open-navigator / scripts /datasources /fec /fec_integration.py
jcbowyer's picture
Clean HuggingFace deployment without binary files
61d29fc
"""
FEC (Federal Election Commission) Data Integration
Track political contributions and their relationship to:
- Nonprofit leadership (board members, executives)
- Policy decisions and grant awards
- Oral health advocacy funding
Data Sources:
1. FEC Bulk Data: Individual contributions, committee finances
2. FEC API: Real-time contribution tracking
3. OpenFEC: RESTful API for contribution searches
Use Cases:
- Map donor networks in oral health advocacy
- Track political influence on grant awards
- Identify politically active nonprofit leaders
- Analyze campaign finance in healthcare policy
API Documentation: https://api.open.fec.gov/developers/
Bulk Data: https://www.fec.gov/data/browse-data/?tab=bulk-data
"""
import requests
import pandas as pd
from typing import Dict, List, Optional, Tuple
from datetime import datetime
from pathlib import Path
from loguru import logger
import time
import zipfile
import io
class OpenFECAPI:
"""Client for OpenFEC API (easier than parsing bulk files)"""
BASE_URL = "https://api.open.fec.gov/v1"
def __init__(self, api_key: Optional[str] = None):
"""
Initialize OpenFEC API client
Args:
api_key: FEC API key (get from https://api.data.gov/signup/)
If None, uses 'DEMO_KEY' with lower rate limits
Note:
Get your free API key at: https://api.data.gov/signup/
DEMO_KEY has strict rate limits (30 requests/hour)
"""
self.api_key = api_key or "DEMO_KEY"
self.session = requests.Session()
self.session.headers.update({
'User-Agent': 'CommunityOne/1.0 (Civic Engagement Platform)'
})
def _make_request(self, endpoint: str, params: Dict = None) -> Dict:
"""Make API request with rate limiting"""
if params is None:
params = {}
params['api_key'] = self.api_key
url = f"{self.BASE_URL}/{endpoint}"
response = self.session.get(url, params=params)
response.raise_for_status()
# Rate limiting
time.sleep(0.2) # 5 requests/second max
return response.json()
def search_individual_contributions(
self,
contributor_name: Optional[str] = None,
contributor_city: Optional[str] = None,
contributor_state: Optional[str] = None,
contributor_employer: Optional[str] = None,
min_amount: Optional[float] = None,
max_amount: Optional[float] = None,
min_date: Optional[str] = None,
max_date: Optional[str] = None,
per_page: int = 100,
page: int = 1
) -> Dict:
"""
Search individual contributions
Args:
contributor_name: Contributor name (partial match)
contributor_city: City
contributor_state: Two-letter state code
contributor_employer: Employer name (partial match)
min_amount: Minimum contribution amount
max_amount: Maximum contribution amount
min_date: Start date (YYYY-MM-DD)
max_date: End date (YYYY-MM-DD)
per_page: Results per page (max 100)
page: Page number
Returns:
API response with contribution records
Example:
>>> api = OpenFECAPI(api_key="your_key")
>>> # Find contributions from nonprofit executives
>>> results = api.search_individual_contributions(
... contributor_employer="Community Health Center",
... contributor_state="MA",
... min_amount=1000
... )
"""
params = {
'per_page': per_page,
'page': page
}
if contributor_name:
params['contributor_name'] = contributor_name
if contributor_city:
params['contributor_city'] = contributor_city
if contributor_state:
params['contributor_state'] = contributor_state
if contributor_employer:
params['contributor_employer'] = contributor_employer
if min_amount:
params['min_amount'] = min_amount
if max_amount:
params['max_amount'] = max_amount
if min_date:
params['min_date'] = min_date
if max_date:
params['max_date'] = max_date
logger.info(f"Searching FEC contributions: {params}")
return self._make_request('schedules/schedule_a/', params)
def get_candidate_info(self, candidate_id: str) -> Dict:
"""Get information about a specific candidate"""
return self._make_request(f'candidate/{candidate_id}/')
def search_candidates(
self,
name: Optional[str] = None,
office: Optional[str] = None, # 'H' (House), 'S' (Senate), 'P' (President)
state: Optional[str] = None,
district: Optional[str] = None,
party: Optional[str] = None, # 'DEM', 'REP', etc.
cycle: Optional[int] = None,
per_page: int = 100
) -> Dict:
"""
Search for candidates
Args:
name: Candidate name (partial match)
office: Office type (H, S, P)
state: Two-letter state code
district: Congressional district (for House)
party: Party code (DEM, REP, etc.)
cycle: Election cycle year
per_page: Results per page
Returns:
API response with candidate records
"""
params = {'per_page': per_page}
if name:
params['name'] = name
if office:
params['office'] = office
if state:
params['state'] = state
if district:
params['district'] = district
if party:
params['party'] = party
if cycle:
params['cycle'] = cycle
return self._make_request('candidates/', params)
def search_committees(
self,
name: Optional[str] = None,
committee_type: Optional[str] = None,
designation: Optional[str] = None,
state: Optional[str] = None,
per_page: int = 100
) -> Dict:
"""
Search for committees
Args:
name: Committee name (partial match)
committee_type: Type (P=Presidential, H=House, S=Senate, etc.)
designation: Designation code
state: Two-letter state code
per_page: Results per page
Returns:
API response with committee records
"""
params = {'per_page': per_page}
if name:
params['name'] = name
if committee_type:
params['committee_type'] = committee_type
if designation:
params['designation'] = designation
if state:
params['state'] = state
return self._make_request('committees/', params)
class FECBulkDataLoader:
"""Load FEC bulk data files (for comprehensive historical analysis)"""
BULK_DATA_URL = "https://www.fec.gov/files/bulk-downloads"
def __init__(self, cache_dir: Path = Path("data/cache/fec")):
self.cache_dir = cache_dir
self.cache_dir.mkdir(parents=True, exist_ok=True)
def download_individual_contributions(
self,
cycle: str = "2024",
force: bool = False
) -> Path:
"""
Download bulk individual contributions file
Args:
cycle: Election cycle (e.g., "2024", "2022")
force: Force re-download even if cached
Returns:
Path to downloaded file
Note:
These files are LARGE (several GB). Consider using the API
for smaller queries or state-specific data.
"""
filename = f"indiv{cycle[-2:]}.zip" # e.g., indiv24.zip
cache_file = self.cache_dir / filename
if cache_file.exists() and not force:
logger.info(f"Using cached file: {cache_file}")
return cache_file
url = f"{self.BULK_DATA_URL}/{cycle}/{filename}"
logger.info(f"Downloading {url} (this may take a while...)")
logger.warning(f"File size is typically 1-5 GB!")
response = requests.get(url, stream=True)
response.raise_for_status()
total_size = int(response.headers.get('content-length', 0))
with open(cache_file, 'wb') as f:
downloaded = 0
for chunk in response.iter_content(chunk_size=8192):
f.write(chunk)
downloaded += len(chunk)
if total_size > 0 and downloaded % (10 * 1024 * 1024) == 0: # Every 10MB
logger.info(f"Downloaded: {downloaded / (1024*1024):.1f} MB / {total_size / (1024*1024):.1f} MB")
logger.info(f"Download complete: {cache_file}")
return cache_file
def parse_individual_contributions(
self,
zip_path: Path,
state_filter: Optional[str] = None,
employer_filter: Optional[str] = None,
min_amount: Optional[float] = None
) -> pd.DataFrame:
"""
Parse individual contributions from bulk file
Args:
zip_path: Path to bulk ZIP file
state_filter: Filter to specific state (e.g., "MA")
employer_filter: Filter by employer name (partial match)
min_amount: Minimum contribution amount
Returns:
DataFrame with contribution records
Note:
This can be memory-intensive for full files. Consider filters.
"""
logger.info(f"Parsing {zip_path}")
with zipfile.ZipFile(zip_path, 'r') as z:
# Find the main data file (usually .txt)
txt_files = [f for f in z.namelist() if f.endswith('.txt')]
if not txt_files:
raise ValueError(f"No .txt file found in {zip_path}")
data_file = txt_files[0]
logger.info(f"Reading {data_file}")
# FEC bulk files are pipe-delimited
with z.open(data_file) as f:
# Read in chunks to handle large files
chunks = []
for chunk in pd.read_csv(
f,
delimiter='|',
dtype=str, # Read as strings first
chunksize=100000,
low_memory=False
):
# Apply filters during read to reduce memory
if state_filter:
chunk = chunk[chunk['STATE'] == state_filter]
if employer_filter and 'EMPLOYER' in chunk.columns:
mask = chunk['EMPLOYER'].str.contains(
employer_filter,
case=False,
na=False
)
chunk = chunk[mask]
if min_amount and 'TRANSACTION_AMT' in chunk.columns:
chunk['TRANSACTION_AMT'] = pd.to_numeric(
chunk['TRANSACTION_AMT'],
errors='coerce'
)
chunk = chunk[chunk['TRANSACTION_AMT'] >= min_amount]
if len(chunk) > 0:
chunks.append(chunk)
if chunks:
df = pd.concat(chunks, ignore_index=True)
logger.info(f"Parsed {len(df):,} records")
return df
else:
logger.warning("No records matched filters")
return pd.DataFrame()
class PoliticalContributionMatcher:
"""Match FEC contributions to nonprofit leadership"""
def __init__(self, fec_api: OpenFECAPI):
self.api = fec_api
def find_nonprofit_leadership_contributions(
self,
officers_df: pd.DataFrame,
state_code: str,
min_amount: float = 200.0,
election_cycle: str = "2024"
) -> pd.DataFrame:
"""
Find political contributions from nonprofit officers
Args:
officers_df: DataFrame with nonprofit officers (from IRS 990)
state_code: State to search (e.g., "MA")
min_amount: Minimum contribution to track
election_cycle: Election cycle year
Returns:
DataFrame matching officers to their political contributions
"""
logger.info(f"Searching for political contributions from {len(officers_df):,} officers")
all_contributions = []
# Group by person name to avoid duplicates
if 'person_name' in officers_df.columns:
unique_names = officers_df['person_name'].dropna().unique()
else:
logger.warning("No 'person_name' column found")
return pd.DataFrame()
for name in unique_names[:100]: # Limit for demo - API rate limits
logger.info(f"Searching: {name}")
try:
results = self.api.search_individual_contributions(
contributor_name=name,
contributor_state=state_code,
min_amount=min_amount,
min_date=f"{election_cycle}-01-01"
)
if results.get('results'):
for contrib in results['results']:
# Enrich with nonprofit context
officer_match = officers_df[
officers_df['person_name'] == name
].iloc[0]
all_contributions.append({
'contributor_name': contrib.get('contributor_name'),
'contributor_city': contrib.get('contributor_city'),
'contributor_state': contrib.get('contributor_state'),
'contributor_employer': contrib.get('contributor_employer'),
'contribution_amount': contrib.get('contribution_receipt_amount'),
'contribution_date': contrib.get('contribution_receipt_date'),
'committee_name': contrib.get('committee', {}).get('name'),
'candidate_name': contrib.get('candidate_name'),
# Nonprofit context
'nonprofit_ein': officer_match.get('ein'),
'nonprofit_name': officer_match.get('organization_name'),
'officer_title': officer_match.get('title'),
'officer_compensation': officer_match.get('compensation')
})
except Exception as e:
logger.warning(f"Error searching {name}: {e}")
continue
time.sleep(1) # Rate limiting
if all_contributions:
df = pd.DataFrame(all_contributions)
logger.info(f"Found {len(df):,} contributions from nonprofit leadership")
return df
else:
return pd.DataFrame()
def analyze_political_influence(
self,
contributions_df: pd.DataFrame,
grants_df: pd.DataFrame
) -> pd.DataFrame:
"""
Analyze potential political influence on grant awards
Compare:
- Which nonprofit leaders donated to campaigns
- Which nonprofits received federal grants
- Timeline: donation → grant award
Args:
contributions_df: Political contributions by nonprofit leaders
grants_df: Federal grants received by nonprofits
Returns:
DataFrame with influence analysis
"""
logger.info("Analyzing political influence patterns")
# Merge contributions with grants by EIN
merged = contributions_df.merge(
grants_df,
left_on='nonprofit_ein',
right_on='ein',
how='inner'
)
if merged.empty:
logger.warning("No matches between contributions and grants")
return pd.DataFrame()
# Calculate time between donation and grant
if 'contribution_date' in merged.columns and 'grant_date' in merged.columns:
merged['contribution_date'] = pd.to_datetime(merged['contribution_date'])
merged['grant_date'] = pd.to_datetime(merged['grant_date'])
merged['days_donation_to_grant'] = (
merged['grant_date'] - merged['contribution_date']
).dt.days
# Aggregate by nonprofit
summary = merged.groupby('nonprofit_ein').agg({
'contribution_amount': 'sum',
'grant_amount': 'sum',
'contributor_name': 'count'
}).reset_index()
summary.columns = [
'ein',
'total_political_donations',
'total_grants_received',
'number_of_donors'
]
logger.info(f"Analyzed {len(summary):,} nonprofits with both donations and grants")
return summary
def main():
"""Example usage"""
import argparse
parser = argparse.ArgumentParser(description="Query FEC political contribution data")
parser.add_argument("--api-key", help="FEC API key (get from https://api.data.gov/signup/)")
parser.add_argument("--contributor", help="Contributor name to search")
parser.add_argument("--employer", help="Employer name to search")
parser.add_argument("--state", help="State code (e.g., MA)")
parser.add_argument("--min-amount", type=float, default=200, help="Minimum contribution amount")
parser.add_argument("--output", type=Path, default=Path("data/gold/fec"), help="Output directory")
args = parser.parse_args()
# Initialize API
api = OpenFECAPI(api_key=args.api_key)
# Search contributions
results = api.search_individual_contributions(
contributor_name=args.contributor,
contributor_employer=args.employer,
contributor_state=args.state,
min_amount=args.min_amount
)
if results.get('results'):
df = pd.DataFrame(results['results'])
print(f"\nFound {len(df):,} contributions")
print(f"\nTotal amount: ${df['contribution_receipt_amount'].sum():,.2f}")
# Save results
args.output.mkdir(parents=True, exist_ok=True)
output_file = args.output / "political_contributions.parquet"
df.to_parquet(output_file, index=False)
print(f"\nSaved to: {output_file}")
else:
print("No contributions found")
if __name__ == "__main__":
main()