A Zhihu-based Dataset for Public Archival Demand on Social Media

The dataset used in this study was collected from Zhihu, consisting of questions, answers, and discussion texts related to topics such as “personnel archives” and “public archives.” It contains a total of 2,331 Chinese texts and their corresponding publication timestamps, covering a period from approximately February 2012 to January 2025. The dataset covers public concerns such as archive inquiry, replacement, transfer, custody, the whereabouts of graduates’ archives, civil service political review, recruitment and appointment, disciplinary sanctions, institutional archive management, and archive digitization.

Each record is a complete user consultation or experience-sharing text with a corresponding timestamp. In addition, the project provides three types of data files: raw texts, publication timestamps, and texts processed through word segmentation and topic modeling, which are aligned line by line. The dataset has been anonymized and can be used for research tasks such as public archival demand identification, topic classification, semantic annotation, and demand trend prediction in social media contexts.