Murreviikko: an Annotated and Normalized Corpus of Dialectal Finnish Tweets

Murreviikko (literally 'Dialect week') is a campaign founded in the University of Eastern Finland to promote the use of Finnish dialects in social media. It started in 2020 and takes place mid-October. The original data was collected from Twitter with the search word murreviikko ('dialect week') and hashtag #murreviikko separately for 2020, 2021 and 2022. The current dataset combines all the original collections. The tweets are dialectologically annotated on two levels: following the East-West division of Finnish dialects, and following a seven-way division of Finnish dialects (South-West, Häme, Southern Ostrobothnia, Central and Northern Ostrobothnia, Far North, Savo, and South-East), appended with the Helsinki slang. There is also a class for dialectal tweets, which are not discernible (NA) because of contrasting or scarce dialectal features. The original tweets are normalized to a phonetic standard, but word order is not altered, or grammar rules of standard Finnish followed otherwise. This means that for instance standard Finnish possessive suffixes (minun kirja-ni 'my book-my') are not added if they are not present in the original tweet (minun kirja). Likewise, dialect words are not corrected to the standard alternative, even if such words would exist (pruukata > pruukata instead of standard tavata). Following the rules of the Twitter API, this repository only includes the tweet id's, dialect annotations and normalizations. The original tweets are available for scientific use by request, as granted by the European Union’s Digital Single Market directive (2019/790).方言周（Murreviikko，字面意为“方言周”）是东芬兰大学（University of Eastern Finland）发起的一项旨在推广芬兰方言在社交媒体中使用的活动。该活动于2020年启动，每年十月中旬举办。原始数据于2020、2021及2022年分别通过推特（Twitter）以搜索词"murreviikko"和话题标签"#murreviikko"采集得到，当前数据集整合了所有原始采集内容。 推文在两个层面开展方言学标注：其一按照芬兰方言的东西分区进行标注，其二按照七分区体系标注，包括西南方言区、海梅（Häme）方言区、南奥斯特罗博特尼亚（Southern Ostrobothnia）方言区、中北奥斯特罗博特尼亚（Central and Northern Ostrobothnia）方言区、极北方言区、萨沃（Savo）方言区、东南方言区，并附加赫尔辛基俚语。此外设有无法辨别方言特征的推文类别（NA），此类推文因方言特征冲突或稀缺而难以判定所属方言区。 原始推文被转换为语音标准化形式，但未调整词序，也未遵循标准芬兰语的其他语法规则。例如，若原始推文中未包含标准芬兰语的属格后缀（如"minun kirja-ni"意为“我的书-我的”），则不会添加该后缀，保留原文为"minun kirja"。同理，即便存在标准语对应替代词，方言词汇也不会被修正为标准形式（如保留"pruukata"而非标准语的"tavata"）。 根据推特应用程序编程接口（Twitter API）的相关规则，本数据集仅包含推文ID、方言标注及规范化文本。原始推文需通过申请方可用于科研用途，且需符合欧盟《数字单一市场指令》（2019/790）的相关许可要求。