FounderSimilarityCalculator.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# IEOR 135: Applied Data Science for Venture Applications\n",
    "## The Holy Grail of Venture Capital\n",
    "\n",
    "Project Team: Julian Chan, Thomas Ferry, Mudit Goyal, Nitin Sampath, Yuan Zhou\n",
    "\n",
    "IPython Notebook: Julian Chan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import json"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load in founder data file that includes information about them such as company name, years of employment, university attended, etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": false
   },
   "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>Full Name</th>\n",
       "      <th>Primary Company</th>\n",
       "      <th>Previous startups?</th>\n",
       "      <th>Consulting before start-up</th>\n",
       "      <th>Standardized University</th>\n",
       "      <th>Standardized Major</th>\n",
       "      <th>Degree Type</th>\n",
       "      <th>Standardized Graduate Institution</th>\n",
       "      <th>Standardized Graduate Studies</th>\n",
       "      <th>Graduate Diploma</th>\n",
       "      <th>Ever served as TA/Teacher/Professor/Mentor?</th>\n",
       "      <th>Years of Employment</th>\n",
       "      <th>Worked as product manager/director/head/VP?</th>\n",
       "      <th>Worked at Google?</th>\n",
       "      <th>Worked at Microsoft?</th>\n",
       "      <th>Worked in Sales?</th>\n",
       "      <th>Stanford or Berkeley</th>\n",
       "      <th>Ivy League</th>\n",
       "      <th>Crunchbase</th>\n",
       "      <th>LinkedIn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Chad Hurley</td>\n",
       "      <td>YouTube</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>Indiana University of Pennsylvania</td>\n",
       "      <td>Design</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>https://www.crunchbase.com/person/chad-hurley</td>\n",
       "      <td>http://www.linkedin.com/profile/view?id=5711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Gwyneth Paltrow</td>\n",
       "      <td>Goop Inc.</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>https://www.crunchbase.com/person/gwyneth-paltrow</td>\n",
       "      <td>https://www.linkedin.com/in/gwyneth-paltrow-48...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Jason Calacanis</td>\n",
       "      <td>Inside.com</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>Fordham University</td>\n",
       "      <td>Psychology</td>\n",
       "      <td>BA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>https://www.crunchbase.com/person/jason-calacanis</td>\n",
       "      <td>http://www.linkedin.com/in/jasoncalacanis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Tony Fadell</td>\n",
       "      <td>Nest Labs</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>University of Michigan</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>BS</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>https://www.crunchbase.com/person/tony-fadell</td>\n",
       "      <td>http://www.linkedin.com/pub/tony-fadell/0/1/380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Matt Mullenweg</td>\n",
       "      <td>Automattic</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>University of Houston</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>https://www.crunchbase.com/person/matt-mullenweg</td>\n",
       "      <td>http://www.linkedin.com/in/mattm</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Full Name Primary Company  Previous startups?  \\\n",
       "0      Chad Hurley         YouTube                   0   \n",
       "1  Gwyneth Paltrow       Goop Inc.                   0   \n",
       "2  Jason Calacanis      Inside.com                   3   \n",
       "3      Tony Fadell       Nest Labs                   1   \n",
       "4   Matt Mullenweg      Automattic                   1   \n",
       "\n",
       "   Consulting before start-up             Standardized University  \\\n",
       "0                           0  Indiana University of Pennsylvania   \n",
       "1                           0                                 NaN   \n",
       "2                           0                  Fordham University   \n",
       "3                           0              University of Michigan   \n",
       "4                           0               University of Houston   \n",
       "\n",
       "  Standardized Major Degree Type Standardized Graduate Institution  \\\n",
       "0             Design         NaN                               NaN   \n",
       "1                NaN         NaN                               NaN   \n",
       "2         Psychology          BA                               NaN   \n",
       "3   Computer Science          BS                               NaN   \n",
       "4                NaN         NaN                               NaN   \n",
       "\n",
       "  Standardized Graduate Studies Graduate Diploma  \\\n",
       "0                           NaN              NaN   \n",
       "1                           NaN              NaN   \n",
       "2                           NaN              NaN   \n",
       "3                           NaN              NaN   \n",
       "4                           NaN              NaN   \n",
       "\n",
       "   Ever served as TA/Teacher/Professor/Mentor?  Years of Employment  \\\n",
       "0                                          0.0                  4.0   \n",
       "1                                          0.0                 12.0   \n",
       "2                                          0.0                  4.0   \n",
       "3                                          0.0                 18.0   \n",
       "4                                          0.0                 15.0   \n",
       "\n",
       "   Worked as product manager/director/head/VP?  Worked at Google?  \\\n",
       "0                                            1                  1   \n",
       "1                                            1                  0   \n",
       "2                                            1                  0   \n",
       "3                                            1                  0   \n",
       "4                                            1                  0   \n",
       "\n",
       "   Worked at Microsoft?  Worked in Sales?  Stanford or Berkeley  Ivy League  \\\n",
       "0                   0.0               0.0                     0           0   \n",
       "1                   0.0               0.0                     0           0   \n",
       "2                   0.0               0.0                     0           0   \n",
       "3                   0.0               0.0                     0           0   \n",
       "4                   0.0               0.0                     0           0   \n",
       "\n",
       "                                          Crunchbase  \\\n",
       "0      https://www.crunchbase.com/person/chad-hurley   \n",
       "1  https://www.crunchbase.com/person/gwyneth-paltrow   \n",
       "2  https://www.crunchbase.com/person/jason-calacanis   \n",
       "3      https://www.crunchbase.com/person/tony-fadell   \n",
       "4   https://www.crunchbase.com/person/matt-mullenweg   \n",
       "\n",
       "                                            LinkedIn  \n",
       "0       http://www.linkedin.com/profile/view?id=5711  \n",
       "1  https://www.linkedin.com/in/gwyneth-paltrow-48...  \n",
       "2          http://www.linkedin.com/in/jasoncalacanis  \n",
       "3    http://www.linkedin.com/pub/tony-fadell/0/1/380  \n",
       "4                   http://www.linkedin.com/in/mattm  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "founders = pd.read_csv(\"data/cleaned_input_data.csv\", encoding=\"ISO-8859-1\")\n",
    "founders.fillna(\"\")\n",
    "founders.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load in Seed and Series Valuation data that includes information about how much their startup was valuated at in different stages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Full Name</th>\n",
       "      <th>Primary Company</th>\n",
       "      <th>Seed Valuation</th>\n",
       "      <th>A Valuation</th>\n",
       "      <th>B Valuation</th>\n",
       "      <th>Valuation Increase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Chad Hurley</td>\n",
       "      <td>YouTube</td>\n",
       "      <td>0</td>\n",
       "      <td>14,000,000.00</td>\n",
       "      <td>88,400,000.00</td>\n",
       "      <td>6.314285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Gwyneth Paltrow</td>\n",
       "      <td>Goop Inc.</td>\n",
       "      <td>0</td>\n",
       "      <td>40,000,000.00</td>\n",
       "      <td>65,000,000.00</td>\n",
       "      <td>1.625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Jason Calacanis</td>\n",
       "      <td>Inside.com</td>\n",
       "      <td>$7,000,000.00</td>\n",
       "      <td>11,250,000.00</td>\n",
       "      <td>0</td>\n",
       "      <td>#VALUE!</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Tony Fadell</td>\n",
       "      <td>Nest Labs</td>\n",
       "      <td>$2,700,000.00</td>\n",
       "      <td>49,210,000.00</td>\n",
       "      <td>288,120,000.00</td>\n",
       "      <td>5.854907539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Matt Mullenweg</td>\n",
       "      <td>Automattic</td>\n",
       "      <td>0</td>\n",
       "      <td>8,650,000.00</td>\n",
       "      <td>238,590,000.00</td>\n",
       "      <td>27.58265896</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Full Name Primary Company  Seed Valuation    A Valuation  \\\n",
       "0      Chad Hurley         YouTube               0  14,000,000.00   \n",
       "1  Gwyneth Paltrow       Goop Inc.               0  40,000,000.00   \n",
       "2  Jason Calacanis      Inside.com  $7,000,000.00   11,250,000.00   \n",
       "3      Tony Fadell       Nest Labs  $2,700,000.00   49,210,000.00   \n",
       "4   Matt Mullenweg      Automattic               0   8,650,000.00   \n",
       "\n",
       "      B Valuation Valuation Increase  \n",
       "0   88,400,000.00        6.314285714  \n",
       "1   65,000,000.00              1.625  \n",
       "2               0            #VALUE!  \n",
       "3  288,120,000.00        5.854907539  \n",
       "4  238,590,000.00        27.58265896  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output = pd.read_csv(\"data/cleaned_output_data.csv\", encoding=\"ISO-8859-1\")\n",
    "output.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load in the weights for each feature generated from a Random Forest Regressor. The weights are determined using importance values from predicting series valuation from information about the founder. The weights are the importance values of each feature determined by the Random Forest Regressor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "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>Feature</th>\n",
       "      <th>Importance</th>\n",
       "      <th>Feature Index</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Previous startups?</td>\n",
       "      <td>0.060639</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Consulting before start-up</td>\n",
       "      <td>0.036717</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Standardized University</td>\n",
       "      <td>0.231523</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Standardized Major</td>\n",
       "      <td>0.198227</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Degree Type</td>\n",
       "      <td>0.087502</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Feature  Importance  Feature Index\n",
       "0          Previous startups?    0.060639              1\n",
       "1  Consulting before start-up    0.036717              2\n",
       "2     Standardized University    0.231523              3\n",
       "3          Standardized Major    0.198227              4\n",
       "4                 Degree Type    0.087502              5"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load in weights for each feature (generated using RandomForestRegressor)\n",
    "weight_df = pd.read_csv(\"data/weights.csv\", encoding=\"ISO-8859-1\")\n",
    "weight_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "weights = weight_df[\"Importance\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create a similarity calculator object. See below for class API:\n",
    "* (PRIVATE METHOD) _weightedContributionToSimilarity: given 2 feature values and the weight of that feature, compute the weighted contribution of this feature to the similarity score\n",
    "    * If the feature is a string, we check for equality and assign a 0 (not equal) or 1 (equal).\n",
    "    * If the feature is a numeric value,\n",
    "        * If absolute difference between the 2 feature values is 0 or 1, assign 1\n",
    "        * If absolute difference between the 2 feature values is > 1, assign $\\frac{1}{difference}$\n",
    "        \n",
    "* (PRIVATE METHOD) _computeWeightedSimilarity: given information about a founder (in a numpy array) not in our database and the index of a founder in our database, compute the weighted similarity between the two founders\n",
    "    * This score is essentially a linear combination of the contributions computed in the function above using the appropriate weights determined by our Random Forest Regressor.\n",
    "\n",
    "\n",
    "* findKClosestFounders: given k (the number of similar founders desired) and information about a founder (in a numpy array) not in our database, find the k closest founders using their similarity score (the greater their similarity score, the closer they are)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class FounderSimilarityCalculator:\n",
    "    def __init__(self, data, weights):\n",
    "        \"\"\"\n",
    "        Input:\n",
    "            data: pandas dataframe of feature values (assumes founder name is first column, company name is second column)\n",
    "            weights: array of weights for each feature to be used in determining \"distance\" between founders\n",
    "        \"\"\"\n",
    "        self.founders = data.iloc[:,0] # the first column is assumed to be founder name\n",
    "        self.companies = data.iloc[:,1] # the second column is assumed to be company name\n",
    "        self.crunchbase = data.iloc[:,18] # the 19th column is assumed to be the founder's CrunchBase link\n",
    "        self.linkedin = data.iloc[:,19] # the 20th column is assumed to be the founder's LinkedIn link\n",
    "        \n",
    "        self.features = data.iloc[:,2:18]\n",
    "        self.weights = weights\n",
    "        self.sum_of_weights = np.sum(weights) # compute the sum of the weights as a normalizing factor for similarity\n",
    "        \n",
    "        assert self.features.shape[1] == len(self.weights) # ensure that the # of weights corresponds to # of features\n",
    "    \n",
    "    def _weightedContributionToSimilarity(self, feature1, feature2, weight):\n",
    "        \"\"\"\n",
    "        Computes the weighted contribution of the current feature to the similarity measure.\n",
    "        \n",
    "        If the two values are the same, then we add 1*weight to the similarity score.\n",
    "        \n",
    "        For features whose values are continuous, the probability of 2 values being equal is very small, so we don't want\n",
    "        to penalize them for being different. We can use a smoother penalty based on how different they are. So, if the \n",
    "        two values are different,\n",
    "            1. Compute the absolute difference between the two values\n",
    "            2. Add 1/difference * weight to the similarity score\n",
    "        \n",
    "        Input:\n",
    "            feature1: feature of founder\n",
    "            feature2: corresponding feature of different founder\n",
    "            weight: weight on the feature\n",
    "        \"\"\"\n",
    "        if isinstance(feature1, str) and isinstance(feature2, str):\n",
    "            if feature1 == feature2:\n",
    "                return weight\n",
    "            else:\n",
    "                return 0\n",
    "        elif isinstance(feature1, float) and isinstance(feature2, float):\n",
    "            if not (np.isnan(feature1) or np.isnan(feature2)):\n",
    "                diff = np.abs(feature1 - feature2)\n",
    "                if diff <= 1:\n",
    "                    return weight\n",
    "                else:\n",
    "                    return 1/diff * weight\n",
    "            else:\n",
    "                return 0\n",
    "        return 0\n",
    "    \n",
    "    def _computeWeightedSimilarity(self, founder1_data, founder2_index):\n",
    "        \"\"\"\n",
    "        Computes the weighted similarity between 2 founders.\n",
    "        \n",
    "        Input:\n",
    "            founder1_data: information (feature values) of founder 1\n",
    "            founder2_index: integer index of founder 2\n",
    "        \"\"\"\n",
    "        features1 = founder1_data\n",
    "        features2 = self.features.iloc[founder2_index,:]\n",
    "        \n",
    "        similarity = 0\n",
    "        for i in range(features1.shape[0]):\n",
    "            similarity += self._weightedContributionToSimilarity(features1[i], features2[i], self.weights[i])\n",
    "        \n",
    "        return similarity/self.sum_of_weights\n",
    "    \n",
    "    def findKClosestFounders(self, k, founder_data):\n",
    "        \"\"\"\n",
    "        Finds the k closest founders in terms of similarity to the founder corresponding to founder_index.\n",
    "        \n",
    "        Input:\n",
    "            k: # of most similar founders\n",
    "            founder_data: information (feature values) of founder of whom we wish to find similar founders\n",
    "        \"\"\"\n",
    "        assert k < self.founders.shape[0]\n",
    "        \n",
    "        similarity = np.zeros(self.founders.shape[0])\n",
    "        \n",
    "        for i in range(self.founders.shape[0]):\n",
    "            similarity[i] = self._computeWeightedSimilarity(founder_data, i)\n",
    "            \n",
    "        min_indices = similarity.argsort()[::-1][:k]\n",
    "        \n",
    "        closest_founders = []\n",
    "        closest_companies = []\n",
    "        similarity_score = []\n",
    "        linkedin_links = []\n",
    "        \n",
    "        for i in min_indices:\n",
    "            closest_founders.append(self.founders.iloc[i])\n",
    "            closest_companies.append(self.companies.iloc[i])\n",
    "            similarity_score.append(similarity[i])\n",
    "            linkedin_links.append(self.linkedin.iloc[i])\n",
    "        \n",
    "        return closest_founders, closest_companies, similarity_score, linkedin_links"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create a FounderSimilarityCalculator instance using the founder data that we read in and the weights we computed using the Random Forest Regressor. Input the data of a founder (who is NOT in our database) that we are interested in. Then find the k closest founders in our database to the external founder."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fsc = FounderSimilarityCalculator(founders, weights)\n",
    "\n",
    "# Find the 5 closest founders\n",
    "founder_data = {\"Full Name\": \"Stephen Torres\",\n",
    "               \"Primary Company\": \"PV Solar Report\",\n",
    "               \"Previous startups?\": 1,\n",
    "               \"Consulting before start-up\": 0,\n",
    "               \"Standardized University\": \"University of California Berkeley\",\n",
    "               \"Standardized Major\": \"Business\",\n",
    "               \"Degree Type\": \"BA\",\n",
    "               \"Standardized Graduate Institution\": \"Cornell University\",\n",
    "               \"Standradized Graduate Studies\": \"Business\",\n",
    "               \"Graduate Diploma\": \"MBA\",\n",
    "               \"Ever served as TA/Teacher/Professor/Mentor?\": 1,\n",
    "               \"Years of Employment\": 9,\n",
    "               \"Worked as product manager/director/head/VP?\": 0,\n",
    "               \"Worked at Google?\": 0,\n",
    "               \"Worked at Microsoft\": 0,\n",
    "               \"Worked in Sales?\": 1,\n",
    "               \"Stanford or Berkeley\": 1,\n",
    "               \"Ivy League\": 1,\n",
    "               \"Crunchbase\": \"\",\n",
    "               \"LinkedIn\": \"https://www.linkedin.com/in/stephendtorres/\"}\n",
    "\n",
    "k = 5\n",
    "closest_founders, closest_companies, similarity_score, linkedin = fsc.findKClosestFounders(k, \n",
    "                                                                            np.array(list(founder_data.values())[2:18]))\n",
    "\n",
    "closest = list(zip(closest_founders, closest_companies, similarity_score, linkedin))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Format the results to write it to a JSON file to be fed into the front-end for display to the end user."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = {}\n",
    "data[\"name\"] = founder_data[\"Full Name\"]\n",
    "data[\"company\"] = founder_data[\"Primary Company\"]\n",
    "data[\"size\"] = 3000\n",
    "data[\"link\"] = founder_data[\"LinkedIn\"]\n",
    "\n",
    "data[\"children\"] = []\n",
    "for tup in closest:\n",
    "    name = tup[0]\n",
    "    company = tup[1]\n",
    "    similarity = round(tup[2], 4)\n",
    "    linkedin = tup[3] if isinstance(tup[3], str) else \"\"\n",
    "    size = tup[2]*(10**4)\n",
    "    \n",
    "    new_dict = {\"name\": name,\n",
    "               \"company\": company,\n",
    "               \"similar\": similarity,\n",
    "               \"link\": linkedin,\n",
    "               \"size\": size,}\n",
    "    data[\"children\"].append(new_dict)\n",
    "\n",
    "with open('front-end-UI/data.json', 'w') as outfile:  \n",
    "    json.dump(data, outfile)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "name : Stephen Torres\n",
      "company : PV Solar Report\n",
      "size : 3000\n",
      "link : https://www.linkedin.com/in/stephendtorres/\n",
      "\n",
      "***************Most similar founders:***************\n",
      "\n",
      "name : Assaf Resnick\n",
      "company : BigPanda\n",
      "similar : 0.4297\n",
      "link : https://www.linkedin.com/in/assafresnick\n",
      "size : 4297.4994157025\n",
      "\n",
      "name : Steven Lam\n",
      "company : GOGOVAN\n",
      "similar : 0.4297\n",
      "link : https://hk.linkedin.com/in/stevenhylam\n",
      "size : 4297.4994157025\n",
      "\n",
      "name : David Niu\n",
      "company : TINYhr\n",
      "similar : 0.3663\n",
      "link : http://www.linkedin.com/in/davidniu\n",
      "size : 3662.5613763374377\n",
      "\n",
      "name : David Yeom\n",
      "company : Hollar\n",
      "similar : 0.333\n",
      "link : \n",
      "size : 3329.5958166704045\n",
      "\n",
      "name : Tariq Hilaly\n",
      "company : Lumity, Inc.\n",
      "similar : 0.333\n",
      "link : https://www.linkedin.com/in/tariqhilaly\n",
      "size : 3329.5958166704045\n",
      "\n"
     ]
    }
   ],
   "source": [
    "with open('front-end-UI/data.json') as infile:\n",
    "    data = json.load(infile)\n",
    "    \n",
    "    for key, val in data.items():\n",
    "        if key != \"children\":\n",
    "            print(key, \":\", val)\n",
    "        else:\n",
    "            children = data[\"children\"]\n",
    "            print(\"\\n***************Most similar founders:***************\\n\")\n",
    "            for founder_info in children:\n",
    "                for key, val in founder_info.items():\n",
    "                    print(key, \":\", val)\n",
    "                print()"
   ]
  }
 ],
 "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": 2
}