新华·舆情即时反馈数据集

数据集为纯文本模态，计划积累10亿条评论，日增300万条。单条平均30字，附带用户ID、发布时间、平台、点赞、转发、情感极性、实体、谣言标签、立场等10余项结构化字段。压缩后文本数据规模约8TB，索引与向量缓存另占2TB。

This is a pure-text modality dataset. It is planned to accumulate 1 billion comments, with a daily increment of 3 million comments. Each comment averages 30 words, and it is equipped with more than 10 structured fields including user ID, release time, platform, likes, reposts, sentiment polarity, entities, rumor labels, stance and others. The compressed text data has a total scale of approximately 8 TB, and the indexes and vector caches occupy another 2 TB.

搜集汇总

背景与挑战

背景概述

该数据集是一个大规模舆情分析资源，专注于收集和处理纯文本评论，计划积累10亿条数据，每日新增300万条，具有显著的实时性和扩展性。每条评论平均30字，并附带用户ID、发布时间、平台、点赞、转发、情感极性、实体、谣言标签、立场等10余项结构化字段，提供多维度的深度分析能力。数据压缩后规模约8TB，索引与向量缓存另占2TB，适用于大规模存储和高效检索应用。