Algo Crypto: an interactive trading-rule demo (synthetic data)
Algo Crypto began as a side project in 2017 and 2018 to study how cryptocurrency prices relate to other data: social mentions, metals prices and US stock prices. The public repository contains the historical Python 3.6 implementation.
Its working part collects data from public web APIs (CryptoCompare, CoinMarketCap, Alpha Vantage and Quandl) with multithreaded Pandas code, stores gzipped JSON in Amazon S3 using boto3, and makes it queryable with AWS Glue and Athena. The repository includes regression experiments, while the machine-learning component remains unfinished.
This page is an educational illustration only: you can adjust a moving-average or momentum rule and backtest it on seeded, generated prices. It is not the original project's algorithm, uses no live data, and is not financial advice.
Features
- Three seeded synthetic markets: crypto-like, NASDAQ-like and commodity-like
- Adjustable moving-average crossover or momentum rule with transaction costs
- Backtest: equity curve, drawdown, summary statistics and trade log
- Every chart has a data table, and the controls work from the keyboard
- Side-by-side pipeline: what the repository does and what this demo does
All prices and results are synthetic and generated in your browser from a seed. Nothing is sent anywhere. Educational only, not financial advice; no live data and no trading.