About SPECULA.NEWS
What is SPECULA.NEWS?
SPECULA.NEWS is a predictive news platform that analyzes current events and generates future timeline predictions. By combining real-time news data with prediction markets and AI analysis, we create comprehensive forecasts of how current events might unfold over the next 6 months.
How It Works
Our platform uses a sophisticated pipeline that processes news articles, prediction markets, and generates AI-powered future scenarios. Here's the step-by-step process:
1
Market Data Collection
We fetch the top 50 most traded prediction markets from Polymarket, a leading prediction market platform. These markets represent real-world events that people are betting on.
2
News Article Collection
We gather the top 100 news articles per category from major news sources. These articles provide the current context and recent developments in various fields.
3
Market Classification
Using sentiment analysis and AI, we categorize prediction markets into relevant categories (politics, technology, sports, etc.) to understand which markets relate to which news topics.
4
Article-Market Mapping
We analyze which news articles are relevant to which prediction markets, creating connections between current events and future predictions.
5
Sentiment Mapping
Every article is analyzed for relevance to each market, sentiment classification (Yes/No/Neutral), confidence level (1-10), strength of effect (1-10), and overall relevancy score (1-10). This creates a comprehensive mapping of how current news affects future predictions.
6
Present Summaries Generation
Using GPT-4.1, we generate structured summaries for each category including titles, subtitles, comprehensive descriptions, key insights, and recent trends. These summaries provide the foundation for future predictions.
7
Future Timeline Generation
Using AI models, we generate future timeline trees starting from the current month, with predictions extending 6 months into the future. Each prediction includes probability estimates, impact assessments, and is generated for three scenarios: optimistic, neutral, and pessimistic.
8
Future Predictions Synthesis
We combine present summaries with future timeline events to generate comprehensive future predictions for each category, including key predictions, risk factors, opportunities, and strategic recommendations.
9
General Summaries Creation
Finally, we synthesize all category-specific data into two general summaries: one for present conditions across all categories, and one for future predictions, providing a comprehensive overview for the homepage.
Data Sources & Processing
- Prediction Markets: Polymarket API - Top 50 most traded markets with real-time price data, volume, and sentiment prices (Yes/No outcomes)
- News Articles: NewsAPI.org - Top 100 articles per category from major news outlets, filtered by relevance and recency (last 7 days)
- Market Classification: gpt-4.1 analysis to categorize markets into 10 news categories with confidence scoring
- Article-Market Analysis: gpt-4.1 relevance analysis with sentiment mapping (Yes/No/Neutral), confidence (1-10), strength (1-10), and relevancy scores (1-10)
- Content Generation: gpt-4.1 for structured summaries, future timeline generation, and prediction synthesis
Categories Covered
We analyze events across 10 comprehensive categories, each with specific search terms and market classifications:
- Sports: Football, basketball, baseball, soccer, tennis, golf, racing, Olympics, and major sporting events
- Finance: Financial markets, banking, Federal Reserve, interest rates, inflation, and economic policies
- Politics: Elections, government, Congress, Senate, presidential politics, and policy changes
- Conflict & War: Military conflicts, international tensions, Ukraine, Russia, Israel, Palestine, Iran, Taiwan
- Crypto: Cryptocurrency, Bitcoin, Ethereum, blockchain technology, DeFi, NFTs
- Technology: Artificial intelligence, machine learning, software, hardware, tech innovations
- Entertainment: Movies, television, music, celebrity news, Hollywood, streaming platforms
- Health & Science: Healthcare, medical research, scientific discoveries, medicine, COVID-19
- Environment: Climate change, global warming, pollution, sustainability, renewable energy
- Business: Corporate news, mergers, acquisitions, startups, venture capital, entrepreneurship
Understanding the Visualizations
Timeline Graphs (Desktop)
Our timeline visualizations show predicted events over time. Each node represents a potential future event, with:
- Color coding: Red for high impact, orange for medium, yellow for low impact
- Opacity: Based on probability - more opaque = higher probability
- Size: Larger nodes indicate more significant events
- Month backgrounds: Alternating background colors to distinguish different time periods
- Interactive features: Drag to navigate, zoom to explore, click for details
Mobile Experience
On mobile devices, we provide an optimized experience:
- Placeholder view: Shows "Better viewed on Desktop" with option to view full timeline
- Full timeline access: "Show me anyway" button loads the complete desktop timeline
- Responsive design: Optimized for touch interaction and smaller screens
Data Visualization Features
Our visualizations include comprehensive data representation:
- Probability scoring: 0-1 scale based on AI analysis and market data
- Impact assessment: High/Medium/Low based on potential significance
- Sentiment analysis: Optimistic, neutral, and pessimistic scenarios
- Market correlation: Links between news events and prediction market outcomes
Project Goal
SPECULA.NEWS aims to democratize access to predictive analytics by combining the wisdom of crowds (prediction markets) with AI-powered analysis. We believe that by making future predictions more accessible and visual, we can help people make better-informed decisions about the future.
This project is created by Romy Haik as an exploration of predictive news and timeline visualization.
Technical Architecture
The platform is built using a sophisticated multi-stage pipeline:
- Data Collection: Python scripts with multi-threading for API calls to Polymarket and NewsAPI.org
- AI Processing: gpt-4.1 for market classification, sentiment analysis, content generation, and future prediction
- Data Processing: JSON-based data structures with comprehensive error handling and rate limiting
- Frontend: HTML, CSS, JavaScript with D3.js for interactive timeline visualizations
- Visualization: D3.js for interactive timeline graphs with drag, zoom, and click functionality
- Responsive Design: Mobile-optimized with placeholder views and touch-friendly interactions
Pipeline Components
- Market Fetcher: Collects top 50 most traded markets with price history and sentiment data
- Market Classifier: gpt-4.1 analysis to categorize markets into 10 news categories
- News Fetcher: Retrieves 100 articles per category with relevance filtering
- Article-Market Analyzer: Multi-threaded gpt-4.1 analysis for sentiment mapping and relevance scoring
- GPT Summarizer: Generates structured present summaries for each category
- Timeline Generator: Creates future timeline trees with optimistic/neutral/pessimistic scenarios
- Future Predictor: Synthesizes present and future data into comprehensive predictions
- General Summarizer: Combines all categories into homepage-ready summaries