loan-analyzer 0.1.0

A comprehensive loan data validation and metrics engine using DuckDB.

Loan Analyzer

A lightweight, SQL-powered engine for validating loan data and computing analytical metrics using DuckDB.

Features

• Schema & Invariant Validation: Built-in SQL tests for uniqueness, reconciliation, principal tracking, and more.
• Recursive State Engine: Tracks monthly loan balances and interest accruals.
• Delinquency Waterfall: Computes Days Past Due (DPD) and classifies loans into delinquency buckets.
• DuckDB Powered: High-performance vectorized execution on local CSV files.

Installation

pip install .

Quick Start

from loan_analyzer import LoanDataLoader, LoanMetricsRunner

# 1. Load and Validate Data
loader = LoanDataLoader("path/to/your/csv_data")
loader.load_data()
is_valid, report = loader.validate()

# 2. Run Metrics
metrics = LoanMetricsRunner(loader)
metrics.run_loan_state()
metrics.run_delinquency()

SQL Resources

The package includes internal SQL resources located in:

• src/loan_analyzer/sql/: Core metrics logic.
• src/loan_analyzer/tests/: Validation tests.