This directory contains tools for analyzing trading performance, debugging AI decisions, and monitoring system behavior.
trading_performance_analyzer.py- Comprehensive trading performance analysisdecision_quality_analyzer.py- AI decision quality scoring and pattern analysisrisk_metrics_calculator.py- Risk-adjusted performance metrics calculation
ai_reasoning_analyzer.py- Analyze AI decision-making patterns and reasoning qualityprompt_performance_tester.py- Test and compare different prompt strategiesdecision_feedback_analyzer.py- Correlation between AI confidence and actual outcomes
market_regime_detector.py- Identify market conditions and regime changessymbol_performance_analyzer.py- Symbol-specific trading performance analysistiming_pattern_analyzer.py- Analysis of trading timing and market hours impact
order_execution_analyzer.py- Order execution quality and failure analysissystem_health_monitor.py- Trading system health and performance monitoringreal_time_dashboard.py- Real-time trading performance dashboard
# Overall trading performance summary
python analysis/trading_performance_analyzer.py --summary
# AI decision quality analysis
python analysis/decision_quality_analyzer.py --recent-days 7
# Symbol-specific performance
python analysis/symbol_performance_analyzer.py --symbol BTCUSDT# Full performance report with charts
python analysis/trading_performance_analyzer.py --full-report --export-charts
# AI reasoning pattern analysis
python analysis/ai_reasoning_analyzer.py --analyze-patterns --export-report
# Market regime analysis
python analysis/market_regime_detector.py --detect-current --historical-analysis# Start real-time dashboard
python analysis/real_time_dashboard.py --port 8080
# System health monitoring
python analysis/system_health_monitor.py --continuousAll analysis tools use the same database and configuration:
- Database:
../data/trading.db - Config:
../config/agent.yaml - Logs:
../logs/
- Console: Colored terminal output with tables and metrics
- JSON: Machine-readable format for integration
- HTML: Rich reports with charts and visualizations
- CSV: Raw data export for external analysis
# Install additional analysis dependencies
pip install plotly dash pandas numpy scikit-learn matplotlib seabornWhen adding new analysis tools:
- Follow the naming convention:
{category}_{tool_name}.py - Include CLI argument support with
--help - Add both summary and detailed modes
- Support multiple output formats
- Update this README with usage examples
- Machine learning model performance tracking
- Automated anomaly detection alerts
- Integration with external monitoring systems
- Custom metric definitions and tracking
- A/B testing framework for trading strategies