CarlosSilva1/xauusd-ticks

该数据集是黄金兑美元(XAU/USD)的逐笔报价数据，覆盖时间从2021年5月24日到2026年5月24日，共计五年。数据以毫秒级精度记录，包含约数亿行。数据格式为Apache Parquet，采用Snappy压缩，并按年份和月份进行Hive风格分区。每条记录包含以下字段：timestamp（UTC时间戳，毫秒精度）、bid_price（最佳买入价，以美元计）、ask_price（最佳卖出价，以美元计）、bid_volume（买入方交易量）、ask_volume（卖出方交易量）。数据集适用于高频交易策略回测、市场微观结构研究、时间序列预测模型训练（如Transformer、LSTM、N-BEATS等）、波动性分析以及执行模拟器校准等用途。数据来源于经纪商数据汇总重建，仅供研究和教育目的使用。Five years of tick-by-tick bid/ask quotes for gold against the US dollar (XAU/USD) at millisecond resolution, covering the period from May 24, 2021, to May 24, 2026. The dataset contains approximately hundreds of millions of rows, stored in Apache Parquet format with Snappy compression, and partitioned in Hive-style by year and month. Each record includes the following columns: timestamp (UTC tick time with millisecond precision), bid_price (best bid price in USD), ask_price (best ask price in USD), bid_volume (bid-side volume), and ask_volume (ask-side volume). The dataset is suitable for backtesting high-frequency trading strategies, market microstructure research, training time-series forecasting models (e.g., transformers, LSTM, N-BEATS), volatility analysis, and calibrating execution simulators. The data is reconstructed from aggregated broker feeds and is provided for research and educational purposes only.