Libraries in Diff posts <i>(2009–August 2023</i>)

Diff (https://diff.wikimedia.org/) is the collaborative, multilingual and multimedia platform for news, updates, and discussions related to the Wikimedia movement, and libraries appear many times within their posts.This dataset is an attempt to gather for the first time, in one place, all those entries that tell the vibrant stories of libraries, librarians and library enthusiasts who contribute to Open Knowledge through Wikimedia projects.<br><br>It was derived from another, larger dataset called Diff posts' titles, authors, full-text, dates, and tags (from 2008-04-11 to 2023-08-31). The rationale behind the three curated library sub-datasets that were derived is described below:<br>1. diff_lib_full-text_lang_data.csv has 134 posts that were identified as related to libraries from the complete dataset, by searching in titles, excerpts and tags<br>2. diff_lib_full-text_lang_false-positives.csv contains those posts that had terms related to libraries but were actually talking about something different (like "libraries" in the sense it is used in computing, a collection of pre-written code to perform specific tasks or "The Wikipedia Library" which is a digital library for active Wikipedia editors). The criteria with which they were filtered out is in the column notes, but they are shared here as well, as many might be considered by other researchers as relevant. 3. diff_lib_full-text_lang_traslations.csv any time an English version of a post was published in Diff, it was added to the first CSV collection (diff_lib_full-text_lang_data.csv) as it is easier for some analysis. However, when a post had a version in another language, this was added to this third dataset. All posts have a language code column as well as another column with the ID of the post to which they are related.Diff（https://diff.wikimedia.org/）是面向维基媒体运动（Wikimedia movement）相关新闻、动态与讨论的协作式多语言多媒体平台，其帖文中多次涉及图书馆主题。本数据集首次尝试将所有讲述图书馆、图书馆员以及通过维基媒体项目助力开放知识（Open Knowledge）的图书馆爱好者的鲜活故事帖文，统一汇聚于一处。 本数据集源自规模更大的Diff帖文数据集，该数据集包含帖文标题、作者、全文、发布日期与标签，时间跨度为2008年4月11日至2023年8月31日。下文将阐述本次筛选生成三个精选图书馆子数据集的依据： 1. diff_lib_full-text_lang_data.csv：从完整数据集中通过标题、摘要与标签检索，共筛选出134条与图书馆相关的帖文，存储于此文件中。 2. diff_lib_full-text_lang_false-positives.csv：包含虽包含图书馆相关术语，但实际讨论主题无关的假阳性帖文（例如计算机领域中“库”指用于实现特定功能的预编写代码集合，或是面向活跃维基百科编辑者的数字图书馆“维基百科图书馆（The Wikipedia Library）”）。筛选剔除此类样本的标准已记录于notes列中，但此处仍一并共享，因为部分研究者或会认为其具备研究价值。 3. diff_lib_full-text_lang_traslations.csv：当Diff平台发布帖文的英文版本时，该版本会被添加至第一个CSV文件（diff_lib_full-text_lang_data.csv）中，以方便部分分析工作。但若某帖文存在其他语言版本，则会被归入此第三个数据集。所有帖文均设有语言代码列，以及用于标识对应帖文的ID列。