sstoeckl/opencryptoassetpricing

Open Crypto Asset Pricing数据集是一个用于加密货币资产定价的开放数据集，基于Liu et al. (2022)的论文《Common Risk Factors in Cryptocurrency》构建。它包含加密货币风险因子（CMKT、CSMB、CMOM），这些因子通过价值加权、30/40/30分界点、2x3双重排序等方法计算，使用精确的周历。数据集还提供576个参数组合的因子变体，涵盖加权方式（如VW、EW）、规模分界点（如中位数、三分位数）、动量回顾期（2、3、4周）、日历类型（如Liu的精确周历、周一至周一日历）、频率（周度、月度）、排除规则（如稳定币、包装币、衍生品）和退市回报处理。此外，包含96个十分位排序的投资组合变体文件。数据来源于CoinMarketCap，通过R包crypto2获取，无生存偏差（包括退市代币），并对周回报设置了9,900%的上限以过滤数据错误。数据集支持R和Python使用，如通过arrow或pandas读取Parquet文件。The Open Crypto Asset Pricing dataset is an open dataset for cryptocurrency asset pricing, constructed following Liu et al. (2022) Common Risk Factors in Cryptocurrency. It includes cryptocurrency risk factors (CMKT, CSMB, CMOM) built using value-weighting, 30/40/30 breakpoints, 2x3 double sort for CMOM, and a sharp-year calendar. The dataset offers 576 parameter combinations for factor variants, covering weighting (e.g., VW, EW), size breakpoints (e.g., median, tercile), momentum lookback (2, 3, 4 weeks), calendar (e.g., Lius sharp, Monday-Monday), frequency (weekly, monthly), exclusions (e.g., all stables+wrapped+derivatives, nostable, none), and delisting returns handling. Additionally, it contains 96 decile-sorted portfolio variant files. Data is sourced from CoinMarketCap via the crypto2 R package, survivorship-bias free (includes delisted coins), with weekly returns capped at 9,900% to filter data errors. The dataset supports usage in R and Python, e.g., reading Parquet files with arrow or pandas.