Weekly wash trading intelligence brief. Monitors cryptocurrency exchanges and DeFi protocols for wash trading, fake volume, and market manipulation signals.
Live: https://joshua-deng.github.io/washwatch/
- Data Collection — Weekly GitHub Actions fetches wash trading and market manipulation signals from CPW Tracker API across multiple entity/topic combinations
- AI Analysis — GitHub Models generates a risk assessment and key findings summary
- Dashboard — Static site deployed to GitHub Pages displays signals with severity classification, trend tracking, and detection methodology
Unlike general cybersecurity monitors, WashWatch focuses exclusively on market integrity:
- Multi-query coverage — Queries wash trading, market manipulation, and fake volume separately for comprehensive signal capture
- Severity classification — Automated triage of signals into critical/high/medium/low
- AI-powered analysis — Weekly risk level assessment with actionable key findings
- Historical tracking —
data/history.jsonaccumulates weekly snapshots for trend analysis over time - Detection methodology — Explains the statistical methods (Benford's Law, volume-volatility correlation, time-of-trade CV analysis) used in professional wash trading detection
- Clone this repo
- Subscribe to CPW API Basic plan (100 free requests/month)
- Add
RAPIDAPI_KEYto repository Settings → Secrets → Actions - Enable GitHub Pages (Settings → Pages → Source: GitHub Actions)
- Workflow runs weekly on Sundays, or trigger manually via Actions tab
| Method | Metric | Suspicious Threshold |
|---|---|---|
| Volume-Volatility Correlation | Pearson r between volume and price volatility | r < 0.4 sustained |
| Benford's Law (K-S Test) | First-digit distribution vs expected | test value > 1.36/√n |
| Time-of-Trade CV | Coefficient of variation of trade counts per second/minute | CV < 0.5 |
| Buy/Sell Ratio | Proportion of buy to sell market orders | Outside 0.4-0.6 |