worldbank/twitter-labor-market

Twitter Labor Market是一个包含28,034条标注推文的数据集，用于推断Twitter用户的就业状态以及检测推文中的工作机会。数据集包括8,876条来自美国用户的推文，12,202条来自墨西哥用户的推文和7,156条来自巴西用户的推文。每条推文被标注为属于五个二元类别之一：`is_unemployed`（是否失业）、`is_hired_1mo`（是否在过去一个月内被雇佣）、`lost_job_1mo`（是否在过去一个月内失去工作）、`job_search`（是否在寻找工作）和`job_offer`（是否包含工作机会）。每条推文由两名标注者标注，最终标签需要至少两名标注者一致。

The Twitter Labor Market dataset is a collection of 28,034 annotated tweets used for inferring the employment status of Twitter users and detecting job offers in tweets. These tweets are from users in the United States, Mexico, and Brazil, spanning from 2007 to 2020. Each tweet is labeled as one of five binary classes, including whether the user is unemployed, was hired in the past month, lost their job in the past month, is looking for a job, or contains a job offer. Each tweet is annotated by two labelers, and the final label requires at least two labelers to provide the same answer.