electricsheepafrica/africa-worldbank-social-protection-labor
收藏资源简介:
该数据集名为“社会保护与劳动——非洲(世界银行WDI)”,包含115,409个观测值,覆盖54个非洲国家,时间跨度为1960年至2025年,涵盖156个不同的指标。数据来源于世界银行的世界发展指标(WDI),这是全球最常引用的全球发展数据参考,编译了来自国家账户、政府机构和合作伙伴组织(如IMF、WHO、FAO、UNESCO、ILO、联合国机构)的官方认可国际统计数据。数据集主题为社会保护与劳动,包括儿童农业就业、农业就业比例、就业人口比率等指标。数据通过世界银行公共API获取,过滤了空值记录,年份规范化为整数,值转换为float64类型。地理覆盖包括54个非洲国家,如南非、埃及、毛里求斯等。数据集模式包括indicator_id(指标代码)、indicator_name(指标名称)、country_iso2(ISO alpha-2国家代码)、country_iso3(ISO alpha-3国家代码,首选连接键)、country_name(国家名称)、year(年份)、value(观测值)、unit(测量单位)和obs_status(观测状态标志)。数据质量方面,记录在获取时过滤了空观测值,指标覆盖因国家和年份而异,并非每个(指标×国家×年份)组合都存在,且仅包含独立主权国家。数据集由Electric Sheep Africa重新打包,旨在为非洲提供统一的、机器学习就绪的数据层,便于研究人员和开发者使用。
This dataset, named Social Protection & Labor — Africa (World Bank WDI), contains 115,409 observations of Social Protection & Labor data across 54 Africa countries, spanning from 1960 to 2025, and covering 156 distinct indicators. The data is sourced from the World Banks World Development Indicators (WDI), which is the worlds most-cited reference for global development data, compiling officially-recognized international statistics from national accounts, government agencies, and partner organizations such as the IMF, WHO, FAO, UNESCO, ILO, and UN agencies. The topic focuses on Social Protection & Labor, including indicators like child employment in agriculture, employment in agriculture as a percentage of total employment, and employment-to-population ratios. Data is pulled directly from the World Banks public API (api.worldbank.org/v2) for every indicator tagged to this topic across all countries in the region, with records having no observed value dropped. Years are normalized to integers, and values are coerced to float64. Geographic coverage includes 54 Africa countries, such as South Africa, Egypt, Mauritius, and others. The dataset schema includes columns: indicator_id (World Bank WDI indicator code), indicator_name (human-readable indicator name), country_iso2 (ISO 3166-1 alpha-2 country code), country_iso3 (ISO 3166-1 alpha-3 country code, preferred join key), country_name (English country name as published by the World Bank), year (year of observation in CE), value (observed indicator value), unit (unit of measurement, often empty), and obs_status (World Bank observation-status flag, e.g., preliminary or estimated). Data quality notes indicate that records with null observations are filtered out at ingest, indicator coverage varies by country and year, not every (indicator × country × year) combination is present, and the dataset is filtered to individual sovereign states only. Repackaged by Electric Sheep Africa, it is part of a mission to provide a unified, ML-ready data layer for Africa on HuggingFace, enabling researchers and developers to start working quickly using load_dataset().




