FeenQR

An Intelligent Quantitative Research Agent System

Tech Stack

---

An intelligent, agentic system built with the Microsoft Semantic Kernel Architecture and C# that performs quantitative research by analyzing YouTube videos, Market trends, and generating trading signals. This system combines the power of AI-driven content analysis with sophisticated trading algorithms.

---

Project Status

Build & Deployment

System Health

Issues & Maintenance

Development Activity

Features

Core Capabilities

• YouTube Video Analysis: Automatically fetches and analyzes financial videos from YouTube (especially Quantopian channel)
• Trading Signal Generation: Creates actionable trading signals using AI and technical analysis
• Risk Management: Comprehensive portfolio risk assessment and position sizing
• Market Data Integration: Real-time market data from Binance and other sources
• Portfolio Management: Track positions, calculate metrics, and manage trades
• Agent Orchestration: Autonomous job scheduling and execution

Advanced Research Agents

• Market Sentiment Agent: Multi-source sentiment analysis (news, social media, fear/greed index, technical)
• Statistical Pattern Agent: Deep mathematical analysis to detect exploitable market patterns

Research & Strategy Development Tools

• Monte Carlo Simulation Framework: Advanced risk modeling with stochastic scenario generation
• Interactive Strategy Builder: Visual strategy development with real-time backtesting
• Research Notebook Environment: Interactive quantitative research platform with code execution

AI-Powered Analysis

• Sentiment Analysis: Extract market sentiment from video content and internet sources
• Technical Insights: Identify trading concepts and strategies from video content
• Signal Generation: Convert insights into executable trading signals
• Risk Assessment: AI-driven portfolio risk evaluation
• Academic Research: Extract trading strategies from research papers

Architecture

Semantic Kernel Integration

The system is built around Microsoft Semantic Kernel, providing:

• Plugin Architecture: Modular AI functions for different domains
• Natural Language Processing: Advanced content analysis capabilities
• Function Calling: Structured AI interactions with trading systems
• Orchestration: Intelligent job scheduling and execution

Core Components

Services Layer

• YouTubeAnalysisService: YouTube integration and content analysis
• TradingSignalService: Signal generation and management
• MarketDataService: Real-time and historical market data
• RiskManagementService: Portfolio risk assessment
• PortfolioService: Position and performance tracking

Research Agent Services

• MarketSentimentAgentService: Multi-source sentiment analysis and market direction
• StatisticalPatternAgentService: Mathematical pattern detection and exploitability analysis

Research & Strategy Development Services

• MonteCarloService: Stochastic simulation and risk modeling
• StrategyBuilderService: Interactive strategy development and optimization
• NotebookService: Interactive research environment and code execution

Plugins Layer (Semantic Kernel Functions)

• PodcastAnalysisPlugin: AI functions for YouTube video content analysis
• YouTubeAnalysisPlugin: Dedicated YouTube video analysis functions
• TradingPlugin: Trading operations and signal management
• MarketDataPlugin: Market data retrieval and analysis
• RiskManagementPlugin: Risk assessment and portfolio management

Research Agent Plugins

• MarketSentimentPlugin: Sentiment analysis and trading signal generation
• StatisticalPatternPlugin: Pattern detection and statistical analysis functions

Research & Strategy Development Plugins

• MonteCarloPlugin: Stochastic simulation and risk analysis functions
• StrategyBuilderPlugin: Interactive strategy development and backtesting
• NotebookPlugin: Research notebook execution and analysis tools

Core Layer

• AgentOrchestrator: Main orchestration engine
• Models: Data structures and business entities

Setup and Configuration

Prerequisites

• .NET 8.0 SDK
• OpenAI API Key
• YouTube Data API Key (v3)
• Binance API Keys (optional, for live trading)

Installation

1. Clone the repository

   cd /Users/misango/codechest/ArithmaxResearchChest

2. Configure Settings Update appsettings.json with your API keys:

   {
     "OpenAI": {
       "ApiKey": "your-openai-api-key",
       "ModelId": "gpt-4o"
     },
     "Spotify": {
       "ClientId": "your-spotify-client-id",
       "ClientSecret": "your-spotify-client-secret"
     },
     "Binance": {
       "ApiKey": "your-binance-api-key",
       "SecretKey": "your-binance-secret-key",
       "TestMode": true
     }
   }

3. Install Dependencies

   dotnet restore

4. Build the Project

   dotnet build

5. Run the Agent

   dotnet run

Web Frontend

The application now includes a web-based frontend built with Blazor Server, providing an interactive dashboard with widgets for various financial analyses.

To run with the web frontend:

# Using the launcher script
./launch.sh web

# Or directly with dotnet
dotnet run -- --web

The web interface will be available at http://localhost:5000 with:

• Dashboard: Overview with market data, sentiment, portfolio, news, and risk widgets
• Market Analysis: Technical and comprehensive analysis tools
• Portfolio: Portfolio management and Alpaca integration
• Research: Academic paper search functionality

Usage Examples

Podcast Analysis

// Analyze a Spotify podcast episode
var episode = await podcastService.AnalyzePodcastAsync("https://open.spotify.com/episode/69tcEMbTyOEcPfgEJ95xos");

Trading Signal Generation

// Generate signals for all tracked symbols
var signals = await signalService.GenerateSignalsAsync();

// Generate signal for specific symbol
var btcSignal = await signalService.GenerateSignalAsync("BTCUSDT");

Risk Management

// Assess portfolio risk
var riskAssessment = await riskService.AssessPortfolioRiskAsync();

// Calculate position size
var positionSize = await riskService.CalculatePositionSizeAsync(signal);

Portfolio Management

// Get portfolio metrics
var metrics = await portfolioService.CalculateMetricsAsync();

// Execute trading signal
var executed = await portfolioService.ExecuteSignalAsync(signal, quantity);

Semantic Kernel Plugin Functions

The system exposes AI functions through Semantic Kernel plugins:

Core Trading Plugins

PodcastAnalysisPlugin

• AnalyzePodcastAsync(): Full podcast analysis with insights and signals
• GetPodcastSentimentAsync(): Extract sentiment score from podcast
• ExtractTradingInsightsAsync(): Get technical trading insights

TradingPlugin

• GenerateTradingSignalsAsync(): Generate signals for symbols
• ExecuteTradingSignalAsync(): Execute a signal with risk management
• GetActiveSignalsAsync(): Retrieve current active signals
• AnalyzeTradingPerformanceAsync(): Performance analysis and metrics

MarketDataPlugin

• GetMarketDataAsync(): Current market data for symbols
• GetHistoricalDataAsync(): Historical price data
• CompareSymbolsAsync(): Compare multiple symbols performance
• GetMarketOverviewAsync(): Overall market summary

RiskManagementPlugin

• AssessPortfolioRiskAsync(): Comprehensive risk assessment
• GetPortfolioSummaryAsync(): Portfolio positions and metrics
• CalculatePositionSizeAsync(): Optimal position sizing
• AnalyzeDiversificationAsync(): Portfolio diversification analysis

Research Agent Plugins

MarketSentimentPlugin

• AnalyzeMarketSentimentAsync(): Multi-source sentiment analysis
• CompareSentimentAcrossAssetsAsync(): Cross-asset sentiment comparison
• GenerateSentimentTradingSignalsAsync(): Sentiment-based trading signals
• GetSentimentAlertsAsync(): Extreme sentiment condition alerts
• GetSentimentTrendAsync(): Historical sentiment trend analysis

StatisticalPatternPlugin

• DetectPatternsAsync(): Comprehensive pattern detection
• DetectSpecificPatternAsync(): Target specific pattern types
• ComparePatternAcrossSymbolsAsync(): Multi-symbol pattern comparison
• DetectAnomaliesAsync(): Statistical anomaly detection
• AnalyzeVolatilityPatternsAsync(): Volatility clustering analysis
• GenerateImplementationRoadmapAsync(): Pattern implementation guidance

Research & Strategy Development Plugins

MonteCarloPlugin

• RunMonteCarloSimulationAsync(): Execute stochastic price simulations
• CalculateValueAtRiskAsync(): Compute VaR with confidence intervals
• GenerateStressTestScenariosAsync(): Portfolio stress testing under extreme conditions
• AnalyzeSimulationResultsAsync(): Statistical analysis of simulation outcomes
• VisualizeSimulationPathsAsync(): Interactive visualization of price paths

StrategyBuilderPlugin

• BuildStrategyAsync(): Construct trading strategies with components
• BacktestStrategyAsync(): Real-time strategy backtesting
• OptimizeParametersAsync(): Genetic algorithm parameter optimization
• ValidateStrategyAsync(): Walk-forward analysis and robustness testing
• GenerateStrategyReportAsync(): Performance attribution and risk metrics

NotebookPlugin

• ExecuteCodeAsync(): Run code cells with real-time output
• CreateVisualizationAsync(): Generate charts and statistical plots
• RunStatisticalAnalysisAsync(): Hypothesis testing and regression analysis
• ExportNotebookAsync(): Export research findings and documentation
• ShareAnalysisAsync(): Collaborative research sharing capabilities

Integration with Existing Strategies

This system builds upon your existing Python quantitative strategies:

Similar Patterns

• EMA/SMA Strategies: Technical indicator integration
• Scalping Strategy: High-frequency signal generation
• Swing Trading: Medium-term position management
• Risk Management: Drawdown control and position sizing
• Portfolio Optimization: Multi-asset strategy execution

Enhanced Capabilities

• AI-Driven Analysis: Semantic understanding of market content
• Automated Orchestration: Self-managing agent workflows
• Natural Language Interface: Conversational trading interactions
• Multi-Source Integration: Podcast, market data, and news analysis

Agent Workflow

1. Initialization: Load configuration and register plugins
2. Scheduled Jobs: Podcast analysis, market data updates
3. Signal Generation: AI-powered trading signal creation
4. Risk Validation: Comprehensive risk checks before execution
5. Portfolio Management: Position tracking and performance monitoring
6. Continuous Learning: Adaptation based on performance feedback

Performance Monitoring

Research Agent Capabilities

Market Sentiment Agent

Multi-Source Sentiment Analysis & Prediction

• Analyzes news sentiment from financial news sources
• Processes social media sentiment (Twitter, Reddit, Discord)
• Monitors Fear & Greed Index and volatility indicators
• Performs technical sentiment analysis
• Generates market direction predictions

Key Features:

• Real-time sentiment aggregation from multiple sources
• Confidence-weighted sentiment scoring
• Contrarian vs momentum signal generation
• Extreme sentiment alert system
• Historical sentiment trend analysis

Statistical Pattern Agent

Deep Mathematical Market Analysis

• Detects mean reversion patterns with statistical significance
• Identifies momentum and trend patterns
• Analyzes seasonal and calendar effects
• Performs volatility clustering analysis (GARCH modeling)
• Detects statistical anomalies and outliers
• Analyzes fractal patterns and self-similarity
• Identifies potential arbitrage opportunities
• Performs regime change detection
• Analyzes market microstructure patterns

Supported Pattern Types:

• Mean Reversion: ADF tests, Hurst exponent, half-life calculations
• Momentum: Autocorrelation analysis, trend strength measurement
• Seasonality: Intraday, day-of-week, monthly effects
• Volatility: GARCH parameters, volatility clustering
• Correlation: Cross-asset correlation breakdowns
• Anomaly: Outlier detection, jump detection
• Fractal: Self-similarity analysis, fractal dimension
• Arbitrage: Price discrepancy identification
• Regime: Market regime change detection
• Microstructure: Order flow analysis, volume patterns

Research & Strategy Development Tools

Monte Carlo Simulation Framework

Advanced Risk Modeling & Scenario Analysis

• Generates thousands of market scenarios using stochastic processes
• Implements Geometric Brownian Motion (GBM) for price simulation
• Performs Value-at-Risk (VaR) calculations with confidence intervals
• Analyzes portfolio stress testing under extreme market conditions
• Provides probability distributions for returns and drawdowns

Key Features:

• Customizable simulation parameters (time horizon, volatility, drift)
• Multi-asset portfolio simulation capabilities
• Historical backtesting integration
• Risk metric calculations (VaR, CVaR, Expected Shortfall)
• Interactive visualization of simulation results

Interactive Strategy Builder

Dynamic Trading Strategy Development

• Visual strategy construction with drag-and-drop components
• Real-time backtesting against historical market data
• Parameter optimization using genetic algorithms
• Walk-forward analysis for strategy robustness testing
• Performance attribution and risk decomposition

Supported Strategy Components:

• Entry Signals: Technical indicators, price patterns, volume analysis
• Exit Rules: Profit targets, stop losses, trailing stops, time exits
• Risk Management: Position sizing, portfolio heat management
• Filters: Market regime filters, volatility filters, correlation filters
• Performance Metrics: Sharpe ratio, Sortino ratio, maximum drawdown

Research Notebook Environment

Interactive Quantitative Research Platform

• Jupyter-like notebook interface for financial analysis
• Code execution with real-time output visualization
• Integration with all FeenQR services and plugins
• Markdown support for documentation and analysis notes
• Export capabilities for sharing research findings

Built-in Analysis Tools:

• Data Visualization: Charts, graphs, and statistical plots
• Statistical Analysis: Hypothesis testing, regression analysis
• Time Series Analysis: ARIMA, GARCH, cointegration testing
• Machine Learning: Model training and evaluation workflows
• Performance Analytics: Strategy benchmarking and comparison

Notebook Plugins:

• MonteCarloPlugin: Simulation execution and result analysis
• StrategyBuilderPlugin: Interactive strategy development
• DataAnalysisPlugin: Statistical and visualization functions
• ModelValidationPlugin: Backtesting and performance evaluation

The system tracks comprehensive metrics:

• Total return and P&L
• Sharpe ratio and volatility
• Maximum drawdown
• Win rate and profit factor
• Portfolio diversification
• Risk-adjusted returns

Risk Management

Built-in risk controls include:

• Maximum drawdown limits
• Position size constraints
• Volatility targeting
• Diversification requirements
• Stop-loss and take-profit management

AI Agent Capabilities

• Autonomous Operation: Self-managing job execution
• Adaptive Learning: Performance-based strategy adjustment
• Multi-Modal Analysis: Text, audio, and numerical data processing
• Contextual Decision Making: Market regime awareness
• Natural Language Interaction: Conversational AI interface

Next Steps

Completed Features

1. Research & Strategy Development Tools: Monte Carlo simulations, interactive strategy builder, and research notebook environment
2. Web Interface: Blazor Server dashboard with comprehensive widgets for market analysis, portfolio management, and research
3. AI Agent Architecture: Full Semantic Kernel integration with modular plugins and autonomous orchestration

In Progress / Future Enhancements

1. Audio Processing: Integrate speech-to-text for actual podcast transcription
2. Live Trading: Enhanced Alpaca integration for production trading
3. Advanced Machine Learning: Deep learning models for price prediction and pattern recognition
4. Multi-Asset Support: Expand beyond crypto to stocks, forex, commodities, and derivatives
5. Real-time Data Pipeline: High-frequency data processing and order book analysis
6. Cloud Deployment: Azure/container deployment with scalable architecture

This system provides a comprehensive quantitative research platform with AI-driven analysis, autonomous trading capabilities, and advanced research tools for professional quantitative finance.