In [9]:

    

%matplotlib inline
from matplotlib import pyplot as plt
import pandas as pd
import numpy as np
import seaborn as sns; sns.set();


    

In [46]:

    

applied = {
            "3M":["Dylan","Lois","Dan","Noah","Swetha","Ikenna","Melanie","Kyle","JF"],
            "DogVacay":['Dan','Dylan','Ikenna','Jimmy','Natalie','Peter','Swetha'],
            'Aurotech':['Ikenna','Kyle','Melanie','Noah','Peter','Cindy','Swetha'],
            'tronc':['Cindy','Dan','Ikenna','JF','Jimmy','Natalie','Peter','Swetha'],
            'InVenture':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Kyle','Melanie','Natalie','Peter','Swetha'],
            'Nielsen':['Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Melanie','Natalie','Noah','Peter','Swetha'],
            'BCG':['Cindy','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Melanie','Natalie','Noah','Peter','Swetha'],
            'Facebook':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Lois','Melanie','Natalie','Noah','Peter','Swetha'],
            'Netflix':['Cindy','Dylan','JF','Jimmy','Kyle','Melanie','Natalie','Peter','Swetha'],
            'Virtu':['Dylan','JF','Noah','Peter'],
            'Amazon':['Amelia','Cindy','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Lois','Natalie','Noah','Peter','Swetha'],
            'Goodyear':['Cindy','Dan','Dylan','Ikenna','Jimmy','Kyle','Melanie','Noah','Peter'],
            'HomeAway':['Cindy','Dan','Dylan','Ikenna','JF','Josh','Kyle','Lois','Melanie','Natalie','Noah','Peter'],
            'Intuit':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Kyle','Lois','Melanie','Natalie','Peter','Swetha'],
            'iSpot.tv':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Lois','Melanie','Natalie','Noah','Peter','Swetha'],
            'Payoff':['Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Kyle','Melanie','Natalie','Peter','Swetha'],
            'Red Bull':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Kyle','Natalie','Peter','Swetha'],
            'Shopify':['Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Kyle','Lois','Melanie','Noah','Peter','Swetha'],
            'Zymergen':['Amelia','Cindy','Dan','Dylan','Ikenna','JF','Jimmy','Josh','Kyle','Lois','Melanie','Natalie','Noah','Peter','Swetha'],
            }

print "{} total companies in queue.".format(len(applied))


    

    




19 total companies in queue.

In [84]:

    

fellows = ['Amelia','Cindy','Dan','Dylan','Ikenna',
           'JF','Jimmy','Josh','Kyle','Lois',
           'Melanie','Natalie','Noah','Peter','Swetha']

indexes = []
arr = np.zeros((len(applied),len(fellows)),dtype=int)
for i,c in enumerate(applied):
    indexes.append(c)
    for person in applied[c]:
        arr[i,fellows.index(person)] = 1


    

In [85]:

    

df = pd.DataFrame(data=arr,
                  index=indexes,
                  columns=fellows)
df.head()


    

    
Out[85]:






  

    

      

      
Amelia
      
Cindy
      
Dan
      
Dylan
      
Ikenna
      
JF
      
Jimmy
      
Josh
      
Kyle
      
Lois
      
Melanie
      
Natalie
      
Noah
      
Peter
      
Swetha
    
  
  

    

      
Red Bull
      
1
      
1
      
1
      
1
      
1
      
1
      
1
      
0
      
1
      
0
      
0
      
1
      
0
      
1
      
1
    
    

      
BCG
      
0
      
1
      
0
      
1
      
1
      
1
      
1
      
1
      
1
      
0
      
1
      
1
      
1
      
1
      
1
    
    

      
Shopify
      
0
      
1
      
1
      
1
      
1
      
1
      
1
      
0
      
1
      
1
      
1
      
0
      
1
      
1
      
1
    
    

      
tronc
      
0
      
1
      
1
      
0
      
1
      
1
      
1
      
0
      
0
      
0
      
0
      
1
      
0
      
1
      
1
    
    

      
Payoff
      
0
      
1
      
1
      
1
      
1
      
1
      
1
      
0
      
1
      
0
      
1
      
1
      
0
      
1
      
1
    
  

In [86]:

    

# Co-occurrence matrix is the product of the matrix and its transpose

coocc = df.T.dot(df)
coocc.head()


    

    
Out[86]:






  

    

      

      
Amelia
      
Cindy
      
Dan
      
Dylan
      
Ikenna
      
JF
      
Jimmy
      
Josh
      
Kyle
      
Lois
      
Melanie
      
Natalie
      
Noah
      
Peter
      
Swetha
    
  
  

    

      
Amelia
      
7
      
7
      
6
      
7
      
7
      
7
      
7
      
4
      
7
      
5
      
5
      
7
      
4
      
7
      
7
    
    

      
Cindy
      
7
      
16
      
12
      
14
      
15
      
14
      
14
      
7
      
15
      
7
      
13
      
13
      
10
      
16
      
14
    
    

      
Dan
      
6
      
12
      
14
      
13
      
14
      
12
      
12
      
5
      
12
      
7
      
11
      
11
      
8
      
13
      
12
    
    

      
Dylan
      
7
      
14
      
13
      
17
      
15
      
15
      
14
      
7
      
15
      
8
      
13
      
13
      
11
      
16
      
14
    
    

      
Ikenna
      
7
      
15
      
14
      
15
      
17
      
14
      
14
      
7
      
15
      
8
      
13
      
13
      
11
      
16
      
15
    
  

In [87]:

    

fig,ax = plt.subplots(1,1,figsize=(14,10))
mask = np.zeros_like(coocc)
mask[np.triu_indices_from(mask)] = True
with sns.axes_style("white"):
    sns.heatmap(coocc,
                mask = mask,
                annot=True, fmt="d",
                square=True,
                ax=ax)
    ax.set_xlabel('Fellow',fontsize=16)
    ax.set_ylabel('Fellow',fontsize=16)
    ax.set_title("Applied to the same company?",fontsize=18);


    

In [71]:

    

# People applied to unequal numbers of companies, though. Maybe I could weight by that. 

allnames = []
for job in applied:
    allnames.extend(applied[job])

from collections import Counter
cntr = Counter(allnames)

weights = {}
for person in cntr:
    weights[person] = 1./cntr[person]

print cntr
print weights


    

    




Counter({'Peter': 18, 'Dylan': 17, 'Ikenna': 17, 'Kyle': 16, 'Swetha': 16, 'Cindy': 16, 'JF': 16, 'Jimmy': 15, 'Dan': 14, 'Natalie': 14, 'Melanie': 14, 'Noah': 12, 'Lois': 8, 'Josh': 7, 'Amelia': 7})
{'Natalie': 0.07142857142857142, 'Noah': 0.08333333333333333, 'Dan': 0.07142857142857142, 'Josh': 0.14285714285714285, 'Kyle': 0.0625, 'Swetha': 0.0625, 'Lois': 0.125, 'Ikenna': 0.058823529411764705, 'Melanie': 0.07142857142857142, 'Dylan': 0.058823529411764705, 'Cindy': 0.0625, 'Jimmy': 0.06666666666666667, 'Amelia': 0.14285714285714285, 'JF': 0.0625, 'Peter': 0.05555555555555555}

In [78]:

    

indexes = []
arr = np.zeros((len(applied),len(fellows)),dtype=float)
for i,c in enumerate(applied):
    indexes.append(c)
    for person in applied[c]:
        arr[i,fellows.index(person)] = weights[person]


    

In [79]:

    

df = pd.DataFrame(data=arr,
                  index=indexes,
                  columns=fellows)
df.head()


    

    
Out[79]:






  

    

      

      
Amelia
      
Cindy
      
Dan
      
Dylan
      
Ikenna
      
JF
      
Jimmy
      
Josh
      
Kyle
      
Lois
      
Melanie
      
Natalie
      
Noah
      
Peter
      
Swetha
    
  
  

    

      
Red Bull
      
0.142857
      
0.0625
      
0.071429
      
0.058824
      
0.058824
      
0.0625
      
0.066667
      
0.000000
      
0.0625
      
0.000
      
0.000000
      
0.071429
      
0.000000
      
0.055556
      
0.0625
    
    

      
BCG
      
0.000000
      
0.0625
      
0.000000
      
0.058824
      
0.058824
      
0.0625
      
0.066667
      
0.142857
      
0.0625
      
0.000
      
0.071429
      
0.071429
      
0.083333
      
0.055556
      
0.0625
    
    

      
Shopify
      
0.000000
      
0.0625
      
0.071429
      
0.058824
      
0.058824
      
0.0625
      
0.066667
      
0.000000
      
0.0625
      
0.125
      
0.071429
      
0.000000
      
0.083333
      
0.055556
      
0.0625
    
    

      
tronc
      
0.000000
      
0.0625
      
0.071429
      
0.000000
      
0.058824
      
0.0625
      
0.066667
      
0.000000
      
0.0000
      
0.000
      
0.000000
      
0.071429
      
0.000000
      
0.055556
      
0.0625
    
    

      
Payoff
      
0.000000
      
0.0625
      
0.071429
      
0.058824
      
0.058824
      
0.0625
      
0.066667
      
0.000000
      
0.0625
      
0.000
      
0.071429
      
0.071429
      
0.000000
      
0.055556
      
0.0625
    
  

In [80]:

    

# Co-occurrence matrix is the product of the matrix and its transpose

coocc = df.T.dot(df)
coocc.head()


    

    
Out[80]:






  

    

      

      
Amelia
      
Cindy
      
Dan
      
Dylan
      
Ikenna
      
JF
      
Jimmy
      
Josh
      
Kyle
      
Lois
      
Melanie
      
Natalie
      
Noah
      
Peter
      
Swetha
    
  
  

    

      
Amelia
      
0.142857
      
0.062500
      
0.061224
      
0.058824
      
0.058824
      
0.062500
      
0.066667
      
0.081633
      
0.062500
      
0.089286
      
0.051020
      
0.071429
      
0.047619
      
0.055556
      
0.062500
    
    

      
Cindy
      
0.062500
      
0.062500
      
0.053571
      
0.051471
      
0.055147
      
0.054688
      
0.058333
      
0.062500
      
0.058594
      
0.054688
      
0.058036
      
0.058036
      
0.052083
      
0.055556
      
0.054688
    
    

      
Dan
      
0.061224
      
0.053571
      
0.071429
      
0.054622
      
0.058824
      
0.053571
      
0.057143
      
0.051020
      
0.053571
      
0.062500
      
0.056122
      
0.056122
      
0.047619
      
0.051587
      
0.053571
    
    

      
Dylan
      
0.058824
      
0.051471
      
0.054622
      
0.058824
      
0.051903
      
0.055147
      
0.054902
      
0.058824
      
0.055147
      
0.058824
      
0.054622
      
0.054622
      
0.053922
      
0.052288
      
0.051471
    
    

      
Ikenna
      
0.058824
      
0.055147
      
0.058824
      
0.051903
      
0.058824
      
0.051471
      
0.054902
      
0.058824
      
0.055147
      
0.058824
      
0.054622
      
0.054622
      
0.053922
      
0.052288
      
0.055147
    
  

In [82]:

    

fig,ax = plt.subplots(1,1,figsize=(14,10))
mask = np.zeros_like(coocc)
mask[np.triu_indices_from(mask)] = True
with sns.axes_style("white"):
    sns.heatmap(coocc,
                mask = mask,
                square=True,
                ax=ax,
                cmap='YlGnBu')
    ax.set_xlabel('Fellow',fontsize=16)
    ax.set_ylabel('Fellow',fontsize=16)
    ax.set_title("Weighted by total number of applications",fontsize=18);


    

In [ ]: