Data Format Guide

StratEvo's evolution engine works with CSV files. This document describes the expected format and directory structure.

CSV Format

Required Columns

Column	Type	Description
date	string	Date/datetime (e.g. 2024-03-24 or 2024-03-24 00:00:00)
open	float	Opening price
high	float	High price
low	float	Low price
close	float	Closing price
volume	float	Trading volume

Optional Columns

Column	Type	Description
code	string	Symbol/ticker code (for multi-stock files)
amount	float	Trading amount in currency
turn	float	Turnover rate

Example: Crypto (hourly)

date,open,high,low,close,volume
2024-03-24 00:00:00,0.6232,0.634,0.6225,0.6333,395848.7683
2024-03-24 01:00:00,0.6331,0.6346,0.6302,0.6332,277089.9413

Example: A-Shares (daily)

date,code,open,high,low,close,volume,amount,turn
2024-03-01,sh.600030,20.22,20.35,20.03,20.24,72393608,1537955150.46,0.6369

Directory Structure

data/
├── crypto/              # One CSV per trading pair
│   ├── BTC_USDT.csv
│   ├── ETH_USDT.csv
│   └── ...
├── a_shares_elite/      # One CSV per stock (elite 100)
│   ├── sh.600030.csv
│   └── ...
└── a_shares_300/        # One CSV per stock (CSI 300)
    ├── sh.000001.csv
    └── ...

Each CSV file = one symbol. The filename (minus .csv) is used as the symbol identifier.

Data Preparation

Download Crypto Data

stratevo download-crypto --symbols BTC/USDT,ETH/USDT --timeframe 1h --days 365

Download A-Share Data

python scripts/download_a_shares.py

Using Your Own Data

1. Create a directory with one CSV per symbol
2. Ensure each CSV has at minimum: date, open, high, low, close, volume
3. Sort rows by date ascending
4. No NaN values in OHLCV columns

Then run evolution:

stratevo evolve --market crypto --data-dir ./my_data/ --generations 100

Data Validation

The data loader automatically: - Removes rows with NaN in any OHLCV field - Removes duplicate dates (keeps first occurrence) - Sorts by date ascending - Warns if fewer than 30 data points (minimum required for backtesting) - Skips symbols with insufficient data