audit-analytics

Audit Analytics projects

Repository for CQA: Audit market concentration in the EU. Reproducible Empirical Accounting Research. Ass I

End-to-end financial transaction anomaly detection system built with Python and Power BI. Flags 3.2% of 284,807 transactions as high-risk using Isolation Forest, Z-Score, and IQR methods with 91% precision. Simulates real-world Big 4 audit analytics workflows.

Financial audit risk analytics dashboard built with Python and Excel — anomaly detection, risk segmentation, and executive visualisation

This Python script generates a dataset of fake financial transactions, designed for audit training and testing purposes. It creates a CSV file containing a mix of normal transactions and various types of irregularities, ranging from simple anomalies to sophisticated patterns that require advanced analytical techniques to detect.

NAAIL OpenLab™ — Evidence-governed AI research and education for accounting, auditing, finance and sustainability. Reproducible benchmarks, multi-agent prototypes and human-reviewed analysis.

Audit analytics: Benford's Law, rule-based red flags and Isolation Forest over a 1M-row transaction ledger, validated against a synthetic anomaly set

AAAS — Accrual Anomaly Audit Screening. Forensic screening for U.S. public-company filings: Beneish M-Score, Altman Z, Dechow F, Jones discretionary accruals, REM and an SEC-enforcement-trained ML ensemble. Free during early access. Proprietary, not open source.

SQL-based fraud analysis of a 6M+ financial transaction dataset using PostgreSQL. The project follows audit-style workflows, preserves raw data integrity, applies performance indexing, and identifies systemic fraud patterns, balance inconsistencies, and high-risk transaction channels.

Anonymised Power Query (M) and SQL logic used for Completeness & Accuracy testing in G-SIB financial audits.

An unsupervised deep learning model (Autoencoder) for detecting accounting fraud and General Ledger anomalies using categorical embeddings.

Procurement risk analytics platform built in R/Shiny using Polish public procurement data (Atlas Przetargów). Features Benford screening, buyer & vendor risk scoring, relationship analytics, concentration monitoring and Dockerized deployment. Developed as an end-to-end analytics product for procurement risk screening and audit support.

Financial audit analytics with consistent master data, explainable anomaly review and reproducible workpapers on a synthetic ledger.

Practical guide and templates for applying data analysis in internal audit

Audit & accounting risk analytics: data validation, Isolation Forest anomaly detection, risk scoring, PostgreSQL and a Power BI dashboard

Ranks the journals in a general ledger for ISA 240 journal entry testing, with the reason for each, a GL-to-TB population check and an Excel working paper. Synthetic data; Python and IDEAScript.

Checked a year of San Francisco's vendor payments ($6.1B) for duplicate payments, duplicate vendor records, and bills paid more often than they should be. Python, SQL, Excel.

Python/pandas Continuous Controls Monitoring (CAAT) suite -- 6 audit checks over a synthetic ERP/HR dataset, surfacing 113 exceptions and $1.4M+ in quantified exposure with composite priority scoring.

Simulates an audit-style data validation workflow on transactional data — SQL and Python exception-based control tests (duplicates, referential integrity, statistical outliers) that flag data-quality issues and generate an exception summary report + chart.

Full-population internal audit tests in DuckDB SQL and Python, measured against a synthetic dataset with known anomalies.