Anteroom Data Model

Macro data intelligence system for historical market analysis, event monitoring, and scenario research.

Anteroom Data Model is a Python-based research system that collects long-range market and macroeconomic data, monitors world-event signals, compares current conditions against historical stress periods, and produces structured scenario summaries through local or API-based language models.

Built by Anteroom Studio as part of its research systems and intelligence tooling.

---

Overview

This project was designed to study market regimes as connected systems rather than isolated charts. It combines historical financial data, live market snapshots, world-event monitoring, lead-lag analysis, and historical similarity matching.

It helps explore questions such as:

• Which indicators historically move before others?
• Which historical stress periods resemble current conditions?
• What categories of news are most relevant to market risk?
• How do equities, commodities, rates, currencies, and crypto behave across similar regimes?

This is a research and analysis tool, not a trading signal service.

---

System Flow

Historical Data + Live Market Data + World News Feeds
                         ↓
Data Normalization and Storage
                         ↓
Correlation and Lead-Lag Engine
                         ↓
Historical Stress-Period Matching
                         ↓
Optional Local/API Model Summary
                         ↓
Terminal Dashboard

---

Core Capabilities

• Historical market and macro data collection
• Live market refresh cycle
• Cross-asset correlation analysis
• Lead-lag relationship detection
• Historical stress-period comparison
• RSS-based world-event monitoring
• Local LLM support through Ollama
• Optional Anthropic API fallback
• Terminal dashboard for live review
• Hardware-aware launcher

---

Data Sources

Source	Coverage
FRED	Inflation, GDP, rates, unemployment, treasury data
Yahoo Finance	Equities, commodities, volatility, market indices
CoinGecko	Bitcoin and Ethereum market data
World Bank	Global macroeconomic data
RSS feeds	World events, market news, energy, crypto, policy, technology

---

Requirements

• Python 3.8+
• 8GB RAM minimum; 12GB+ recommended
• Around 2GB+ local storage for datasets
• Optional Ollama installation for local summaries
• Optional Anthropic API key for cloud model summaries

---

Setup

Install dependencies:

pip install -r requirements.txt

Create local environment settings:

cp .env.example .env

Optional local configuration:

ANTHROPIC_API_KEY=
ANTEROOM_DATA_PATH=./anteroom_data
ANTEROOM_USE_LOCAL_LLM=true
ANTEROOM_LOCAL_LLM_MODEL=phi3:mini

Never commit .env or real credentials.

---

Usage

Download historical data:

python3 data_collector.py

Run the analysis engine:

python3 correlation_engine.py

Run world-event analysis:

python3 news_brain.py

Launch the live dashboard:

python3 dashboard.py

Use the hardware-aware launcher:

python3 zai_launcher.py

The launcher file name is retained for compatibility. It can be renamed later after repository migration.

---

Local Model Support

If Ollama is installed, the system can use a local model without API costs.

ollama pull phi3:mini

Recommended starting configuration:

ANTEROOM_USE_LOCAL_LLM=true
ANTEROOM_LOCAL_LLM_MODEL=phi3:mini

Larger systems can use models such as mistral:7b, llama3:8b, or larger variants depending on available RAM/VRAM.

---

Project Structure

File	Purpose
config.py	Safe runtime configuration and environment loading
data_collector.py	Historical and live market data collection
correlation_engine.py	Correlation, lead-lag, and historical similarity analysis
news_brain.py	RSS-based world-event monitoring and market-impact mapping
dashboard.py	Terminal dashboard for live review
zai_launcher.py	Hardware-aware launcher retained for compatibility
.env.example	Safe local environment template
.gitignore	Keeps local datasets, caches, and secrets out of Git

---

Safety and Scope

Anteroom Data Model is intended for research, education, and internal experimentation. Its outputs may be incomplete, stale, or incorrect depending on data-source availability, local configuration, and model behavior.

This project does not provide financial, investment, legal, or professional advice. Always verify outputs independently before using them in any real-world decision.

---

Studio

Anteroom Studio
Research systems, intelligence interfaces, and experimental software.