five

electricsheepafrica/africa-wash-nigeria

收藏
Hugging Face2026-04-26 更新2026-05-03 收录
下载链接:
https://hf-mirror.com/datasets/electricsheepafrica/africa-wash-nigeria
下载链接
链接失效反馈
官方服务:
资源简介:
Water Point Data Exchange (NGA) 数据集是一个关于尼日利亚农村水点(如水井、泉水、水龙头)的众包数据集合,由Water Point Data Exchange平台发布,数据来源于人道主义数据交换(HDX)。该数据集基于WPdx-Plus子集,经过清洗和标准化处理,遵循WPdx数据标准,适用于分析和决策支持。数据集包含两个主要文件:wpdx_adm_region_analysis提供各行政级别的人口服务概况(包括已服务、未服务和未绘制区域)以及基于水点记录年龄的数据质量分析;wpdx_water_points提供每个水点的详细数据,包括WPdx数据标准参数以及修复优先级、服务差距/新建建议和状态预测等分析结果。数据集中每一行代表国家级汇总数据,时间覆盖由report_date和created_timestamp列指示,地理范围限定为尼日利亚(NGA)。数据集包含99,246行和40列,涵盖地理、时间、人口统计、标识符和其他变量,用于支持农村基本水服务访问分析、水点修复优先级评估和服务差距识别。数据集已由Electric Sheep Africa转换为Parquet格式,并分为训练集(79,396行)和测试集(19,849行),适用于机器学习任务。

The Water Point Data Exchange (NGA) dataset is a crowdsourced collection focusing on rural water points (e.g., wells, springs, tapstands) in Nigeria, published by the Water Point Data Exchange platform and sourced from the Humanitarian Data Exchange (HDX). Based on the WPdx-Plus subset, the dataset is cleaned and harmonized using the WPdx Data Standard, making it robust and analysis-ready. It includes two primary files: wpdx_adm_region_analysis provides an overview of population served, unserved, and uncharted for each administrative level, along with data quality analysis based on water point record age; wpdx_water_points offers detailed water point-level data, including all WPdx Data Standard parameters and results from Rehab Priority, Service Gap/New Construction, and Status Prediction analyses. Each row in the dataset represents country-level aggregates, with temporal coverage indicated by report_date and created_timestamp columns, and geographic scope limited to Nigeria (NGA). The dataset contains 99,246 rows and 40 columns, covering geographic, temporal, demographic, identifier, and other variables, and is used to support insights on rural basic water service access, prioritized water point repair recommendations, and identification of service gaps. Curated by Electric Sheep Africa into Parquet format, it is split into train (79,396 rows) and test (19,849 rows) sets for machine learning applications.
提供机构:
electricsheepafrica
二维码
社区交流群
二维码
科研交流群
商业服务