AdamAtractor/24_crypto_orderbook_30lvl

---
license: cc-by-4.0
task_categories:
- time-series-forecasting
- reinforcement-learning
tags:
- finance
- crypto
- algorithmic-trading
- orderbook
- level-2
- market-microstructure
- quant
- hyperliquid
size_categories:
- 100K<n<1M
---

# 📊 Institutional Crypto L2 Orderbook Depth (30 Levels, 1m) - Free Sample

> **⚠️ NOTE: This is a 7-day FREE SAMPLE dataset.** > You are currently viewing 7 days of continuous, high-fidelity market data for 24 crypto assets.
> Stop wasting weeks on data engineering. Get the complete, institutional-grade dataset featuring **12+ months of continuous history** directly at 👉 **[ImbalanceLabs.com](https://imbalancelabs.com/)**

---

## 🛑 Stop Trading on "Liquidity Illusions"

Standard OHLCV candles hide the true market intent, spread, spoofing walls, and slippage. When training Deep Learning (LSTM, Transformers) or Reinforcement Learning (RL) trading agents, standard candles often lead to overfitting and account liquidations in live environments because the models cannot evaluate the real cost of execution.

This dataset provides a sample of **Cleaned, Normalized, and Time-Aligned Level 2 Orderbook Depth Data**. It extracts the raw microstructural edge from the **Hyperliquid DEX** (the leading L1 Perpetual Decentralized Exchange) and transforms it into noise-free, time-aligned 1-minute bars.

## 🗂️ Dataset Structure (What's Inside?)

This specific sample contains **3 files per asset** for 24 major perpetual futures (BTC, ETH, SOL, ADA, APT, ARB, ATOM, AVAX, BNB, DOGE, DOT, ETC, FIL, INJ, LINK, LTC, NEAR, OP, SEI, SUI, TIA, WIF, XLM, XRP).

For every asset, you get:

**1. `[ASSET]_1m_depth30.csv` (Ultra-Deep Microstructure)**
* 1-minute time-aligned orderbook snapshots.
* **30 levels of depth** (bids and asks).
* Includes cumulative passive volume and basis-point distance from the mid-price for all 30 levels.
* Perfect for high-frequency feature engineering, liquidity profiling, and RL environments.

**2. `[ASSET]_1m_ohlcv.csv` (Standard 1-minute Flow)**
* 1-minute aggregated standard candles.
* Includes mid-price tracking (`open`, `high`, `low`, `close`) and taker flow volume proxy.

**3. `[ASSET]_5m_ohlcv.csv` (Standard 5-minute Flow)**
* 5-minute aggregated standard candles.
* Clean, time-aligned pricing data ready to be merged with your depth indicators.

## ⚡ Why This Data is Different?

* **No Clock Drift:** Time-aligned bars, eliminating the asynchrony and timestamp issues of raw WebSocket feeds.
* **Extreme Depth:** **30 levels** of bid/ask liquidity mapped out mathematically.
* **Granular Resolution:** **1-minute** snapshots for high-fidelity backtesting.
* **Ready for AI/ML:** Feed directly into Pandas, DuckDB, RL environments, or XGBoost.
* **True Market Intent:** Escaping CEX noise and spoofing. DEX orderbooks reflect genuine liquidity and institutional positioning.

## 🚀 Accelerate Your Research

Building L2 depth pipelines from raw decentralized exchange feeds requires custom parsers, strict timestamp alignment, and massive compute.

Skip the heavy data engineering. Download the ready-to-use CSV datasets and jump straight to training models, backtesting TWAP/VWAP strategies, and extracting alpha.

👉 **[Download the Full 12-Month Dataset at ImbalanceLabs.com](https://imbalancelabs.com/)**

---
### License & Legal Disclaimer
This sample dataset is provided under the CC-BY-4.0 license for research and analytical purposes. The data constitutes an Aggregated Liquidity and Orderbook Depth Index (Derived Data) — a proprietary, mathematically derived analytical product. It does not constitute financial advice. Imbalance Labs is not affiliated with any cryptocurrency exchange.
---
license: cc-by-4.0
---

license: CC-BY-4.0
task_categories:
- 时间序列预测
- 强化学习
tags:
- 金融
- 加密货币
- 算法交易
- 订单簿
- 二级（Level 2）行情
- 市场微观结构
- 量化交易
- Hyperliquid
size_categories:
- 10万<n<100万

# 📊 机构级加密货币二级订单簿深度数据（30档，1分钟粒度）- 免费样本

> **⚠️ 注意：本数据集为7天免费样本。**
> 您当前查看的是24种加密资产的7天连续高保真市场数据。
> 无需耗费数周进行数据工程工作，可直接前往👉 **[ImbalanceLabs.com](https://imbalancelabs.com/)** 获取包含12个月以上连续历史数据的完整机构级数据集。

---

## 🛑 切勿基于“流动性假象”进行交易
标准OHLCV（Open-High-Low-Close-Volume）蜡烛图隐藏了真实的市场意图、买卖价差、虚假挂单墙与滑点。在训练深度学习（LSTM、Transformer）或强化学习（RL）交易智能体时，标准蜡烛图往往会导致模型过拟合，在实盘环境中造成账户爆仓，因为模型无法评估真实的交易执行成本。

本数据集提供经过清洗、归一化与时间对齐的二级订单簿深度数据样本。其从**Hyperliquid去中心化交易所（DEX，Decentralized Exchange，该领域头部的一级永续合约去中心化交易所）**提取原始微观结构信息，并将其转换为无噪声、时间对齐的1分钟粒度K线数据。

## 🗂️ 数据集结构（内含内容）
本专属样本包含24种主流永续合约（BTC、ETH、SOL、ADA、APT、ARB、ATOM、AVAX、BNB、DOGE、DOT、ETC、FIL、INJ、LINK、LTC、NEAR、OP、SEI、SUI、TIA、WIF、XLM、XRP）的**单资产3个文件**。

针对每一种资产，您将获得：
**1. `[ASSET]_1m_depth30.csv`（超深度微观结构数据）**
* 1分钟时间对齐的订单簿快照
* **30档深度**（买单与卖单）
* 包含30个档位的累计被动成交成交量与相对于中间价的基点距离
* 适用于高频特征工程、流动性画像分析与强化学习环境搭建

**2. `[ASSET]_1m_ohlcv.csv`（标准1分钟行情流数据）**
* 1分钟聚合的标准K线数据
* 包含中间价跟踪（`open`、`high`、`low`、`close`）与主动成交单（taker）流量代理指标

**3. `[ASSET]_5m_ohlcv.csv`（标准5分钟行情流数据）**
* 5分钟聚合的标准K线数据
* 经过清洗、时间对齐的定价数据，可直接与您的深度指标合并使用

## ⚡ 本数据的独特优势
* **无时钟漂移**：时间对齐的K线，消除了原始WebSocket馈送的异步性与时间戳问题
* **超深度覆盖**：通过数学映射得到**30档**买卖盘流动性数据
* **精细分辨率**：**1分钟**粒度快照，支持高保真回测
* **适配AI/ML工作流**：可直接导入Pandas、DuckDB、强化学习环境或XGBoost
* **真实市场意图**：摆脱中心化交易所（CEX，Centralized Exchange）的噪声与虚假挂单，去中心化交易所（DEX）订单簿反映真实的流动性与机构持仓动向

## 🚀 加速您的研究工作
从原始去中心化交易所数据流构建二级订单簿管道，需要定制解析器、严格的时间戳对齐与大量计算资源。
无需投入繁重的数据工程工作，即可下载可直接使用的CSV数据集，直接开始模型训练、时间加权平均成交价（TWAP，Time-Weighted Average Price）/成交量加权平均成交价（VWAP，Volume-Weighted Average Price）策略回测与阿尔法因子提取。

👉 **前往ImbalanceLabs.com下载完整12个月数据集**[ImbalanceLabs.com](https://imbalancelabs.com/)

---
### 许可证与法律声明
本样本数据集基于CC-BY-4.0许可证提供，仅用于研究与分析用途。本数据属于聚合流动性与订单簿深度指数（衍生数据）——一种专有、通过数学推导得到的分析产品，不构成投资建议。Imbalance Labs与任何加密货币交易所无关联关系。
---
license: cc-by-4.0