4noKKxERcT2FlYhIitKs6yIikjIyYOS6wkpERNxPYSUiIq6nsBIR\nEddTWImIpDgnA+ZbUliJiIjrKaxERFKUhq6LiIi4iMJKRERcT2ElIiKup7ASEUlZmXPQSmElIpLi\nMmDkusJKRETcT2ElIpKiMqcTUGElIiIpQGElIiKup7ASERHXU1iJiIjrKaxERFKUJzxmPRBI/6EW\nCisRkRTVG1bpn1UKKxGRVOUJf4IHMmD6dYWViEiKctQNKCIibufxhMIqqJaViIi4lQZYiIhISvA4\njgZYiIiIu3k8jgZYiIiIu3k84M+AppXCSkQkhXkch6DCSkRE3Cx0zEphJSIiLhY6ZpXsKuJPYSUi\nksIcR0PXRUTE5dQNKCIirufxOGpZiYiIu+k8KxERcT2PjlmJiIjbabolERFxPY+jY1YiIuJyjsfR\nJUJERMTdPI6uFCwiIi7nOA6BQLKriD+FlYhICtNJwSIi4no6KVhERFzP49F5ViIi4nIexyEIaT8i\nUGElIpLCPI4DpP+IQIWViEgKc8Kf4uk+IjCisDLGnGWMKQrfv9YY8zNjzJT4liYiIkNRy+pIdwCH\njTFzgC8CO4Ffxq0qERGJSG9Ypfkgi0jDymetDQIXAndZa28BSuNXloiIRMLjUcuqryxjzKnAlcBf\nw8u88SmRSUuzAAAeQklEQVRJREQipZbVkW4Cfg68aq1dZ4yZCWyOX1kiIhIJT88Ai/TOKrIiXO+v\n1trHex5YazcZYz4fp5pERCRCalkd6VFjTG+3nzFmHPB0fEoSEZFIOeGw0knBIc8AvwYwxlSGH98U\np5pERCRCvd2Aad6yiqgb0Fr7A2PMj40xtwNnAN+11j462DbGmInAfUAVEAR+Ya39sTGmDPg9UANs\nB66y1jaGt7kBuBbwA5+z1i4f1l6JiGQInWcFGGNm9/wAdwOnAy8Cb4WXDcYHfNFaOxtYBFwX3uYr\nwPPW2hnA8+HHhJ+7GpgDLAV+2rfrUUREjvVeWCW5kDgbqmX1VD/LPhj+CQJTB9rQWlsL1Ibvtxhj\nNgDjgcuBs8Or3Uso/P4nvPxBa20nsM0YsxlYCLwa4b6IiGScnvOs/GmeVoOGlbU2JlMqGWNqgAXA\na0BVOMgA9hHqJoRQkK3os9nu8DIRETlKaWkBWVleigpzASgpyaeysjjJVcVPpEPXMcYUABP6bmOt\nXR/BdkXAI8DnrbXNxpje56y1QWNMen8dEBGJg8bGNgB83T4A6upbKM5J/bnJBwrcSCeyvQ7YDzxH\nqGvwKeDJCLbLJhRU9/cZkLE/PPS9Zwh8XXj5HmBin80nhJeJiMgAsrNCH+NdvvSedj3SltUXgbnW\n2h2RvrAxxiE02e0Ga+0P+zz1BLAMuDV8+3if5b8zxvwQqAZmACsjfT8RkUyUkxUah9bt8ye5kviK\nNKz2RRNUYacDHwPeMca8FV52I6GQesgYcy2wA7gKIDyN00PAekIjCa+z1qb3b19EZISywi2rbrWs\nAHjWGPM94EGgo2fhYMesrLV/B5wBnl4ywDY3AzdHWJOISMbLUTfgET4evv1Qn2WDDl0XEZH4U8uq\nj1gNYRcRkdjqbVl1p/dRk0HDyhiTa63tDA9bP4a1ti0+ZYmISCSy1bICQrNHnAi0Eur263sMKogu\nwCgiklQ9owE7M7llZa09MXyb+meaiYikoZwMaVkphEREUlh2ds8xq/QOq6GOWdUT6u47mgMErbVj\n4lKViIhEpLcbMMNPCj45IVWIiMiw5GRrNCA9s1YYY75srf1e3+eMMV8GvtfvhiIikhDvDV1P727A\nSI9ZXR3hMhERSaiBJgpKL0MdszoPOB+oDk+31GMUmfIbEhFxsWD4cvZOmn8iD9Wy6uK9c6wO9/nZ\nCHwgvqWJiMhQAuGw8qR5Wg11zOpvwN+MMY9Ya9cmqCYREYlQOKtwPBkcVn18vO8VfntYa78c23JE\nRCQa77WsklxInEU6wKJvF6AfuBCojFdRIiISmZ6WVUZ3A/aw1n6z72NjzC3AH+JSkYiIRCzQO8Ai\nvcNquNMttQKTYlmIiIhELxgIdwOm+eR5EbWsjDHf571plzzAScCGeBUlIiKRCfQMsEjzllWkAyx6\nhq8D+ICfAY/GpSIREYlYUEPXQ4wxZwBnAieEF70NvGyt7Y5nYSIiMrSelpUnzYcDDtrLaYy5HHiA\n0GCK88I/jwD3G2OuiH95IiIymEyZwWKoltXXgKXW2nV9lr1ljHkZuA/4Y9wqExERCRtq/Ej+UUEF\nQHg2i7z4lCQiIpHyhrv//P7+Lj2YPoYKqxxjTPbRC40xuUBufEoSEZFIeb09YZXZlwh5HLjPGDOq\nZ4ExZjTw6/BzIiKSRN7wCVb+QHq3rIY6ZnUDcBewyxizObxsOvBw+DkREUmirHDLypfmLauhZl3v\nAq41xnwTOJ7QNaze6bmCsIiIJFdOtheAjq4Mvqx9D2vtTmBnnGsREZEo5WZ7yc7y0Hy4K9mlxFWa\nzyYlIpL+CvOyaFJYiYiImxXmZdPS1kUgjQdZKKxERFJcYV4WgSC0tqfvLHgKKxGRFFdckANAQ1NH\nkiuJH4WViEiKKx8VmlCo9sDhJFcSPworEZEUV14SCqs99QorERFxqTGl+QBsrW1OciXxo7ASEUlx\nudleKkflsa22OW1nslBYiYikgfGVRXT7Amzdm56tK4WViEgamDKuBIB3th5IciXxobASEUkDk6uK\n8HocVm+q7716cDpRWImIpIGcbC/Txo9i74E2du5vTXY5MaewEhFJE3NqygD4xzu1Sa4k9hRWIiJp\nYuq4Yorys3l5TS2HO9Jr6iWFlYhImvB6PZxkKuns9vPCm3uSXU5MKaxERNLI/OkV5GZ7Wb5yZ1q1\nrhRWIiJpJDfby6I5VRzu8PHE37cnu5yYUViJiKSZk2ZWMrooh+ff3M2e+vQYGaiwEhFJM1leD+8/\ncQKBQJC7n9qQFlMwKaxERNLQ9PGjmDuljB37Wnjq1R3JLmfEFFYiImlqyYkTKC7I5ol/bGPd9oPJ\nLmdEnFSelqOhqSN1i48Tu6Mx2SVgJpcmuwSRlFUxKs+JZL36+pYgQEeXj+21LQOut6fhMA88/y75\nOV6+9olTqBydH6NK46Oysrjf/VdYpTg3hFOk4hViQ/0OFJ6SSmIdVgBvb25g+eu7mDimiBs/dhK5\n2d4YVBofA4VVVqILkdhIpZDqcXTNww2RaPe97/oKLslEJ0yvYF9jG29vPsCvn97Iv146G8eJKBNd\nQ2GVglIxqPoTaXjFcn97XkuhJZnm3BMn0HCog9fW72dyVTFLT52U7JKiom7AFJIuIeUWCixxo3h0\nA/Zobe/mvuWWwx3dXH/VfOZMKRtBpfExUDegRgOmCAVV7Ol3KpmmKD+bK86YgsdxuOvxtdQ1tiW7\npIgprFKAPlTjR79byTTVFYWcf8pE2jp8/OSRd+jo8iW7pIjomJXLxerDdGOE51jMqnFft4CIxNbx\nU8vZ39jOm5vq+eWTG/jMB+bicfmAC4WVi40kqCINp0i3S+cQszsadfxKMs45C8bTcKidVZvqefKV\n7Vx2+pRklzQodQOmmY3bDw47qCJ53Xi8togkntfjcNkZUygpzOHxv29jy56mZJc0KIWVSw2nVZWo\nIFFwiaSHgtwsLl40mWAQ7n5qPZ3d/mSXNCCFlQtFG1TJDI50CS0NtJBMNXFMESebSvYfbOeJv29L\ndjkDitsxK2PMPcAlQJ21dm54WRnwe6AG2A5cZa1tDD93A3At4Ac+Z61dHq/a0olbgmLj9oNpfVxL\nJJ2dOa8au+sQz76xiyUnTaCsJC/ZJR0jni2rXwNLj1r2FeB5a+0M4PnwY4wxs4GrgTnhbX5qjHHv\n5FVxlMrf8N0SnMOhARaSybKzPJxx/Dh8/iB/fNmdrau4hZW19iXg6E+vy4F7w/fvBa7os/xBa22n\ntXYbsBlYGK/a0oUbw8GNNYnI0ObUlFFekscr6/bRdLgr2eUcI9FD16ustbXh+/uAqvD98cCKPuvt\nDi/LKNG0qhQKIpmttLSArCwv7Z0+DrXHZmDE4nnjePLv21izrZErz5kek9eMlaSdZ2WtDRpjMmpu\nPxGRWGkMT5XU0eXj0KHYTJs0ZUwRXo/D8hXbOXNu1dAbxEFlZXG/yxM9GnC/MWYcQPi2Lrx8DzCx\nz3oTwssyRjq1qtxe39F0vEokJD83i8lVxextOExjS2eyyzlCosPqCWBZ+P4y4PE+y682xuQaY6YA\nM4CVCa4tI9idjUf8iIj0NamqCIANO9z1pTNuYWWMeQB4NXTX7DbGXAvcCpxnjHkXODf8GGvtOuAh\nYD3wDHCdtda9Z6fFWKJGAPYXTgosEelr4phQWG2L8LIjiRK3Y1bW2n8e4KklA6x/M3BzvOpxq+Gc\nABzzGnY2YiZlZleYugBFjtRzjtW+g+66fIhmsMgQQ7WgMrGFpaASOVZutpei/GxqDxxOdilHUFgl\nUSqfACwi6ausJJeDzZ10drnnaIzCKkncNlFtpnUDqlUlMrDycFfgfhddSVhhJSIiR+g5blV7QGGV\n0ZLRqkp0yynVzrUSkfeUFecCuOq4lcIqwXScSkTcrqcbcG+DwkqioFaKiCRScUE2BXlZbNnbnOxS\neimsEshtgyp6xKuL0M0hqxauyMAcx2F8RSGNLZ0caOpIdjmAwiphkh1UmXge1VAUWCIDmxSeyWLN\nloYkVxKisHKpdAkqN7euRGRgMyeOBuD1jXVDrJkYCqsESMaUStFMVhvvMHNzYKl1JdK/4oIcxlcU\nYnceov5Qe7LLUVjFW7SX/hjpB/twZ1PP5G5CBZZI/+bPqCAILF+5M9mlKKzcYiQhFatLfsQzsNzc\nuoJQYCm0RI503KRSSgpzeHlNLU2tyb2+lcIqjiL98BvOB3m8rkkV78BKhdASkRCPx2HR7Cq6fQEe\nemFLcmtJ6runsXgEVaIumpjJx7BAgSXS17yp5VSV5fPqun1sTOK/DYVVEkX6oZ2Mq/rG+z0VWCKp\nweNxOP/kSQDc8+cNHO7oTk4dSXnXNBfJB10kH9ZuuPS8AktExpUXsHhOFQ1NHdzz1AaCwWDCa1BY\nxVgsgsoNIdVXPOtRYImkhtPnjmNSVRGr323gqVd3JPz9FVYJFklQuZXbQjRRFFgioe7ASxfXUFyQ\nzaMvbeXlNXsT+v5ZCX23NDfUh9pgQRVNCGzZWR/xun1Nm1Q5rO2O1lNrrOYU3Lj9ILNqymLyWiIS\nP4X52Vx19nTuf24T9z69kaL8bBbMiM3nylCcZPQ9xkpDU4drio93UA03oAYTi/CK5SS4bg4sXVlY\nEqViVJ4TyXr19S1BgI4uH9trW+Jb1FH2Nhzm93/dTJAgn7l8Lgtmxi6wKiuL+91/dQMmwHCDasvO\n+t6feIjF62dq16BIJquuKOTK903FweHOP65NyPyBCqsYGO4xjaGCKpFGGlwKLJHMMrmqmA+dM40s\nr8PPHl8b92NY6gaMgcHCaqBW1UAf7tGERd2ezQM+N2b89IhfZyDD6SYcabegW7sC1Q0oiZIK3YB9\n1R44zMMvbqGjy88HzpzCJafV4DgR7UK/BuoGVFiNUCKDarBwisRIAiya4BpJYCmsJNOlWlgBHGju\n4OEXNtPc1s05C8bzkfNm4vEML7B0zCrBYhlUdXs2jzio+r7OcF4rmi7CkXQJuv28KxE5VnlJHh85\nz1A5Oo8XVu/hrj+updvnj+l7KKxGINpjVf19iA8WArEKqcFeO9rXT0RguY1aVSJDKy7I5polM5k4\npohVm+r5we/foi2GUzMprIYp2u6/aD68owmRxr2Wxr024tce6ftBZgaWiAwtN8fLh86ehpk4mk27\nmvjO/W/S2BKbS4vopOAk6u9Df6jQGCiYBgus0moTUT097x3Jsa0tO+tjdpKxiKSPLK+HS0+roeDN\n3ax+t4FbfrOK6z98AuPKC0f0uhpgMQyxaFVFE1QjbTn1FWlwQWShNVRgDWewhZsGWagLUBItFQdY\n9CcYDLJi/X5eXlNLYX4WX/jQfKZWlwy5nQZYJECsgyoWXXwjec1IugaH6hJUV6BIZnIch8VzxrJ0\n4STaOnzc9uBqtu9rHvbrKayiNNJJTaMJqniKZWCJiAxk3rRyLllcQ2e3nx88+Ba76lqH9To6ZhUj\nkbSqIgmqoULk4N6NEdVTVj1ryHV63muorsG6PZtjcpJxqlEXoEhsHDe5FL8/wJ9f28ltD67ma8tO\noXxUXlSvoZZVFGJ9qYhogurg3o0RB1W068e7FSciMndqOUtOHE9LWzd3PPYOXd3RnYelsIqB4bSq\nIg2qaENquNsrsI6kVpVI7J04s5Ljp5axY18L9y2P7jNHYRUHsQyqWBlp6OnYlYiMlOM4nHfyRMaW\nFfDK2n2sfjfyuVAVVhEaqAswHtMDjTRYhkutKxGJtyyvh4sWTcLjwP3PbqKzK7LuQA2wiLGRtqqG\nCqnG2kFO/h039DlUB/dujGjwxdESNcDCDedYqQtQJL4qRuVzyqwqXtuwn6df28EVZ04dchu1rCIQ\nq1bVSIKqsdYOGlR91xlqvUS22mJ5JWERSR+L51aRl+PlhdV78PkDQ66vsIqjWFxAMZLwSRduaFWJ\nSGLkZHmZO6WMlrZu3rBDX2lYYTWEaIarDzZbw3BaVSMJqUQFXLrND6guQJHEOWF6BQCr7NBf7BVW\nwzRUF+Bgrap4B1UsX0NEJF7KinMpzMti696hp2FSWMVILOfAS5eQieZ4lRu6ANWqEkksx3EYV15I\nY0snh1oHv5SIwmoQ8RpY0VcyhqhHa6CRgOnWBSgiiVeYFxqU3t7pG3Q9hVUcRNMFeMzzSWxVRXP5\nEBGRWHCc0BVBAoHBr/iksIqBaAZWZAp1AYpIJDq6Qi2qnGzvoOsprKIUyxkr4t0FONBJwtGcFJwJ\nXYAKKpHkqT3QRlF+NhVDzMKusBrAcGdYH0kXYDKpC1BEEq2xpZOmw11MrS7p7Q4ciMJqhNx6Jdxo\nWlUDBZVaVSISTys37Adg0eyqIddVWMVRso5XxbP7LxKaYklEhtJ0uIt3th2kqjSfhccprGIqHjOs\nx1q0QRVt959aVSIyUj5/gCf+sY1AIMilp9fg8QzeBQgKq37F43hVpCKZOX2g7WIVVIm8hH2yRgIq\nqESSIxgM8uwbu6g90MZpc8eyeM7YiLZTWLlQNIE1WEhBbIMqklZVKnQBKqhEkiMYDPLC6j28s/Ug\nk8cW8/ELzJADK3roelYjMNLBFWXVswYcvj7cFlbf1x5IvIJKRGQg3b4AT63YwaZdhxhXXsBnrzx+\nyHOr+lJYJVBptTlm+PpggTVcse72izSo3N6qUotKJDkOtXby5Cvb2XugDTNxNP/5weMpzMuO6jUU\nVi4Qi8AaaqRfvFtTwwmqRB2vUkiJJEcwGOStzQ28+NZeun0BFs2u4pMXHUd2VvRHoBRWI2AmlR7R\nFThtUuURgyzGjJ9+zPD1/lpX8F7YRBNaww2ontoGE++gijcFlEhyHWrt5JnXdrKzrpWC3CyWLTUs\nnjM24mNUR1NYJcFAgQXRnQs12OsPxC0hFY9WlQJKJPk6unysWLefVZvq8QeCnDCtnI8vnUVpce6I\nXldhFWf9ta5g8MAajqHOl3JLSMWDQkok+Xz+AG+928Ar6/bR0eWnrDiXfzp7GqfOrhp2a6ovhVUU\nZtWUDXli8NFdgTB4YPWINrgiPZk3mcek+hPLFpVCSiT5fP4A67c3smL9Pg61dpGf6+VDZ09jyUkT\nohrtNxQnGBz8GiKJZoxZCvwY8AJ3W2tvHWjdhqaOuBUfzYUX+xvCPtAJwomYgsmto/tiFVQKKUln\nFaPyImqG1Ne3BCHU7ba9tiW+RfWjq9vP21sO8PrGOlrbu8nyOpyzYAKXnl5DUX50I/36qqws7nf/\nXdWyMsZ4gTuB84DdwOvGmCesteuTW1n0+mthwXtBEsvQimTGiWQPP1dQiaSH9k4fqzbV8+amejq6\n/ORme7hg4UTOP2XSiI9LDcZVYQUsBDZba7cCGGMeBC4HXBNW/XUFHj0qsMdAgQXHBkwk4RXtNEjp\nElCgkBJJtpa2LlZurGPN5gN0+wMU5mdxxRlTeP9JE0bUkoqU28JqPLCrz+PdwKnJKMRMLh2wKzDa\nwIKh5w2M1Xx8yQyoeJ03paASOVZpaQFZWV7aO30cavfH7X3qG9t46a09vBUe3Vc+Ko8rz57O+adO\nJi83cRHitrBKaT0BMFhoQWwmvD36NSMRi4BK5MSzCimRgTU2tgGhY1aHDrXF/PX3HWxjxfr9bNp1\nCICxZQVcuGgSi+eMJcvroaW5nXgcKausLO53udvCag8wsc/jCeFlSRFt66p3uwFaWT0SNc+e286D\nioaCSiTxgsEgO+taeW39frbvC0XR5LHFXLxoMifOrIzoUh7x4raweh2YYYyZQiikrgauSWZBQwUW\n9D9CsG9QJOpqwsMNp2QHU18KKZHECwaDbN7TzIr1+6g9EGqlHTe5lIsWT2b25NKYnCc1Um4cun4R\ncDuhoev3WGtvHmjdeA5dP9pQ17iK5sKMIw2vkXbnuSmceiikREISOXQ9GAyyrbaFl9/Zy/6D7QAs\nmFHBRYsnM6161LBec6QGGrruurCKRiLDCiK7KKMbrybsxnACBZRIfxIVVjv3t/Dymlr2NBwGYOFx\nY7j0tBrGVxZF/VqxlBLnWbndYF2CPfoGQ7KCS+EkIgM52NzBc6t29x6TWjCjgivOnMrEMckNqaEo\nrKIUSWD1ODo0YhleyQ4kBY9Iaun2BVixfh8rN9ThDwSZXVPKlWdNY2p1SbJLi4jCahh6PqgjDa0e\nyQ6YwSh8RNLX1r3NPPvGLpoOd1FanMs1587gxJmVrhg4ESmF1QgMN7SSTcEkkhkCgSAvr6nltQ37\n8Xgclp46ictOryEvJ/U++lOvYhfq++HvpuBSKIlkrsMd3fzple3s3N9K5eh8rvvAXCZV9X/CbSpQ\nWMVYfwERzwBTIInI0RpbOnnwr+/S0tbN/OkV/Mslx1GQF//5++JJYZUAChQRSZSWtm4eemEzLW3d\nXHHGFC45vQZPCh2bGojCSkQkTbR3+nj4xc00He7i8jOmcNkZU5JdUsx4kl2AiIjExnOrdtPQ1MG5\nJ0/gstNrkl1OTCmsRETSwNa9zWzY0cjUcSVc/f4ZKTUsPRIKKxGRFBcIBHlu1S48Dnx8qUnq7Ojx\norASEUlx22qbOdTaxZknVKf08PTBKKxERFLcmq0HADh7/vgkVxI/CisRkRTW1e1ny95mJlQWMnls\neraqQGElIpLSdu5vJRAIctxk9849GgsKKxGRFLatthmAaeNTY/b04VJYiYiksLpDoSv8TkjyRRPj\nTWElIpLCDjZ3AlBWkpvkSuJLYSUiksKaWjvJz/Wm5GU/oqGwEhFJYe1dfgpTfEb1SCisRERSWGe3\nn4Lc9G5VATjBYDDZNYiIiAxKLSsREXE9hZWIiLiewkpERFxPYSUiIq6nsBIREddTWImIiOsprERE\nxPXS/0yyGDPGLAV+DHiBu621tya5pLgwxmwHWgA/4LPWnmyMKQN+D9QA24GrrLWNSSpxxIwx9wCX\nAHXW2rnhZQPuozHmBuBaQr+Tz1lrlyeh7BEZYJ+/AXwaqA+vdqO19s/h59JhnycC9wFVQBD4hbX2\nx+n+t043allFwRjjBe4ELgRmA/9sjJmd3Kri6hxr7Xxr7cnhx18BnrfWzgCeDz9OZb8Glh61rN99\nDP+drwbmhLf5afj/h1Tza47dZ4Afhf/W8/sEVbrssw/4orV2NrAIuC68b+n+t04rCqvoLAQ2W2u3\nWmu7gAeBy5NcUyJdDtwbvn8vcEUSaxkxa+1LwMGjFg+0j5cDD1prO62124DNhP5/SCkD7PNA0mWf\na621b4bvtwAbgPGk+d863SisojMe2NXn8e7wsnQUBJ4zxqwyxvxreFmVtbY2fH8foW6VdDPQPqb7\n3/6zxpg1xph7jDGl4WVpt8/GmBpgAfAamfu3TkkKKxnIGdba+YS6PK8zxpzV90lrbZBQoKWtTNjH\nsLuAqcB8oBb4QXLLiQ9jTBHwCPB5a21z3+cy6G+dshRW0dkDTOzzeEJ4Wdqx1u4J39YBjxHqBtlv\njBkHEL6tS16FcTPQPqbt395au99a67fWBoD/470ur7TZZ2NMNqGgut9a+2h4ccb9rVOZwio6rwMz\njDFTjDE5hA7CPpHkmmLOGFNojCnuuQ+cD6wltK/LwqstAx5PToVxNdA+PgFcbYzJNcZMAWYAK5NQ\nX8z1fGCHfYDQ3xrSZJ+NMQ7wS2CDtfaHfZ7KuL91KtMlQqJkjLkIuJ3Q0PV7rLU3J7mkmDPGTCXU\nmoLQ6Q2/s9bebIwpBx4CJgE7CA31jfRgvesYYx4AzgYqgP3A14E/MsA+GmO+CnyK0Oiyz1trn05C\n2SMywD6fTagLMEhoCPe/9RzLSZN9PgN4GXgHCIQX30jouFXa/q3TjcJKRERcT92AIiLiegorERFx\nPYWViIi4nsJKRERcT2ElIiKup7CStGeM2W6MmTvEOi8aYy4J3/+WMebDEbzuN4wxt8WqThEZmC4R\nInIUa+3Xkl2DiBxJYSUZwxjzIqFZSBYD1cBD1tpjLnNijPk18Ia19o7wtZ4MMIrQ/HlbgA9Za9uO\n2uZ44H7gs4Rm9f4d702M+py19gtx2CWRjKFuQMk0k4CzCM28/S/GmBkRbHMycA1wHJANfKTvk8aY\ncwmF09XW2r+Fn99irT3eWns88K0Y1i+SkdSykkzzcHjC1iZjzAZgGvDuENsst9YeAjDGvBbepsf5\nhC7Qd761dm942QrgC8aY7wN/A3SVWZERUstKMk1Hn/t+IvvCNtg2mwjNE9lzNWWsta8SarmtAj4G\nvDDcYkUkRGElMjLbCbWuvtMzgjA8U3eztfZB4HrgJGOM/q2JjIC6AUVGyFq7yxizBFhujMkHHOB6\nY4yf0BfCfw93PYrIMGnWdRERcT11TYiIiOsprERExPUUViIi4noKKxERcT2FlYiIuJ7CSkREXE9h\nJSIirvf/AzH3nJqEiqGHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c2eb7f710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.jointplot(x=\"Inlinks\", y=\"Outlinks\", data=results_df, kind = \"kde\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Identify implicit clusters using Clustering algos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Typically, number of clusters are chosen with a plot of within-cluster sum-of-squares-of-dstances vs number of clusters. \n",
    "Here for simplicity, we use a simple heuristic to fix the number of clusters in advnace."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10\n"
     ]
    }
   ],
   "source": [
    "n_clusters = min( int( round(np.sqrt(len(index)/2)) ), 10 ) # not more than 10 clusters\n",
    "print(n_clusters)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3a. K-Means clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "K Means is a point-assignment based clustering-algorithm: here we start with k points chosen randomly as centroids,  assign the remaining points to k centroids by using certain distance measure (Euclidean / Cosine / Jaccardi). Then we compute new centroids, re-assign remaining points and repeat, until there is convergence of centroids. \n",
    "\n",
    "Here we are more interested in clustering inherent in who-follows-me (network-driven) graph, rather than who-do-I-follow (self-driven). So the approach is to represent the nodes as Observations and whether other nodes follow them or not (1 or 0) as Features (One observation per row, and one feature per column) Thus the input matrix must be such that any value in cell implies whether node in the column follows node in the row."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0 : 45 Ids</th>\n",
       "      <th>1 : 67 Ids</th>\n",
       "      <th>2 : 39 Ids</th>\n",
       "      <th>3 : 28 Ids</th>\n",
       "      <th>4 : 21 Ids</th>\n",
       "      <th>5 : 17 Ids</th>\n",
       "      <th>6 : 15 Ids</th>\n",
       "      <th>7 : 24 Ids</th>\n",
       "      <th>8 : 18 Ids</th>\n",
       "      <th>9 : 103 Ids</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Edward Snowden</td>\n",
       "      <td>Tim Urban</td>\n",
       "      <td>John D. Cook</td>\n",
       "      <td>Hilary Mason</td>\n",
       "      <td>OpenAI</td>\n",
       "      <td>Elon Musk</td>\n",
       "      <td>WIRED</td>\n",
       "      <td>Mike Bostock</td>\n",
       "      <td>NassimNicholasTaleb</td>\n",
       "      <td>Charlie Rose Show</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Nate Silver</td>\n",
       "      <td>Tim Ferriss</td>\n",
       "      <td>Lynn Cherny</td>\n",
       "      <td>dj patil</td>\n",
       "      <td>Microsoft Research</td>\n",
       "      <td>Bill Gates</td>\n",
       "      <td>Eric Schmidt</td>\n",
       "      <td>Gilad Lotan</td>\n",
       "      <td>Neil deGrasse Tyson</td>\n",
       "      <td>Irene Au</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Sundar Pichai</td>\n",
       "      <td>GV</td>\n",
       "      <td>Ben Hamner</td>\n",
       "      <td>Jeff Hammerbacher</td>\n",
       "      <td>Andrew Ng</td>\n",
       "      <td>Marc Andreessen</td>\n",
       "      <td>TechCrunch</td>\n",
       "      <td>Jeffrey Heer</td>\n",
       "      <td>Max Roser</td>\n",
       "      <td>Steven Brooks</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Patrick Collison</td>\n",
       "      <td>Mike Bloomberg</td>\n",
       "      <td>David Smith</td>\n",
       "      <td>Drew Conway</td>\n",
       "      <td>Yann LeCun</td>\n",
       "      <td>Tim O'Reilly</td>\n",
       "      <td>Steven Levy</td>\n",
       "      <td>Martin Wattenberg</td>\n",
       "      <td>Steven Pinker</td>\n",
       "      <td>Udacity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Nick Bilton</td>\n",
       "      <td>Garry Tan</td>\n",
       "      <td>Fernando Perez</td>\n",
       "      <td>Wes McKinney</td>\n",
       "      <td>Google Research</td>\n",
       "      <td>Chris Dixon</td>\n",
       "      <td>Chris Anderson</td>\n",
       "      <td>Andrew Gelman</td>\n",
       "      <td>Glenn Greenwald</td>\n",
       "      <td>briankrebs</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Benedict Evans</td>\n",
       "      <td>Joe Gebbia</td>\n",
       "      <td>Andreas Mueller</td>\n",
       "      <td>Monica Rogati</td>\n",
       "      <td>Sebastian Thrun</td>\n",
       "      <td>Fred Wilson</td>\n",
       "      <td>Steve Case</td>\n",
       "      <td>Fernanda Viégas</td>\n",
       "      <td>Atul Gawande</td>\n",
       "      <td>Christopher Soghoian</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Brad Stone</td>\n",
       "      <td>Yves Behar</td>\n",
       "      <td>Kaggle</td>\n",
       "      <td>Hadley Wickham</td>\n",
       "      <td>Fei-Fei Li</td>\n",
       "      <td>Ev Williams</td>\n",
       "      <td>Joi Ito</td>\n",
       "      <td>Steven Strogatz</td>\n",
       "      <td>EFF</td>\n",
       "      <td>IDEO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Josh Kopelman</td>\n",
       "      <td>Ashlee Vance</td>\n",
       "      <td>John Foreman</td>\n",
       "      <td>Sean J. Taylor</td>\n",
       "      <td>Andrej Karpathy</td>\n",
       "      <td>Reid Hoffman</td>\n",
       "      <td>Lessig</td>\n",
       "      <td>Duncan Watts</td>\n",
       "      <td>Kevin Kelly</td>\n",
       "      <td>Stanford University</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>DAVE MORIN</td>\n",
       "      <td>Dustin Moskovitz</td>\n",
       "      <td>Jake VanderPlas</td>\n",
       "      <td>Nathan Yau</td>\n",
       "      <td>Shivon Zilis</td>\n",
       "      <td>Bill Gurley</td>\n",
       "      <td>John Markoff</td>\n",
       "      <td>Kenneth Cukier</td>\n",
       "      <td>Tim Berners-Lee</td>\n",
       "      <td>Gary Marcus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Hunter Walk</td>\n",
       "      <td>Balaji S. Srinivasan</td>\n",
       "      <td>Scientific Python</td>\n",
       "      <td>Peter Skomoroch</td>\n",
       "      <td>Pedro Domingos</td>\n",
       "      <td>Paul Graham</td>\n",
       "      <td>Bill Gross</td>\n",
       "      <td>Sinan Aral</td>\n",
       "      <td>Edward Tufte</td>\n",
       "      <td>Matías Duarte</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         0 : 45 Ids            1 : 67 Ids         2 : 39 Ids  \\\n",
       "0    Edward Snowden             Tim Urban       John D. Cook   \n",
       "1       Nate Silver           Tim Ferriss        Lynn Cherny   \n",
       "2     Sundar Pichai                    GV         Ben Hamner   \n",
       "3  Patrick Collison        Mike Bloomberg        David Smith   \n",
       "4       Nick Bilton             Garry Tan     Fernando Perez   \n",
       "5    Benedict Evans            Joe Gebbia    Andreas Mueller   \n",
       "6        Brad Stone            Yves Behar             Kaggle   \n",
       "7     Josh Kopelman          Ashlee Vance       John Foreman   \n",
       "8        DAVE MORIN      Dustin Moskovitz    Jake VanderPlas   \n",
       "9       Hunter Walk  Balaji S. Srinivasan  Scientific Python   \n",
       "\n",
       "          3 : 28 Ids          4 : 21 Ids       5 : 17 Ids      6 : 15 Ids  \\\n",
       "0       Hilary Mason              OpenAI        Elon Musk           WIRED   \n",
       "1           dj patil  Microsoft Research       Bill Gates    Eric Schmidt   \n",
       "2  Jeff Hammerbacher           Andrew Ng  Marc Andreessen      TechCrunch   \n",
       "3        Drew Conway          Yann LeCun     Tim O'Reilly     Steven Levy   \n",
       "4       Wes McKinney     Google Research      Chris Dixon  Chris Anderson   \n",
       "5      Monica Rogati     Sebastian Thrun      Fred Wilson      Steve Case   \n",
       "6     Hadley Wickham          Fei-Fei Li      Ev Williams         Joi Ito   \n",
       "7     Sean J. Taylor     Andrej Karpathy     Reid Hoffman          Lessig   \n",
       "8         Nathan Yau        Shivon Zilis      Bill Gurley    John Markoff   \n",
       "9    Peter Skomoroch      Pedro Domingos      Paul Graham      Bill Gross   \n",
       "\n",
       "          7 : 24 Ids           8 : 18 Ids           9 : 103 Ids  \n",
       "0       Mike Bostock  NassimNicholasTaleb     Charlie Rose Show  \n",
       "1        Gilad Lotan  Neil deGrasse Tyson              Irene Au  \n",
       "2       Jeffrey Heer            Max Roser         Steven Brooks  \n",
       "3  Martin Wattenberg        Steven Pinker               Udacity  \n",
       "4      Andrew Gelman      Glenn Greenwald            briankrebs  \n",
       "5    Fernanda Viégas         Atul Gawande  Christopher Soghoian  \n",
       "6    Steven Strogatz                  EFF                  IDEO  \n",
       "7       Duncan Watts          Kevin Kelly   Stanford University  \n",
       "8     Kenneth Cukier      Tim Berners-Lee           Gary Marcus  \n",
       "9         Sinan Aral         Edward Tufte         Matías Duarte  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "est = KMeans(max_iter = 100000, n_clusters = n_clusters, n_init = 200, init='k-means++')  \n",
    "results_df['kmeans'] = est.fit_predict(binaryMap_cfr)\n",
    "\n",
    "top_nodes_in_cluster(results_df.sort_values( by = 'PageRank', ascending =False), \n",
    "                     'kmeans', n_clusters)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3b. Spectral clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One problem in using K-Means algorithm to cluster the given social-graph is 'Curse of Dimensionality' i.e. at higher-dimensions (here ~400 dimensions), metric like 'Euclidean distance' or 'Centre' would have little meaning in the context of non-convex adjacency matrix.\n",
    "\n",
    "On the other hand, spectral clustering attempts to partition the graph such that number of edges which connect different components are minimized. Below is the output from Spectral Clustering."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0 : 34 Ids</th>\n",
       "      <th>2 : 19 Ids</th>\n",
       "      <th>3 : 27 Ids</th>\n",
       "      <th>4 : 55 Ids</th>\n",
       "      <th>5 : 11 Ids</th>\n",
       "      <th>6 : 10 Ids</th>\n",
       "      <th>7 : 187 Ids</th>\n",
       "      <th>8 : 14 Ids</th>\n",
       "      <th>9 : 13 Ids</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>dj patil</td>\n",
       "      <td>Lynn Cherny</td>\n",
       "      <td>Tim O'Reilly</td>\n",
       "      <td>Nate Silver</td>\n",
       "      <td>Hilary Mason</td>\n",
       "      <td>Elon Musk</td>\n",
       "      <td>Edward Snowden</td>\n",
       "      <td>Andrew Ng</td>\n",
       "      <td>Luke Wroblewski</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Jeff Hammerbacher</td>\n",
       "      <td>Ben Hamner</td>\n",
       "      <td>Ev Williams</td>\n",
       "      <td>Sundar Pichai</td>\n",
       "      <td>Drew Conway</td>\n",
       "      <td>Bill Gates</td>\n",
       "      <td>WIRED</td>\n",
       "      <td>Yann LeCun</td>\n",
       "      <td>Matías Duarte</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Monica Rogati</td>\n",
       "      <td>Olivier Grisel</td>\n",
       "      <td>Bill Gurley</td>\n",
       "      <td>Patrick Collison</td>\n",
       "      <td>Wes McKinney</td>\n",
       "      <td>Marc Andreessen</td>\n",
       "      <td>Eric Schmidt</td>\n",
       "      <td>Fei-Fei Li</td>\n",
       "      <td>Julie Zhuo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Peter Skomoroch</td>\n",
       "      <td>Fernando Perez</td>\n",
       "      <td>Om Malik</td>\n",
       "      <td>Benedict Evans</td>\n",
       "      <td>Hadley Wickham</td>\n",
       "      <td>Chris Dixon</td>\n",
       "      <td>Vinod Khosla</td>\n",
       "      <td>Andrej Karpathy</td>\n",
       "      <td>Jon Wiley</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Michael E. Driscoll</td>\n",
       "      <td>Andreas Mueller</td>\n",
       "      <td>Kara Swisher</td>\n",
       "      <td>Y Combinator</td>\n",
       "      <td>Sean J. Taylor</td>\n",
       "      <td>Fred Wilson</td>\n",
       "      <td>TechCrunch</td>\n",
       "      <td>Stanford NLP Group</td>\n",
       "      <td>Daniel Burka</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>bradford cross</td>\n",
       "      <td>Jake VanderPlas</td>\n",
       "      <td>Chris Sacca</td>\n",
       "      <td>Emily Chang</td>\n",
       "      <td>Nathan Yau</td>\n",
       "      <td>Reid Hoffman</td>\n",
       "      <td>Chris Anderson</td>\n",
       "      <td>John Platt</td>\n",
       "      <td>Braden Kowitz</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Jeremy Howard</td>\n",
       "      <td>Scientific Python</td>\n",
       "      <td>Steven Levy</td>\n",
       "      <td>Ben Thompson</td>\n",
       "      <td>John Myles White</td>\n",
       "      <td>Paul Graham</td>\n",
       "      <td>OpenAI</td>\n",
       "      <td>Richard</td>\n",
       "      <td>Google Design</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Mike Olson</td>\n",
       "      <td>Sebastian Raschka</td>\n",
       "      <td>Nick Bilton</td>\n",
       "      <td>Brian Chesky</td>\n",
       "      <td>chris wiggins</td>\n",
       "      <td>Ben Horowitz</td>\n",
       "      <td>Steve Case</td>\n",
       "      <td>Oren Etzioni</td>\n",
       "      <td>Josh Brewer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Anthony Goldbloom</td>\n",
       "      <td>Trey Causey</td>\n",
       "      <td>Josh Kopelman</td>\n",
       "      <td>🍪Steven Sinofsky ॐ</td>\n",
       "      <td>Ben Lorica 罗瑞卡</td>\n",
       "      <td>Sam Altman</td>\n",
       "      <td>Tim Urban</td>\n",
       "      <td>Ian Goodfellow</td>\n",
       "      <td>Jake Knapp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Joe Hellerstein</td>\n",
       "      <td>Joshua Bloom</td>\n",
       "      <td>DAVE MORIN</td>\n",
       "      <td>Tony Fadell</td>\n",
       "      <td>Josh Wills</td>\n",
       "      <td>Max Levchin</td>\n",
       "      <td>Joi Ito</td>\n",
       "      <td>Nando de Freitas</td>\n",
       "      <td>John Zeratsky</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            0 : 34 Ids         2 : 19 Ids     3 : 27 Ids          4 : 55 Ids  \\\n",
       "0             dj patil        Lynn Cherny   Tim O'Reilly         Nate Silver   \n",
       "1    Jeff Hammerbacher         Ben Hamner    Ev Williams       Sundar Pichai   \n",
       "2        Monica Rogati     Olivier Grisel    Bill Gurley    Patrick Collison   \n",
       "3      Peter Skomoroch     Fernando Perez       Om Malik      Benedict Evans   \n",
       "4  Michael E. Driscoll    Andreas Mueller   Kara Swisher        Y Combinator   \n",
       "5       bradford cross    Jake VanderPlas    Chris Sacca         Emily Chang   \n",
       "6        Jeremy Howard  Scientific Python    Steven Levy        Ben Thompson   \n",
       "7           Mike Olson  Sebastian Raschka    Nick Bilton        Brian Chesky   \n",
       "8    Anthony Goldbloom        Trey Causey  Josh Kopelman  🍪Steven Sinofsky ॐ   \n",
       "9      Joe Hellerstein       Joshua Bloom     DAVE MORIN         Tony Fadell   \n",
       "\n",
       "         5 : 11 Ids       6 : 10 Ids     7 : 187 Ids          8 : 14 Ids  \\\n",
       "0      Hilary Mason        Elon Musk  Edward Snowden           Andrew Ng   \n",
       "1       Drew Conway       Bill Gates           WIRED          Yann LeCun   \n",
       "2      Wes McKinney  Marc Andreessen    Eric Schmidt          Fei-Fei Li   \n",
       "3    Hadley Wickham      Chris Dixon    Vinod Khosla     Andrej Karpathy   \n",
       "4    Sean J. Taylor      Fred Wilson      TechCrunch  Stanford NLP Group   \n",
       "5        Nathan Yau     Reid Hoffman  Chris Anderson          John Platt   \n",
       "6  John Myles White      Paul Graham          OpenAI             Richard   \n",
       "7     chris wiggins     Ben Horowitz      Steve Case        Oren Etzioni   \n",
       "8    Ben Lorica 罗瑞卡       Sam Altman       Tim Urban      Ian Goodfellow   \n",
       "9        Josh Wills      Max Levchin         Joi Ito    Nando de Freitas   \n",
       "\n",
       "        9 : 13 Ids  \n",
       "0  Luke Wroblewski  \n",
       "1    Matías Duarte  \n",
       "2       Julie Zhuo  \n",
       "3        Jon Wiley  \n",
       "4     Daniel Burka  \n",
       "5    Braden Kowitz  \n",
       "6    Google Design  \n",
       "7      Josh Brewer  \n",
       "8       Jake Knapp  \n",
       "9    John Zeratsky  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import cluster\n",
    "\n",
    "spectral = cluster.SpectralClustering(n_clusters=n_clusters,  n_init = 500,\n",
    "                                          eigen_solver='arpack',\n",
    "                                          affinity=\"nearest_neighbors\")\n",
    "spectral.fit(binaryMap_cfr)\n",
    "results_df['spectral'] = spectral.labels_.astype(np.int)\n",
    "\n",
    "top_nodes_in_cluster(results_df.sort_values( by = 'PageRank', ascending =False), \n",
    "                     'spectral', n_clusters)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Nodes mainly into Deep Learning community have grouped into Cluster 8, Python Machine Learning community into Cluster 2, Design community into Cluster 9, general Data Science community into Cluster 0 and 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One smaller clusters (1) wasn't shown above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>PageRank</th>\n",
       "      <th>UserName</th>\n",
       "      <th>Inlinks</th>\n",
       "      <th>Outlinks</th>\n",
       "      <th>Followers</th>\n",
       "      <th>Friends</th>\n",
       "      <th>Location</th>\n",
       "      <th>Created</th>\n",
       "      <th>kmeans</th>\n",
       "      <th>spectral</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FullName</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>JAMA</th>\n",
       "      <td>0.000658</td>\n",
       "      <td>JAMA_current</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>205154</td>\n",
       "      <td>559</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>May-2009</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FSI Stanford</th>\n",
       "      <td>0.000643</td>\n",
       "      <td>FSIStanford</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>6614</td>\n",
       "      <td>285</td>\n",
       "      <td>Stanford, CA</td>\n",
       "      <td>Dec-2010</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Adam Johnson</th>\n",
       "      <td>0.000638</td>\n",
       "      <td>AJInsight</td>\n",
       "      <td>4</td>\n",
       "      <td>26</td>\n",
       "      <td>27945</td>\n",
       "      <td>1232</td>\n",
       "      <td>New York</td>\n",
       "      <td>Jul-2010</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PNAS</th>\n",
       "      <td>0.000595</td>\n",
       "      <td>PNASNews</td>\n",
       "      <td>4</td>\n",
       "      <td>17</td>\n",
       "      <td>66720</td>\n",
       "      <td>1110</td>\n",
       "      <td>Washington, DC</td>\n",
       "      <td>Feb-2011</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IQSS @ Harvard</th>\n",
       "      <td>0.000558</td>\n",
       "      <td>IQSS</td>\n",
       "      <td>5</td>\n",
       "      <td>4</td>\n",
       "      <td>4241</td>\n",
       "      <td>360</td>\n",
       "      <td>Cambridge, MA</td>\n",
       "      <td>May-2009</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                PageRank      UserName  Inlinks  Outlinks  Followers  Friends  \\\n",
       "FullName                                                                        \n",
       "JAMA            0.000658  JAMA_current        4         3     205154      559   \n",
       "FSI Stanford    0.000643   FSIStanford        5         5       6614      285   \n",
       "Adam Johnson    0.000638     AJInsight        4        26      27945     1232   \n",
       "PNAS            0.000595      PNASNews        4        17      66720     1110   \n",
       "IQSS @ Harvard  0.000558          IQSS        5         4       4241      360   \n",
       "\n",
       "                      Location   Created  kmeans  spectral  \n",
       "FullName                                                    \n",
       "JAMA               Chicago, IL  May-2009       9         1  \n",
       "FSI Stanford      Stanford, CA  Dec-2010       9         1  \n",
       "Adam Johnson          New York  Jul-2010       9         1  \n",
       "PNAS            Washington, DC  Feb-2011       9         1  \n",
       "IQSS @ Harvard   Cambridge, MA  May-2009       9         1  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df [results_df['spectral'].isin([1])].sort_values(by = 'PageRank', ascending =False).set_index('FullName').head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3c. Affinity Propagation clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Unlike K-Means or Spectral clustering, affinity propagation doesn't require the number of clusters to be estimated beforehand. Here the algorithm finds 'exemplars' i.e. members of dataset that are representative of clusters. and tries to find clusters around them.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "128 affinity clusters.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2 : 87 Ids</th>\n",
       "      <th>6 : 34 Ids</th>\n",
       "      <th>32 : 23 Ids</th>\n",
       "      <th>38 : 11 Ids</th>\n",
       "      <th>41 : 11 Ids</th>\n",
       "      <th>111 : 42 Ids</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Tim Urban</td>\n",
       "      <td>Adam Grant</td>\n",
       "      <td>Lynn Cherny</td>\n",
       "      <td>John D. Cook</td>\n",
       "      <td>Jeff Jordan</td>\n",
       "      <td>Yves Behar</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Charlie Rose Show</td>\n",
       "      <td>Steven Strogatz</td>\n",
       "      <td>Ben Hamner</td>\n",
       "      <td>chris wiggins</td>\n",
       "      <td>Dustin Moskovitz</td>\n",
       "      <td>Ashlee Vance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Irene Au</td>\n",
       "      <td>Tim Harford</td>\n",
       "      <td>Olivier Grisel</td>\n",
       "      <td>Jeremy Howard</td>\n",
       "      <td>Balaji S. Srinivasan</td>\n",
       "      <td>WikiLeaks</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>John Carmack</td>\n",
       "      <td>Gary Marcus</td>\n",
       "      <td>Fernando Perez</td>\n",
       "      <td>Xavier Amatriain</td>\n",
       "      <td>Horace Dediu</td>\n",
       "      <td>Seth Godin</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>The Information</td>\n",
       "      <td>DHH</td>\n",
       "      <td>Andreas Mueller</td>\n",
       "      <td>Anthony Goldbloom</td>\n",
       "      <td>Jessica Livingston</td>\n",
       "      <td>Tom Hulme</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Paul Ford</td>\n",
       "      <td>BJ Fogg</td>\n",
       "      <td>Jake VanderPlas</td>\n",
       "      <td>Kaggle</td>\n",
       "      <td>David Marcus</td>\n",
       "      <td>Udacity</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>briankrebs</td>\n",
       "      <td>Joe Blitzstein</td>\n",
       "      <td>Scientific Python</td>\n",
       "      <td>John Foreman</td>\n",
       "      <td>Frank Chen</td>\n",
       "      <td>IDEO</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Christopher Soghoian</td>\n",
       "      <td>Brian Nosek</td>\n",
       "      <td>Sebastian Raschka</td>\n",
       "      <td>Edwin Chen</td>\n",
       "      <td>Kleiner Perkins</td>\n",
       "      <td>Scott Dadich</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>John Hagel</td>\n",
       "      <td>Martin Varsavsky</td>\n",
       "      <td>Trey Causey</td>\n",
       "      <td>Mark Faridani</td>\n",
       "      <td>Amir Efrati</td>\n",
       "      <td>ReadWrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Sarah Frier</td>\n",
       "      <td>Russell Roberts</td>\n",
       "      <td>Joshua Bloom</td>\n",
       "      <td>Shane Conway</td>\n",
       "      <td>Scott Kupor</td>\n",
       "      <td>Peter Diamandis</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             2 : 87 Ids        6 : 34 Ids        32 : 23 Ids  \\\n",
       "0             Tim Urban        Adam Grant        Lynn Cherny   \n",
       "1     Charlie Rose Show   Steven Strogatz         Ben Hamner   \n",
       "2              Irene Au       Tim Harford     Olivier Grisel   \n",
       "3          John Carmack       Gary Marcus     Fernando Perez   \n",
       "4       The Information               DHH    Andreas Mueller   \n",
       "5             Paul Ford           BJ Fogg    Jake VanderPlas   \n",
       "6            briankrebs    Joe Blitzstein  Scientific Python   \n",
       "7  Christopher Soghoian       Brian Nosek  Sebastian Raschka   \n",
       "8            John Hagel  Martin Varsavsky        Trey Causey   \n",
       "9           Sarah Frier   Russell Roberts       Joshua Bloom   \n",
       "\n",
       "         38 : 11 Ids           41 : 11 Ids     111 : 42 Ids  \n",
       "0       John D. Cook           Jeff Jordan       Yves Behar  \n",
       "1      chris wiggins      Dustin Moskovitz     Ashlee Vance  \n",
       "2      Jeremy Howard  Balaji S. Srinivasan        WikiLeaks  \n",
       "3   Xavier Amatriain          Horace Dediu       Seth Godin  \n",
       "4  Anthony Goldbloom    Jessica Livingston        Tom Hulme  \n",
       "5             Kaggle          David Marcus          Udacity  \n",
       "6       John Foreman            Frank Chen             IDEO  \n",
       "7         Edwin Chen       Kleiner Perkins     Scott Dadich  \n",
       "8      Mark Faridani           Amir Efrati        ReadWrite  \n",
       "9       Shane Conway           Scott Kupor  Peter Diamandis  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cluster import AffinityPropagation\n",
    "\n",
    "af = AffinityPropagation(preference=-50).fit(binaryMap_cfr)\n",
    "results_df['affinity'] = af.labels_\n",
    "n_clusters_affinity = len(af.cluster_centers_indices_)\n",
    "\n",
    "print(str(n_clusters_affinity) + \" affinity clusters.\")\n",
    "\n",
    "top_nodes_in_cluster(results_df.sort_values( by = 'PageRank', ascending =False), \n",
    "                     'affinity', n_clusters_affinity)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3d. Principal Component Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To handle the 'curse of dimensionality' problem inherent in high-dimension data, PCA is generally used to represent the high-dimensional data in a fewer number of dimensions - this helps in better visualizing of data as well as faster computation while running clustering algorithm."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Explained-variance and Proportion of Explained-variance in 3 dimensions [dim1 dim2 dim3]\n",
      "[ 7.09266174  3.16212677  1.21891396] [ 0.18363161  0.08186862  0.03155813]\n"
     ]
    }
   ],
   "source": [
    "pca = PCA(n_components=3)\n",
    "Xproj = pca.fit_transform(binaryMap_cfr)\n",
    "\n",
    "results_df['dim1'] = Xproj[:,0]\n",
    "results_df['dim2'] = Xproj[:,1]\n",
    "results_df['dim3'] = Xproj[:,2]\n",
    "\n",
    "results_df = results_df.sort_values( by = 'PageRank', ascending =False)\n",
    "results_df.to_csv(cluster_filename)\n",
    "\n",
    "print(\"Explained-variance and Proportion of Explained-variance in 3 dimensions [dim1 dim2 dim3]\")\n",
    "print(pca.explained_variance_, pca.explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A simpler visualization of Spectral-clustering outcome as rendered in 2 dimensions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHMAAAKBCAYAAADUX+dfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAIAKUeYAAAAAVMg/AZ3IRECCva3cAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c2eb205f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Spectral clustering | Plot the ego-network in 2 dimensions\n",
    "\n",
    "plt.figure(num=None, figsize=(20, 10), dpi=80, facecolor='w', edgecolor='k')\n",
    "plt.scatter(results_df['dim1'], results_df['dim2'], s = 100  ,c= results_df['spectral'], \n",
    "            alpha=0.5, cmap=plt.cm.get_cmap('nipy_spectral', 10))\n",
    "plt.colorbar();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "More on Clustering Algorithms:  http://scikit-learn.org/stable/modules/clustering.html"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## 4. Recommend new friends to follow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have PageRank in place for nodes in social-graph, we can ask for recommendations on the basis of top-ranked nodes in the graph. E.g. To get 20 recommendations, after looking at friends of top PageRank scoring nodes in my network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4a. Reco: After looking at top nodes in full ego-network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "'Who-To-Follow' reco after looking at 187 friends' friends:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Freq</th>\n",
       "      <th>Created</th>\n",
       "      <th>Followers</th>\n",
       "      <th>Friends</th>\n",
       "      <th>Location</th>\n",
       "      <th>UserName</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FullName</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Medium</th>\n",
       "      <td>67</td>\n",
       "      <td>May-2012</td>\n",
       "      <td>2188119</td>\n",
       "      <td>78</td>\n",
       "      <td>San Francisco, CA, US</td>\n",
       "      <td>Medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Barack Obama</th>\n",
       "      <td>67</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>88892730</td>\n",
       "      <td>629775</td>\n",
       "      <td>Washington, DC</td>\n",
       "      <td>BarackObama</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Aaron Levie</th>\n",
       "      <td>52</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>1626688</td>\n",
       "      <td>423</td>\n",
       "      <td>Palo Alto</td>\n",
       "      <td>levie</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Clay Shirky</th>\n",
       "      <td>51</td>\n",
       "      <td>May-2007</td>\n",
       "      <td>365919</td>\n",
       "      <td>749</td>\n",
       "      <td>Shanghai</td>\n",
       "      <td>cshirky</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mitch Kapor</th>\n",
       "      <td>47</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>116400</td>\n",
       "      <td>553</td>\n",
       "      <td>Oakland</td>\n",
       "      <td>mkapor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The New York Times</th>\n",
       "      <td>45</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>37377094</td>\n",
       "      <td>894</td>\n",
       "      <td>New York City</td>\n",
       "      <td>nytimes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jack</th>\n",
       "      <td>45</td>\n",
       "      <td>Mar-2006</td>\n",
       "      <td>4055017</td>\n",
       "      <td>2791</td>\n",
       "      <td>California, USA</td>\n",
       "      <td>jack</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>marissamayer</th>\n",
       "      <td>44</td>\n",
       "      <td>Nov-2008</td>\n",
       "      <td>1705305</td>\n",
       "      <td>350</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>marissamayer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Stewart Butterfield</th>\n",
       "      <td>43</td>\n",
       "      <td>Sep-2006</td>\n",
       "      <td>65193</td>\n",
       "      <td>2418</td>\n",
       "      <td>West coast</td>\n",
       "      <td>stewart</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Anil Dash</th>\n",
       "      <td>40</td>\n",
       "      <td>Dec-2006</td>\n",
       "      <td>613133</td>\n",
       "      <td>4684</td>\n",
       "      <td>NYC</td>\n",
       "      <td>anildash</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dick costolo</th>\n",
       "      <td>40</td>\n",
       "      <td>May-2007</td>\n",
       "      <td>1739649</td>\n",
       "      <td>681</td>\n",
       "      <td>San Francisco</td>\n",
       "      <td>dickc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chris Anderson</th>\n",
       "      <td>40</td>\n",
       "      <td>Nov-2008</td>\n",
       "      <td>184874</td>\n",
       "      <td>977</td>\n",
       "      <td>Berkeley</td>\n",
       "      <td>chr1sa</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>The Economist</th>\n",
       "      <td>39</td>\n",
       "      <td>May-2007</td>\n",
       "      <td>20816240</td>\n",
       "      <td>159</td>\n",
       "      <td>London</td>\n",
       "      <td>TheEconomist</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>President Obama</th>\n",
       "      <td>39</td>\n",
       "      <td>Jun-2013</td>\n",
       "      <td>15459612</td>\n",
       "      <td>79</td>\n",
       "      <td>Washington, D.C.</td>\n",
       "      <td>POTUS44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Farhad Manjoo 🐣</th>\n",
       "      <td>39</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>158093</td>\n",
       "      <td>3674</td>\n",
       "      <td>I'm right over here (it's not NYC)</td>\n",
       "      <td>fmanjoo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>John Maeda</th>\n",
       "      <td>38</td>\n",
       "      <td>Jul-2008</td>\n",
       "      <td>439273</td>\n",
       "      <td>21724</td>\n",
       "      <td>WordPressLandia</td>\n",
       "      <td>johnmaeda</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>David Pogue</th>\n",
       "      <td>38</td>\n",
       "      <td>Oct-2007</td>\n",
       "      <td>1564476</td>\n",
       "      <td>1794</td>\n",
       "      <td>USA</td>\n",
       "      <td>Pogue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hanna Wallach</th>\n",
       "      <td>37</td>\n",
       "      <td>Sep-2012</td>\n",
       "      <td>7537</td>\n",
       "      <td>795</td>\n",
       "      <td>Brooklyn, NY</td>\n",
       "      <td>hannawallach</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tim Cook</th>\n",
       "      <td>37</td>\n",
       "      <td>Jul-2013</td>\n",
       "      <td>4925005</td>\n",
       "      <td>56</td>\n",
       "      <td>Cupertino</td>\n",
       "      <td>tim_cook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hillary Clinton</th>\n",
       "      <td>37</td>\n",
       "      <td>Apr-2013</td>\n",
       "      <td>15590192</td>\n",
       "      <td>762</td>\n",
       "      <td>New York, NY</td>\n",
       "      <td>HillaryClinton</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Freq   Created  Followers  Friends  \\\n",
       "FullName                                                  \n",
       "Medium                 67  May-2012    2188119       78   \n",
       "Barack Obama           67  Mar-2007   88892730   629775   \n",
       "Aaron Levie            52  Mar-2007    1626688      423   \n",
       "Clay Shirky            51  May-2007     365919      749   \n",
       "Mitch Kapor            47  Mar-2007     116400      553   \n",
       "The New York Times     45  Mar-2007   37377094      894   \n",
       "jack                   45  Mar-2006    4055017     2791   \n",
       "marissamayer           44  Nov-2008    1705305      350   \n",
       "Stewart Butterfield    43  Sep-2006      65193     2418   \n",
       "Anil Dash              40  Dec-2006     613133     4684   \n",
       "dick costolo           40  May-2007    1739649      681   \n",
       "Chris Anderson         40  Nov-2008     184874      977   \n",
       "The Economist          39  May-2007   20816240      159   \n",
       "President Obama        39  Jun-2013   15459612       79   \n",
       "Farhad Manjoo 🐣        39  Mar-2007     158093     3674   \n",
       "John Maeda             38  Jul-2008     439273    21724   \n",
       "David Pogue            38  Oct-2007    1564476     1794   \n",
       "Hanna Wallach          37  Sep-2012       7537      795   \n",
       "Tim Cook               37  Jul-2013    4925005       56   \n",
       "Hillary Clinton        37  Apr-2013   15590192      762   \n",
       "\n",
       "                                               Location        UserName  \n",
       "FullName                                                                 \n",
       "Medium                            San Francisco, CA, US          Medium  \n",
       "Barack Obama                             Washington, DC     BarackObama  \n",
       "Aaron Levie                                   Palo Alto           levie  \n",
       "Clay Shirky                                    Shanghai         cshirky  \n",
       "Mitch Kapor                                     Oakland          mkapor  \n",
       "The New York Times                        New York City         nytimes  \n",
       "jack                                    California, USA            jack  \n",
       "marissamayer                          San Francisco, CA    marissamayer  \n",
       "Stewart Butterfield                          West coast         stewart  \n",
       "Anil Dash                                           NYC        anildash  \n",
       "dick costolo                              San Francisco           dickc  \n",
       "Chris Anderson                                 Berkeley          chr1sa  \n",
       "The Economist                                    London    TheEconomist  \n",
       "President Obama                        Washington, D.C.         POTUS44  \n",
       "Farhad Manjoo 🐣      I'm right over here (it's not NYC)         fmanjoo  \n",
       "John Maeda                              WordPressLandia       johnmaeda  \n",
       "David Pogue                                         USA           Pogue  \n",
       "Hanna Wallach                              Brooklyn, NY    hannawallach  \n",
       "Tim Cook                                      Cupertino        tim_cook  \n",
       "Hillary Clinton                            New York, NY  HillaryClinton  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "discover_Friends_toFollow(ids_of_interest = index,\n",
    "                          friend_list = index, \n",
    "                          prop = .5 , count =  20).fillna('').set_index('FullName')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Recommendations on the basis of a specific cluster outcome. \n",
    "E.g. Nodes to follow on the basis of top PageRank nodes in 'Spectral Clustering::Data' clusters. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Reco: 4b. After looking at Data-Science clustsers (Spectral clustering)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "'Who-To-Follow' reco after looking at  39 friends' friends:\n",
      "... Cache-Miss for 3 nodes. Updating Cache...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Freq</th>\n",
       "      <th>Created</th>\n",
       "      <th>Followers</th>\n",
       "      <th>Friends</th>\n",
       "      <th>Location</th>\n",
       "      <th>UserName</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FullName</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Pete Warden</th>\n",
       "      <td>21</td>\n",
       "      <td>May-2008</td>\n",
       "      <td>12142</td>\n",
       "      <td>1559</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>petewarden</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hanna Wallach</th>\n",
       "      <td>19</td>\n",
       "      <td>Sep-2012</td>\n",
       "      <td>7537</td>\n",
       "      <td>795</td>\n",
       "      <td>Brooklyn, NY</td>\n",
       "      <td>hannawallach</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Barack Obama</th>\n",
       "      <td>19</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>88892730</td>\n",
       "      <td>629775</td>\n",
       "      <td>Washington, DC</td>\n",
       "      <td>BarackObama</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alex Smola</th>\n",
       "      <td>19</td>\n",
       "      <td>Jun-2008</td>\n",
       "      <td>8768</td>\n",
       "      <td>48</td>\n",
       "      <td>Mountain View and Pittsburgh</td>\n",
       "      <td>smolix</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hal Daumé III</th>\n",
       "      <td>18</td>\n",
       "      <td>Jun-2010</td>\n",
       "      <td>8702</td>\n",
       "      <td>1017</td>\n",
       "      <td>Washington DC</td>\n",
       "      <td>haldaume3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chris Diehl</th>\n",
       "      <td>17</td>\n",
       "      <td>Apr-2008</td>\n",
       "      <td>4302</td>\n",
       "      <td>999</td>\n",
       "      <td>San Francisco Bay Area, CA</td>\n",
       "      <td>ChrisDiehl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Julie Steele</th>\n",
       "      <td>17</td>\n",
       "      <td>Mar-2008</td>\n",
       "      <td>4199</td>\n",
       "      <td>913</td>\n",
       "      <td>NYC and Providence</td>\n",
       "      <td>jsteeleeditor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Joseph Adler</th>\n",
       "      <td>17</td>\n",
       "      <td>Apr-2008</td>\n",
       "      <td>1583</td>\n",
       "      <td>623</td>\n",
       "      <td>Silicon Valley</td>\n",
       "      <td>jadler</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>joseph reisinger</th>\n",
       "      <td>17</td>\n",
       "      <td>Jul-2009</td>\n",
       "      <td>1933</td>\n",
       "      <td>686</td>\n",
       "      <td>現在地</td>\n",
       "      <td>josephreisinger</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Daniel Tunkelang</th>\n",
       "      <td>16</td>\n",
       "      <td>Aug-2008</td>\n",
       "      <td>8234</td>\n",
       "      <td>20</td>\n",
       "      <td>Mountain View</td>\n",
       "      <td>dtunkelang</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KDnuggets</th>\n",
       "      <td>16</td>\n",
       "      <td>Feb-2009</td>\n",
       "      <td>84083</td>\n",
       "      <td>397</td>\n",
       "      <td>Brookline, MA, USA</td>\n",
       "      <td>kdnuggets</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Demis Hassabis</th>\n",
       "      <td>16</td>\n",
       "      <td>Jun-2013</td>\n",
       "      <td>69914</td>\n",
       "      <td>21</td>\n",
       "      <td></td>\n",
       "      <td>demishassabis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ryan Adams</th>\n",
       "      <td>16</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>10417</td>\n",
       "      <td>674</td>\n",
       "      <td></td>\n",
       "      <td>ryan_p_adams</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hugo Larochelle</th>\n",
       "      <td>16</td>\n",
       "      <td>Jun-2015</td>\n",
       "      <td>22563</td>\n",
       "      <td>350</td>\n",
       "      <td></td>\n",
       "      <td>hugo_larochelle</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Oriol Vinyals</th>\n",
       "      <td>16</td>\n",
       "      <td>Oct-2015</td>\n",
       "      <td>18911</td>\n",
       "      <td>42</td>\n",
       "      <td>London, England</td>\n",
       "      <td>OriolVinyalsML</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Jack Clark</th>\n",
       "      <td>16</td>\n",
       "      <td>Oct-2009</td>\n",
       "      <td>18015</td>\n",
       "      <td>2506</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>jackclarkSF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fernando Pereira</th>\n",
       "      <td>16</td>\n",
       "      <td>Oct-2010</td>\n",
       "      <td>8045</td>\n",
       "      <td>186</td>\n",
       "      <td>Menlo Park, CA, USA</td>\n",
       "      <td>earnmyturns</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Philip (flip) Kromer</th>\n",
       "      <td>15</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>4773</td>\n",
       "      <td>1417</td>\n",
       "      <td>Austin, TX</td>\n",
       "      <td>mrflip</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Josh Reich</th>\n",
       "      <td>15</td>\n",
       "      <td>Dec-2008</td>\n",
       "      <td>5315</td>\n",
       "      <td>1278</td>\n",
       "      <td>Portland</td>\n",
       "      <td>i2pi</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Sean Gourley</th>\n",
       "      <td>15</td>\n",
       "      <td>Dec-2008</td>\n",
       "      <td>8062</td>\n",
       "      <td>1947</td>\n",
       "      <td>San Francisco, New Zealand</td>\n",
       "      <td>sgourley</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Doug Cutting</th>\n",
       "      <td>15</td>\n",
       "      <td>Jul-2007</td>\n",
       "      <td>22676</td>\n",
       "      <td>116</td>\n",
       "      <td>California, USA</td>\n",
       "      <td>cutting</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML Hipster</th>\n",
       "      <td>15</td>\n",
       "      <td>Jul-2012</td>\n",
       "      <td>12401</td>\n",
       "      <td>0</td>\n",
       "      <td>Some local minima</td>\n",
       "      <td>ML_Hipster</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Panos Ipeirotis</th>\n",
       "      <td>14</td>\n",
       "      <td>Oct-2008</td>\n",
       "      <td>3988</td>\n",
       "      <td>259</td>\n",
       "      <td>New York, NY</td>\n",
       "      <td>ipeirotis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Beau Cronin</th>\n",
       "      <td>14</td>\n",
       "      <td>Jun-2009</td>\n",
       "      <td>3705</td>\n",
       "      <td>291</td>\n",
       "      <td>Neck deep in AWS</td>\n",
       "      <td>beaucronin</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Florian Leibert</th>\n",
       "      <td>14</td>\n",
       "      <td>Jan-2009</td>\n",
       "      <td>11076</td>\n",
       "      <td>1741</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>flo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mark Reid</th>\n",
       "      <td>14</td>\n",
       "      <td>Feb-2008</td>\n",
       "      <td>4625</td>\n",
       "      <td>718</td>\n",
       "      <td>Sunnyvale, CA</td>\n",
       "      <td>mdreid</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Claudia Perlich</th>\n",
       "      <td>14</td>\n",
       "      <td>Feb-2012</td>\n",
       "      <td>2732</td>\n",
       "      <td>105</td>\n",
       "      <td>NYC</td>\n",
       "      <td>claudia_perlich</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Alistair Croll</th>\n",
       "      <td>14</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>22391</td>\n",
       "      <td>11537</td>\n",
       "      <td>Montreal, Quebec</td>\n",
       "      <td>acroll</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Joseph Turian</th>\n",
       "      <td>14</td>\n",
       "      <td>Dec-2007</td>\n",
       "      <td>2873</td>\n",
       "      <td>991</td>\n",
       "      <td>San Francisco</td>\n",
       "      <td>turian</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Neha Narkhede</th>\n",
       "      <td>14</td>\n",
       "      <td>Sep-2009</td>\n",
       "      <td>11158</td>\n",
       "      <td>678</td>\n",
       "      <td>Mountain View</td>\n",
       "      <td>nehanarkhede</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Freq   Created  Followers  Friends  \\\n",
       "FullName                                                   \n",
       "Pete Warden             21  May-2008      12142     1559   \n",
       "Hanna Wallach           19  Sep-2012       7537      795   \n",
       "Barack Obama            19  Mar-2007   88892730   629775   \n",
       "Alex Smola              19  Jun-2008       8768       48   \n",
       "Hal Daumé III           18  Jun-2010       8702     1017   \n",
       "Chris Diehl             17  Apr-2008       4302      999   \n",
       "Julie Steele            17  Mar-2008       4199      913   \n",
       "Joseph Adler            17  Apr-2008       1583      623   \n",
       "joseph reisinger        17  Jul-2009       1933      686   \n",
       "Daniel Tunkelang        16  Aug-2008       8234       20   \n",
       "KDnuggets               16  Feb-2009      84083      397   \n",
       "Demis Hassabis          16  Jun-2013      69914       21   \n",
       "Ryan Adams              16  Jan-2015      10417      674   \n",
       "Hugo Larochelle         16  Jun-2015      22563      350   \n",
       "Oriol Vinyals           16  Oct-2015      18911       42   \n",
       "Jack Clark              16  Oct-2009      18015     2506   \n",
       "Fernando Pereira        16  Oct-2010       8045      186   \n",
       "Philip (flip) Kromer    15  Mar-2007       4773     1417   \n",
       "Josh Reich              15  Dec-2008       5315     1278   \n",
       "Sean Gourley            15  Dec-2008       8062     1947   \n",
       "Doug Cutting            15  Jul-2007      22676      116   \n",
       "ML Hipster              15  Jul-2012      12401        0   \n",
       "Panos Ipeirotis         14  Oct-2008       3988      259   \n",
       "Beau Cronin             14  Jun-2009       3705      291   \n",
       "Florian Leibert         14  Jan-2009      11076     1741   \n",
       "Mark Reid               14  Feb-2008       4625      718   \n",
       "Claudia Perlich         14  Feb-2012       2732      105   \n",
       "Alistair Croll          14  Mar-2007      22391    11537   \n",
       "Joseph Turian           14  Dec-2007       2873      991   \n",
       "Neha Narkhede           14  Sep-2009      11158      678   \n",
       "\n",
       "                                          Location         UserName  \n",
       "FullName                                                             \n",
       "Pete Warden                      San Francisco, CA       petewarden  \n",
       "Hanna Wallach                         Brooklyn, NY     hannawallach  \n",
       "Barack Obama                        Washington, DC      BarackObama  \n",
       "Alex Smola            Mountain View and Pittsburgh           smolix  \n",
       "Hal Daumé III                        Washington DC        haldaume3  \n",
       "Chris Diehl             San Francisco Bay Area, CA       ChrisDiehl  \n",
       "Julie Steele                    NYC and Providence    jsteeleeditor  \n",
       "Joseph Adler                        Silicon Valley           jadler  \n",
       "joseph reisinger                               現在地  josephreisinger  \n",
       "Daniel Tunkelang                     Mountain View       dtunkelang  \n",
       "KDnuggets                       Brookline, MA, USA        kdnuggets  \n",
       "Demis Hassabis                                        demishassabis  \n",
       "Ryan Adams                                             ryan_p_adams  \n",
       "Hugo Larochelle                                     hugo_larochelle  \n",
       "Oriol Vinyals                      London, England   OriolVinyalsML  \n",
       "Jack Clark                       San Francisco, CA      jackclarkSF  \n",
       "Fernando Pereira               Menlo Park, CA, USA      earnmyturns  \n",
       "Philip (flip) Kromer                    Austin, TX           mrflip  \n",
       "Josh Reich                                Portland             i2pi  \n",
       "Sean Gourley            San Francisco, New Zealand         sgourley  \n",
       "Doug Cutting                       California, USA          cutting  \n",
       "ML Hipster                       Some local minima       ML_Hipster  \n",
       "Panos Ipeirotis                       New York, NY        ipeirotis  \n",
       "Beau Cronin                       Neck deep in AWS       beaucronin  \n",
       "Florian Leibert                  San Francisco, CA              flo  \n",
       "Mark Reid                            Sunnyvale, CA           mdreid  \n",
       "Claudia Perlich                                NYC  claudia_perlich  \n",
       "Alistair Croll                    Montreal, Quebec           acroll  \n",
       "Joseph Turian                        San Francisco           turian  \n",
       "Neha Narkhede                        Mountain View     nehanarkhede  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "favorite_cluster_df = results_df [results_df['spectral'].isin([0,2,5,8])]\n",
    "favorite_cluster_list = list(favorite_cluster_df.index)\n",
    "\n",
    "discover_Friends_toFollow(ids_of_interest = favorite_cluster_list, \n",
    "                          friend_list = index, \n",
    "                          prop = .5, count = 30).fillna('').set_index('FullName')  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4c.Reco:  After looking at Design clustser (Spectral clustering)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "'Who-To-Follow' reco after looking at  13 friends' friends:\n",
      "... Cache-Miss for 18 nodes. Updating Cache...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Freq</th>\n",
       "      <th>Created</th>\n",
       "      <th>Followers</th>\n",
       "      <th>Friends</th>\n",
       "      <th>Location</th>\n",
       "      <th>UserName</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FullName</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Nicholas Jitkoff</th>\n",
       "      <td>8</td>\n",
       "      <td>Jan-2007</td>\n",
       "      <td>4605</td>\n",
       "      <td>204</td>\n",
       "      <td>Palo Alto, CA</td>\n",
       "      <td>alcor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>John Maeda</th>\n",
       "      <td>8</td>\n",
       "      <td>Jul-2008</td>\n",
       "      <td>439273</td>\n",
       "      <td>21724</td>\n",
       "      <td>WordPressLandia</td>\n",
       "      <td>johnmaeda</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Joshua Porter</th>\n",
       "      <td>7</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>33903</td>\n",
       "      <td>817</td>\n",
       "      <td>Newburyport, MA</td>\n",
       "      <td>bokardo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Khoi Vinh</th>\n",
       "      <td>7</td>\n",
       "      <td>Dec-2006</td>\n",
       "      <td>347073</td>\n",
       "      <td>2347</td>\n",
       "      <td>New York, NY</td>\n",
       "      <td>khoi</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Brett Lider</th>\n",
       "      <td>7</td>\n",
       "      <td>Dec-2008</td>\n",
       "      <td>6953</td>\n",
       "      <td>947</td>\n",
       "      <td>San Francisco, California</td>\n",
       "      <td>bl</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Margaret Stewart</th>\n",
       "      <td>7</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>12846</td>\n",
       "      <td>862</td>\n",
       "      <td>Bay Area, CA</td>\n",
       "      <td>mags</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Paul Stamatiou 📷</th>\n",
       "      <td>7</td>\n",
       "      <td>Jan-2007</td>\n",
       "      <td>33711</td>\n",
       "      <td>599</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>Stammy</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GV Design</th>\n",
       "      <td>7</td>\n",
       "      <td>Aug-2011</td>\n",
       "      <td>17799</td>\n",
       "      <td>515</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>GVDesignTeam</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Marc Hemeon</th>\n",
       "      <td>6</td>\n",
       "      <td>Jan-2009</td>\n",
       "      <td>16957</td>\n",
       "      <td>3322</td>\n",
       "      <td>Southern California</td>\n",
       "      <td>hemeon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Michael Leggett</th>\n",
       "      <td>6</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>7999</td>\n",
       "      <td>209</td>\n",
       "      <td>Seattle</td>\n",
       "      <td>leggett</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>M.G. Siegler</th>\n",
       "      <td>6</td>\n",
       "      <td>Jan-2007</td>\n",
       "      <td>190837</td>\n",
       "      <td>996</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>mgsiegler</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Jenna Bilotta</th>\n",
       "      <td>6</td>\n",
       "      <td>Apr-2009</td>\n",
       "      <td>12558</td>\n",
       "      <td>287</td>\n",
       "      <td>San Francisco, CA</td>\n",
       "      <td>jenna</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Noah Levin</th>\n",
       "      <td>6</td>\n",
       "      <td>Nov-2008</td>\n",
       "      <td>5295</td>\n",
       "      <td>848</td>\n",
       "      <td>New York, NY</td>\n",
       "      <td>nlevin</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Geoff Teehan</th>\n",
       "      <td>6</td>\n",
       "      <td>Oct-2008</td>\n",
       "      <td>22642</td>\n",
       "      <td>810</td>\n",
       "      <td>Palo Alto, CA</td>\n",
       "      <td>gt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Jonathan Lee</th>\n",
       "      <td>6</td>\n",
       "      <td>Apr-2009</td>\n",
       "      <td>2771</td>\n",
       "      <td>610</td>\n",
       "      <td></td>\n",
       "      <td>hifromjonathan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Google</th>\n",
       "      <td>6</td>\n",
       "      <td>Feb-2009</td>\n",
       "      <td>18020558</td>\n",
       "      <td>204</td>\n",
       "      <td>Mountain View, CA</td>\n",
       "      <td>Google</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Paul Adams</th>\n",
       "      <td>6</td>\n",
       "      <td>Jan-2007</td>\n",
       "      <td>22637</td>\n",
       "      <td>387</td>\n",
       "      <td>Dublin</td>\n",
       "      <td>Padday</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dustin senos</th>\n",
       "      <td>6</td>\n",
       "      <td>Sep-2006</td>\n",
       "      <td>11846</td>\n",
       "      <td>721</td>\n",
       "      <td>West Coast</td>\n",
       "      <td>dustin</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Doug Bowman</th>\n",
       "      <td>6</td>\n",
       "      <td>Mar-2007</td>\n",
       "      <td>232398</td>\n",
       "      <td>500</td>\n",
       "      <td>San Diego</td>\n",
       "      <td>stop</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Katie Dill</th>\n",
       "      <td>6</td>\n",
       "      <td>Apr-2009</td>\n",
       "      <td>4751</td>\n",
       "      <td>413</td>\n",
       "      <td>San Francisco</td>\n",
       "      <td>lil_dill</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Freq   Created  Followers  Friends  \\\n",
       "FullName                                               \n",
       "Nicholas Jitkoff     8  Jan-2007       4605      204   \n",
       "John Maeda           8  Jul-2008     439273    21724   \n",
       "Joshua Porter        7  Mar-2007      33903      817   \n",
       "Khoi Vinh            7  Dec-2006     347073     2347   \n",
       "Brett Lider          7  Dec-2008       6953      947   \n",
       "Margaret Stewart     7  Mar-2007      12846      862   \n",
       "Paul Stamatiou 📷     7  Jan-2007      33711      599   \n",
       "GV Design            7  Aug-2011      17799      515   \n",
       "Marc Hemeon          6  Jan-2009      16957     3322   \n",
       "Michael Leggett      6  Mar-2007       7999      209   \n",
       "M.G. Siegler         6  Jan-2007     190837      996   \n",
       "Jenna Bilotta        6  Apr-2009      12558      287   \n",
       "Noah Levin           6  Nov-2008       5295      848   \n",
       "Geoff Teehan         6  Oct-2008      22642      810   \n",
       "Jonathan Lee         6  Apr-2009       2771      610   \n",
       "Google               6  Feb-2009   18020558      204   \n",
       "Paul Adams           6  Jan-2007      22637      387   \n",
       "dustin senos         6  Sep-2006      11846      721   \n",
       "Doug Bowman          6  Mar-2007     232398      500   \n",
       "Katie Dill           6  Apr-2009       4751      413   \n",
       "\n",
       "                                   Location        UserName  \n",
       "FullName                                                     \n",
       "Nicholas Jitkoff              Palo Alto, CA           alcor  \n",
       "John Maeda                  WordPressLandia       johnmaeda  \n",
       "Joshua Porter               Newburyport, MA         bokardo  \n",
       "Khoi Vinh                      New York, NY            khoi  \n",
       "Brett Lider       San Francisco, California              bl  \n",
       "Margaret Stewart               Bay Area, CA            mags  \n",
       "Paul Stamatiou 📷          San Francisco, CA          Stammy  \n",
       "GV Design                 San Francisco, CA    GVDesignTeam  \n",
       "Marc Hemeon             Southern California          hemeon  \n",
       "Michael Leggett                     Seattle         leggett  \n",
       "M.G. Siegler              San Francisco, CA       mgsiegler  \n",
       "Jenna Bilotta             San Francisco, CA           jenna  \n",
       "Noah Levin                     New York, NY          nlevin  \n",
       "Geoff Teehan                  Palo Alto, CA              gt  \n",
       "Jonathan Lee                                 hifromjonathan  \n",
       "Google                    Mountain View, CA          Google  \n",
       "Paul Adams                           Dublin          Padday  \n",
       "dustin senos                     West Coast          dustin  \n",
       "Doug Bowman                       San Diego            stop  \n",
       "Katie Dill                    San Francisco        lil_dill  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "discover_Friends_toFollow(ids_of_interest = list(results_df [results_df['spectral'] == 9].index), \n",
    "                          friend_list = index, \n",
    "                          prop = 1, count = 20).fillna('').set_index('FullName')  "
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}