cra-scraper 🏛️

CRA exam ratings scraper and analyzer.

Search the FFIEC CRA ratings database programmatically, parse Performance Evaluation PDFs from OCC, FDIC, and Federal Reserve, and identify banks that are prime CDFI partnership opportunities — using free public data from federal regulators.

Why cra-scraper?

The Community Reinvestment Act has been law for nearly 50 years and federal regulators publish exam results for every bank — but there's no programmatic way to access this data. Banks, CDFIs, fair lending researchers, and journalists all rely on manual PDF downloads from three different regulator websites. cra-scraper makes CRA exam data queryable in Python for the first time.

Installation

pip install cra-scraper

# For PE PDF parsing
pip install cra-scraper[pdf]

Quickstart

from crascraper import (
    search_ratings, find_partnership_opportunities,
    rating_distribution, build_history, find_downgrades,
    summary_report,
)

# Search FFIEC ratings. Raises FFIECAccessError on failure (e.g. HTTP 403 —
# see note below); never returns placeholder/sample data.
ratings = search_ratings(bank_name="Federal")

# Compute rating distribution
dist = rating_distribution(ratings)
print(dist)

# Find banks with low ratings — CDFI partnership opportunities
opps = find_partnership_opportunities(ratings, state="IL")
print(opps)

# Track rating history per bank
histories = build_history(ratings)
for bank_id, hist in histories.items():
    print(f"{hist.bank_name}: {hist.trend}")

# Identify downgrades
downgrades = find_downgrades(ratings)
print(downgrades)

# Generate a full Markdown summary report
report = summary_report(ratings)
print(report)

A note on cloud/datacenter environments

search_ratings() scrapes www.ffiec.gov, which sits behind Cloudflare. Requests from cloud/datacenter IP ranges (Google Colab, CI runners, hosted notebooks, most cloud VMs) are blocked with HTTP 403 at the edge. This is an IP-reputation block, not something a User-Agent or header change can clear, and it is not transient. When this happens the call raises FFIECAccessError with an explanation — it does not return placeholder data. The scrape works from a residential internet connection; run it from one to obtain the data.

Parse Performance Evaluation PDFs

from crascraper import parse_pe_pdf

pe = parse_pe_pdf("/path/to/performance_eval.pdf")
print(pe.bank_name)
print(pe.overall_rating)
print(pe.lending_test_rating)
print(pe.assessment_areas)

CRA Rating Categories

Federal Regulators Covered

• OCC — Office of the Comptroller of the Currency (national banks)
• FRS — Federal Reserve System (state member banks)
• FDIC — Federal Deposit Insurance Corporation (state non-member banks)

Use Cases

• CDFIs identifying bank partnership opportunities (banks under regulatory pressure to improve community lending)
• Banks benchmarking their CRA performance against peers
• Researchers studying CRA effectiveness and rating trends over time
• Journalists investigating bank performance in low-income communities
• Regulators analyzing exam outcomes across asset sizes and regions

Data Sources

• FFIEC CRA Ratings Database — ffiec.gov/craratings
• OCC CRA Search — occ.gov
• FDIC CRAPES — crapes.fdic.gov
• Federal Reserve CRA — federalreserve.gov

All free public data — no API keys required.

Running Tests

PYTHONPATH=. pytest tests/ -v

28 tests across all modules.

License

MIT 2026 Jaypatel1511