Fraudit

Government Spending Fraud Detection System for Texas

Fraudit is an open-source tool that aggregates Texas state government spending data from multiple sources and applies fraud detection algorithms to identify suspicious patterns, duplicate payments, contract splitting, and vendors on federal exclusion lists.

Features

• Multi-Source Data Ingestion

  • Texas Comptroller payments (data.texas.gov)
  • Centralized Master Bidders List (CMBL)
  • LBB Contract Database
  • USASpending.gov federal grants
  • TxSmartBuy purchase orders
  • SAM.gov federal exclusions (debarred contractors)
  • Texas Ethics Commission campaign finance
  • State employee salary data
  • Sales tax permits

• Fraud Detection Rules

  • Contract splitting (clustering payments below thresholds)
  • Duplicate payment detection
  • Vendor name/address clustering
  • Debarred vendor screening (SAM.gov cross-reference)
  • Ghost vendor detection
  • Employee-vendor self-dealing
  • Pay-to-play pattern detection
  • Fiscal year-end spending rush
  • Related party transactions

• Terminal UI Dashboard

  • Real-time data visualization
  • Alert management
  • Vendor investigation tools
  • Sync status monitoring

Quick Start

One-Line Install (Linux)

git clone https://github.com/yourusername/fraudit.git
cd fraudit
./run.sh

The setup script will:

1. Install PostgreSQL if needed
2. Create the database
3. Set up Python virtual environment
4. Prompt for API keys (SAM.gov, Socrata)
5. Initialize database tables
6. Launch the terminal UI

API Keys

Fraudit works best with these API keys (the setup will prompt you):

Key	Required	Purpose	Get It
SAM.gov API Key	Recommended	Federal exclusions list	sam.gov/data-services
Socrata App Token	Optional	Faster data.texas.gov access	data.texas.gov

Without the SAM.gov key, you can manually download the exclusions ZIP file.

Usage

Terminal UI

./run.sh
# Select option 1 or just press Enter

Keyboard shortcuts:

• d - Dashboard
• v - Vendors
• a - Alerts
• s - Sync status
• q - Quit

CLI Commands

# Activate virtual environment
source .venv/bin/activate

# Sync data sources
fraudit sync run              # Smart sync (skip completed)
fraudit sync run --all        # Force full sync
fraudit sync run -s cmbl      # Sync specific source

# Run fraud detection
fraudit analyze run           # Run all detection rules
fraudit analyze run -r debarment  # Run specific rule
fraudit analyze rules         # List all detection rules

# Manage alerts
fraudit alerts list
fraudit alerts list --severity high
fraudit alerts show 123

# Search vendors
fraudit vendors search "ACME"
fraudit vendors show 1234567890000

# View configuration
fraudit config --show

Installation (Manual)

Requirements

• Python 3.10+
• PostgreSQL 14+
• Linux (tested on Arch, Ubuntu, Fedora)

Step by Step

# Clone repository
git clone https://github.com/yourusername/fraudit.git
cd fraudit

# Create virtual environment
python -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -e .

# Set up PostgreSQL
sudo systemctl start postgresql
createdb fraudit

# Copy and edit config
cp config.yaml.example config.yaml
# Edit config.yaml with your settings

# Or use environment variables / .env file
export SAM_API_KEY="your_key"
export TEXASAUDIT_SOCRATA_TOKEN="your_token"

# Initialize database
fraudit init

# Run sync and analysis
fraudit sync run
fraudit analyze run

# Start UI
fraudit tui

Configuration

Create a .env file or set environment variables:

# API Keys
SAM_API_KEY=your_sam_gov_api_key
TEXASAUDIT_SOCRATA_TOKEN=your_socrata_token

# Database (optional, defaults work for local PostgreSQL)
TEXASAUDIT_DB_HOST=/run/postgresql
TEXASAUDIT_DB_NAME=fraudit
TEXASAUDIT_DB_USER=your_username

Or edit config.yaml:

database:
  host: /run/postgresql  # Unix socket path or hostname
  port: 5432
  name: fraudit
  user: your_username

sync:
  sources:
    - cmbl
    - socrata_payments
    - sam_exclusions
    # ... more sources

detection:
  thresholds:
    contract_splitting_min: 45000
    contract_splitting_max: 50000
    vendor_name_similarity: 0.85

Project Structure

fraudit/
├── fraudit/
│   ├── ingestion/      # Data source ingestors
│   │   ├── cmbl.py           # Centralized Master Bidders List
│   │   ├── socrata.py        # data.texas.gov payments
│   │   ├── sam_exclusions.py # Federal exclusions
│   │   ├── salaries.py       # State employee salaries
│   │   ├── ethics.py         # Campaign finance
│   │   └── ...
│   ├── detection/      # Fraud detection rules
│   │   ├── contract_splitting.py
│   │   ├── debarment.py
│   │   ├── duplicates.py
│   │   ├── employee_vendor.py
│   │   ├── ghost_vendors.py
│   │   └── ...
│   ├── database/       # SQLAlchemy models
│   ├── tui/            # Terminal UI (Textual)
│   ├── normalization/  # Name/address normalization
│   └── cli.py          # Command line interface
├── config.yaml         # Configuration
├── run.sh              # Setup and launch script
└── data/               # Downloaded files and reports

Data Sources

Source	Records	Update Frequency
CMBL (Vendors)	~55,000	Daily
State Payments	~2,000,000	Weekly
LBB Contracts	~15,000	Weekly
SAM.gov Exclusions	~140,000	Daily
Campaign Finance	Variable	As filed
Employee Salaries	~300,000	Annually

Detection Rules

Rule	Description	Alert Severity
debarment	Vendor matches federal exclusion list	HIGH
contract-splitting	Payments clustered below $50K threshold	MEDIUM-HIGH
duplicate-payments	Same vendor/amount/date payments	MEDIUM
vendor-clustering	Related vendors by name/address	LOW-MEDIUM
ghost-vendors	Payments to vendors not in CMBL	MEDIUM
employee-vendor	Employee names matching vendor owners	HIGH
pay-to-play	Campaign contributions correlating with contracts	HIGH
fiscal-year-rush	Year-end spending spikes	MEDIUM

Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests: pytest
5. Submit a pull request

Legal Notice

This tool is for research and transparency purposes. Users are responsible for:

• Verifying findings before taking action
• Following applicable laws regarding public records
• Using data ethically and responsibly

False positives are expected - always investigate before drawing conclusions.

License

MIT License - see LICENSE for details.

Acknowledgments

• Texas Comptroller of Public Accounts
• data.texas.gov (Socrata)
• SAM.gov
• USASpending.gov
• Texas Ethics Commission