Stroll.Theta.Sixty — 1-Minute Bar Resolution Storage

Purpose: High-resolution 1-minute OHLCV bars for options backtesting (2018-01-01 → 2025-08-29)

Architecture: Hourly Parquet files for efficient storage and selective loading

Primary source: github.com/revred/Stroll.Alpha (generation engine)

Options data: github.com/revred/Stroll.Theta.DB (main dataset)

This repo: github.com/revred/Stroll.Theta.Sixty (1-minute bars only)

Structure

Stroll.Theta.Sixty/
├─ README.md
├─ schema.md                    # Parquet schema definition
├─ manifest.json                # Index of available data
└─ bars/
   └─ YYYY/MM/DD/
      ├─ SPX_HH.parquet        # 1-minute bars for hour HH
      ├─ XSP_HH.parquet
      ├─ VIX_HH.parquet
      ├─ QQQ_HH.parquet
      ├─ GLD_HH.parquet
      └─ USO_HH.parquet

Features

• 1-minute resolution during market hours (9:30 AM - 4:00 PM ET)
• Parquet compression (~70% smaller than raw data)
• Hourly partitioning for selective loading and GitHub file size compliance
• 6 major symbols: SPX, XSP, VIX, QQQ, GLD, USO
• 7+ years coverage: January 2018 through August 2025

Usage

# Load specific hour
import pandas as pd
bars = pd.read_parquet('bars/2024/03/15/SPX_14.parquet')  # 2-3 PM

# Load full trading day
from pathlib import Path
day_files = Path('bars/2024/03/15').glob('SPX_*.parquet')
full_day = pd.concat([pd.read_parquet(f) for f in day_files])

File Specifications

• ~60 bars per file (1 hour of 1-minute data)
• ~2MB per Parquet file (compressed)
• ~96MB per trading day (8 hours × 6 symbols × 2MB)
• <100MB GitHub file limit compliance

Data Quality

• Synchronized with options data from Stroll.Theta.DB
• Market hours filtering (excludes pre/post-market)
• Holiday calendar integration
• Missing data handling with forward-fill

— Updated: 2025-09-01