Internal market intelligence platform — GCR Capital
Started as a few quick Python scripts. Now growing into a proper platform. Naming it ARIA going forward.
ARIA aggregates live price feeds, runs technical analysis, scores news sentiment, monitors watchlists, and fires alerts — so the desk spends less time pulling data and more time on decisions.
| Component | Stack | Purpose |
|---|---|---|
src/ |
TypeScript / Node.js | Core platform — API, services, indicators, CLI |
tools/ |
Python 3 | Original scripts (still used for quick lookups) |
ml/ |
Python 3.8 / PyTorch 1.7 | BiLSTM sentiment model for news scoring |
native/ |
C | Fast rolling statistics (called via ctypes) |
- Live and historical OHLCV data (Alpha Vantage)
- Technical indicators: SMA, EMA, RSI, MACD, Bollinger Bands, VWAP, ATR
- Portfolio tracking with cost basis and unrealised P&L
- Configurable price/indicator alerts (console + email)
- News aggregation with ticker filtering
- Headline sentiment scoring via trained BiLSTM model
- WebSocket server for real-time price streaming to clients
- Stock screener with rule-based filters
- Basic event-driven backtester
- Interactive CLI
- Node.js >= 14
- Python >= 3.8
- Redis 6+
- Alpha Vantage API key
# install TS deps
npm install
cp .env.example .env
# fill in ALPHA_VANTAGE_API_KEY at minimum
npm run dev # ts-node, no build step needed
npm run cli # interactive terminal interface
# build C extension
cd native && make
# sentiment model
cd ml
pip install -r requirements.txt
python sentiment/inference.py --headline "Inflation fears push treasury yields higher"Internal use only — do not distribute outside the desk.