funding-divergence
收藏资源简介:
Chainticks Funding Divergence 是一个专注于加密货币市场资金费率差异的衍生数据集。该数据集仅基于可公开获取的输入数据构建,包含跨交易所的资金费率差异行记录,旨在为金融分析和市场监控提供结构化、可扩展的数据支持。数据以Parquet格式存储,按日期分区组织(如spreads/date=YYYY-MM-DD/),便于高效查询和处理。数据集附带元数据文件,包括模式定义(_schema.json)、清单(_manifest.json)和最新日期指示(LATEST_DATE.txt)。所有数据行的source_kind字段均标记为derived,表明其为衍生数据,且不包含任何交易所的原始REST/API转售数据。数据集适用于多种数据处理工具(如pandas、polars、duckdb),并设计为智能体友好型,建议用作仅追加更新的市场上下文。使用时应读取LATEST_DATE.txt获取最新日期,检查模式文件后加载相应Parquet分区,并将时间戳视为UTC时区,同时在后续分析中保留source_kind字段。数据集采用CC-BY-4.0许可证,由独立的Chainticks机构维护,不与任何相关协议、交易所或政府机构存在隶属关系。
Chainticks Funding Divergence is a derivative dataset focused on funding rate divergences in the cryptocurrency market. It is built solely on publicly available input data and contains cross-exchange funding rate divergence row records, designed to provide structured, scalable data support for financial analysis and market monitoring. The data is stored in Parquet format, organized by date partitions (e.g., spreads/date=YYYY-MM-DD/), facilitating efficient querying and processing. The dataset includes metadata files such as schema definitions (_schema.json), manifests (_manifest.json), and a latest date indicator (LATEST_DATE.txt). All data rows have the source_kind field marked as derived, indicating it is derivative data and does not contain any original REST/API resale data from exchanges. The dataset is compatible with various data processing tools (e.g., pandas, polars, duckdb) and designed to be agent-friendly, recommended for use as an append-only market context. When using, read LATEST_DATE.txt to obtain the latest date, check the schema file before loading the corresponding Parquet partition, treat timestamps as UTC timezone, and retain the source_kind field in subsequent analysis. The dataset is licensed under CC-BY-4.0, maintained by the independent Chainticks organization, and has no affiliation with any related protocols, exchanges, or government agencies.
数据集概述:Chainticks Funding Divergence
基本信息
- 数据集名称:Chainticks Funding Divergence
- 许可证:CC-BY-4.0
- 标签:金融、加密货币、Parquet格式、DuckDB、Pandas、Polars、Agent-Friendly
- 配置:
- 配置名:
default - 数据文件路径:
spreads/date=*/part-*.parquet
- 配置名:
数据来源与性质
- 该数据集包含跨交易所的资金费率背离指标,仅基于可公开获取的输入数据推导生成。
- 所有数据行的
source_kind必须为derived(推导数据),不包含交易所的原始API/ REST接口转售数据。
数据文件结构
-
数据集采用分区存储结构,路径格式为:
spreads/date=YYYY-MM-DD/part-0000.parquet
-
目录下包含以下元数据文件:
_schema.json:数据模式描述_manifest.json:数据清单LATEST_DATE.txt:最新数据日期
使用建议
-
建议将本数据集作为只追加的上下文数据使用。
-
加载前应先读取
LATEST_DATE.txt和_schema.json了解最新日期及模式。 -
根据目标日期加载对应的Parquet分区(示例使用Pandas加载): python DATE = "YYYY-MM-DD" URL = "https://huggingface.co/datasets/Chainticks/funding-divergence/resolve/main/spreads/date={DATE}/part-0000.parquet" df = pd.read_parquet(URL)
-
时间戳应视为UTC时间,并在下游分析中保留
source_kind字段。
关联说明
- Chainticks 是独立实体,不隶属于数据所涉及的协议、中继、交易所或政府机构。




