In [1]:

    

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline


    

In [2]:

    

leaderboard = pd.read_csv('transfer-learning-on-stack-exchange-tags-publicleaderboard.csv')
leaderboard.sample(4)


    

    
Out[2]:






  

    

      

      
TeamId
      
TeamName
      
SubmissionDate
      
Score
    
  
  

    

      
1264
      
508529
      
No Print Good Product
      
2017-03-13 02:17:12
      
0.08511
    
    

      
107
      
415012
      
connectwithghosh
      
2016-11-28 05:22:28
      
0.06519
    
    

      
771
      
421196
      
NTU_r05921004_FourPeople
      
2017-01-05 08:52:29
      
0.14867
    
    

      
1401
      
563016
      
ObserverL
      
2017-03-23 12:14:37
      
0.13107
    
  

In [12]:

    

plt.figure(figsize=(14,6))
leaderboard['Score'].hist(bins=50)
plt.xlabel('F1 Score')


    

    
Out[12]:





<matplotlib.text.Text at 0x7f614288f048>






    

In [4]:

    

leaderboard[leaderboard.Score > 0.9]


    

    
Out[4]:






  

    

      

      
TeamId
      
TeamName
      
SubmissionDate
      
Score
    
  
  

    

      
1169
      
456730
      
Kilorad
      
2017-03-03 07:04:59
      
0.93974
    
    

      
1289
      
514346
      
Test
      
2017-03-15 01:37:11
      
0.97295
    
    

      
1340
      
456730
      
Kilorad
      
2017-03-19 12:53:43
      
0.94087
    
    

      
1341
      
456730
      
Kilorad
      
2017-03-19 13:02:53
      
0.94198
    
    

      
1342
      
456730
      
Kilorad
      
2017-03-19 14:04:52
      
0.94635
    
    

      
1361
      
442101
      
Mice Labs
      
2017-03-20 08:19:27
      
0.91404
    
  

In [5]:

    

topByTeam = leaderboard.sort_values('Score', ascending=False).groupby('TeamId').first().reset_index()


    

In [14]:

    

plt.figure(figsize=(14,6))
topByTeam.Score.hist(bins=30)
plt.xlabel("F1 Score")


    

    
Out[14]:





<matplotlib.text.Text at 0x7f6142579240>






    

In [ ]: