YFI (YouTube Finance Intelligence) is a creator intelligence research platform designed to transform financial media into structured, measurable investment intelligence.
The system continuously analyzes financial creator content, extracts investment signals and narratives, measures post-publication performance, and evaluates creators using statistical performance metrics, signal quality analytics, creator rankings, and narrative lead-time modeling.
Rather than focusing solely on price action, YFI studies the information flows that often precede market movements.
Most market participants analyze charts.
YFI analyzes information.
The platform was built to investigate questions that traditional market tools rarely address:
- Which creators consistently identify opportunities before the broader market?
- Which narratives emerge before widespread adoption?
- Which signals generate measurable alpha after publication?
- Which creators demonstrate repeatable predictive performance?
- How does narrative propagation evolve across financial communities?
- Can information flow be measured and ranked before it is reflected in price?
The goal is to build a data-driven framework for evaluating creator influence, signal quality, and narrative timing.
Latest verified system audit:
| Dataset | Volume |
|---|---|
| Financial Videos Processed | 17,378+ |
| Creator Transcripts | 6,130+ |
| Extracted Investment Signals | 15,320+ |
| Narrative Observations | 20,842+ |
| Unique Assets Tracked | 1,661+ |
| Financial Creators Analyzed | 30 |
| Operational & Analytics Tables | 15 |
| Pipeline Operation | Continuous |
Creator Content
↓
Transcript Harvesting
↓
Gemini Signal Extraction
↓
Signal Performance Attribution
↓
Signal Quality Analytics
↓
Narrative Extraction
↓
Narrative Lead-Time Modeling
↓
Creator Ranking Engine
↓
Behavioral Intelligence Database
Collects and processes creator transcripts and source content for downstream analysis.
Uses Google's Gemini models to analyze creator transcripts and extract structured investment intelligence including:
- Tickers
- Investment theses
- Catalysts
- Opportunity signals
- Emerging narratives
The extraction system utilizes transcript chunking, batching, and retry handling to process large content volumes.
Measures post-publication signal performance using historical market data.
Current evaluation windows include:
- 7-Day Performance
- 30-Day Performance
- 90-Day Performance
Each extracted signal is linked to future market outcomes, creating a measurable feedback loop between creator recommendations and asset performance.
Computes quantitative performance metrics including:
- Mean Alpha
- Standard Deviation
- Information Ratio
- T-Statistic
- Sample Size Validation
This allows signals to be evaluated using statistical performance measures rather than engagement metrics.
Identifies recurring themes, market narratives, and emerging discussion clusters across creator communities.
Narratives are normalized, tracked, and stored for longitudinal analysis.
Measures how early creators identify narratives before broader narrative activation.
The system detects narrative adoption spikes and calculates lead-time advantages for creators who identify trends before they become widely discussed.
Ranks creators using historical signal performance metrics including:
- Signal Count
- Win Rate
- Average Alpha (7D, 30D, 90D)
- Historical Predictive Performance
The objective is to evaluate creator effectiveness using measurable outcomes rather than popularity.
YFI currently generates and maintains:
- Creator Rankings
- Signal Performance Attribution
- Signal Quality Analytics
- Asset Sentiment Datasets
- Narrative Momentum Analytics
- Narrative Lead-Time Analytics
- Creator Lead-Time Analytics
- Behavioral Intelligence Datasets
Primary Database:
data/yfi.db
Core Tables:
signals
signal_performance
transcripts
narratives
creators
creator_rankings
signal_quality
asset_sentiment
narrative_momentum
narrative_leadtime
creator_narrative_leadtime
video_queue
Real-time operational dashboard providing visibility into:
- Dataset growth
- Pipeline throughput
- Queue health
- Signal generation
- Narrative generation
- Creator activity
- Asset activity
- Operational monitoring
Operational command center used to:
- Start and stop pipeline execution
- Run system diagnostics
- Monitor queue health
- Track dataset growth
- Review creator distribution
- Monitor failure rates
- Validate transcript integrity
- Perform operational health checks
- Python
- SQLite
- Google Gemini API
- Natural Language Processing
- Behavioral Finance
- Market Data Analysis
- Statistical Analytics
- Creator Intelligence Research
Active Development
YFI is an ongoing research effort focused on understanding how creator-generated information, investment narratives, and market outcomes interact over time.
The platform serves as a testing ground for:
- Creator intelligence systems
- Narrative analytics
- Information-flow research
- Behavioral finance experimentation
- AI-assisted signal extraction
- Market intelligence infrastructure
This project is intended for research, analytics, and educational purposes.
YFI does not provide investment advice and should not be used as the sole basis for financial decisions.