contrib_all.py

"""
contrib_all.py
Spring 2022 PJW

Select contribution data from the FEC file and save it
in pickled form for analysis by a subsequent script.
"""

import pandas as pd
import matplotlib.pyplot as plt

#
#  Read the CSV file inside the zip archive
#

zipname = 'contributions.zip'
pklname = 'contrib_all_pkl.zip'

raw = pd.read_csv(zipname,dtype=str)

n_now = len(raw)
print( 'raw records:', n_now )

raw = raw.rename( columns={'TRANSACTION_PGI':'PGI'} )

#
#  Convert the transactions into floats
#

raw['amt'] = raw['TRANSACTION_AMT'].astype(float)

#
#  Build a date variable and trim off early records
#

ymd = pd.to_datetime(raw['TRANSACTION_DT'],format="%m%d%Y")

raw['date'] = ymd.dt.to_period('M')

n_last = n_now

year_ok = ymd.dt.year >= 2019
contrib = raw[ year_ok ]

n_now = len(contrib)
print( 'dropped due to year:', n_last-n_now )

#
#  Pick out the records for the primary and general elections
#  in 2020
#

n_last = n_now
contrib = contrib.query( 'PGI=="P2020" or PGI=="G2020"')

n_now = len(contrib)
print( 'dropped due to PGI', n_last-n_now )

print( 'final records:', n_now )

#
#  Trim down the varibles to those of interest
#

keepvars = ['CMTE_ID','STATE','ZIP_CODE','PGI','date','amt']

trimmed = contrib[keepvars]

#
#  Write it out in pickled form
#

trimmed.to_pickle(pklname)

#
#  Print some information about the dataframe
#

print('records:', len(trimmed) )
print('variables:', list(trimmed.columns) )

#
#  Plot primary and general election contributions by month
#

grouped = trimmed.groupby(['date','PGI'])
by_date_pgi = grouped['amt'].sum()/1e6

by_date_wide = by_date_pgi.unstack('PGI')

fig1, ax1 = plt.subplots(dpi=300)
fig1.suptitle('Individual Contributions')
by_date_wide.plot(ax=ax1)
ax1.set_xlabel('Date')
ax1.set_ylabel('Million Dollars')
fig1.tight_layout()
fig1.savefig('by_month.png')