Candidate party classification

In [45]:

    

import pandas as pd


    

In [46]:

    

df = pd.read_csv("/home/ben/Code/django-calaccess-processed-data/example/django.log", sep="|")


    

In [47]:

    

df.head()


    

    
Out[47]:






  

    

      

      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 578 Candidacy objects
    
  
  

    

      
0
      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 342 CandidateContest objects
    
    

      
1
      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 166 BallotMeasureContest objects
    
    

      
2
      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 2 RetentionContest objects
    
    

      
3
      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 55 Election objects
    
    

      
4
      
DEBUG
      
19/Jul/2017 01:10:40
      
__init__
      
Flushing 2 Membership objects
    
  

In [48]:

    

df.columns = ['level', 'time', 'logger', 'message']


    

In [49]:

    

df = df[df.logger == 'candidates']


    

In [50]:

    

def method(message):
    if 'based' in message:
        return message.split("based on")[-1]
    else:
        return message.split("after")[-1]


    

In [51]:

    

df['reason'] = df.message.apply(method)


    

In [52]:

    

df.reason.value_counts()


    

    
Out[52]:





 Form 501 party                   2192
 failing to find a match            71
 office                             19
 checking its scraped filer id       9
 Form 501 filer id                   9
 correction                          4
Name: reason, dtype: int64

In [53]:

    

df[df.reason.str.contains("correction")]


    

    
Out[53]:






  

    

      

      
level
      
time
      
logger
      
message
      
reason
    
  
  

    

      
1707
      
DEBUG
      
19/Jul/2017 01:14:58
      
candidates
      
WALLS, JIMELLE L. party set to NO PARTY PREFER...
      
correction
    
    

      
1715
      
DEBUG
      
19/Jul/2017 01:14:59
      
candidates
      
WINSTON, ALMA MARIE party set to REPUBLICAN ba...
      
correction
    
    

      
1775
      
DEBUG
      
19/Jul/2017 01:15:07
      
candidates
      
WAHL, BERNT party set to NO PARTY PREFERENCE b...
      
correction
    
    

      
2037
      
DEBUG
      
19/Jul/2017 01:15:43
      
candidates
      
RODRIGUEZ, GREG party set to DEMOCRATIC based ...
      
correction