Robust Stock Backtesting & ML Trading Tool

This production-ready tool ingests enriched historical data via Alpha Vantage and NewsAPI, publishes data to Kafka, and saves it to CSV. It then uses the enriched data to:

 - Simulate historical trades via a backtesting engine (with realistic commission and trade size).
 - Train an ML (LSTM) model on multiple features (price, volume, technical indicators, news sentiment) to help refine trading strategies.
 - Provide an interactive Streamlit UI to adjust parameters, trigger ingestion, run ML training, and simulate backtests.

Setup Instructions

Obtain API Keys:

• Alpha Vantage API Key
• NewsAPI.org Key
• Kafka broker access (e.g., free tier via Confluent Cloud)

Configure the Tool:

• Edit config/config.yaml with your API keys, Kafka settings, and simulation parameters.

Install Dependencies:

   pip install -r requirements.txt

Data Ingestion:

Run the ingestion script to fetch and enrich historical data:

   python data_ingestion/producer.py

This script fetches data from Alpha Vantage, enriches it with news sentiment from NewsAPI, publishes each record to Kafka, and saves all data to data/historical_data.csv.

Launch the User Interface:

   streamlit run ui/app.py

Use the sidebar to trigger data ingestion, train the ML model, or run a backtest.

Backtesting & ML:

• Adjust strategy parameters via the UI.
• Run backtests to see simulated trade performance.
• Train the ML model to forecast future prices and help refine strategies.

Built With

• alpha-vantage-api
• apache-kafka
• dockerfile
• google-cloud-run
• lstm
• news-api
• python
• streamlit